NuttX is not affected by GHSA-93v7-287q-qx4g: it has a flat VFS, the
FAT driver used for /fs/microsd does not support symlinks, and
CONFIG_PSEUDOFS_SOFTLINKS is off by default. Running realpath() and
parent canonicalization there is dead weight that costs flash for no
security benefit.
Move the canonicalize_path helper, the in-root check, the new
includes, and O_NOFOLLOW behind #ifndef __PX4_NUTTX so the NuttX
build is functionally identical to main, except that _workOpen and
_workWrite now also call _validatePathIsWritable for write paths
(matching the pattern already used by the other writable opcodes).
On POSIX/SITL the symlink-resolution hardening is unchanged.
Drop the now-unused _validatePathIsInRoot declaration from the
header.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
The previous commits in this branch blocked all writes under
PX4_STORAGEDIR/etc/ as a defence against attacker-controlled boot
hook files (rc.txt, config.txt, extras.txt). That subtree is the
documented user customization point for PX4 startup, and QGC's
MAVFTP file manager is the supported way to upload it. Blocking
writes there breaks a legitimate workflow on top of an already
unauthenticated channel where the real mitigation is MAVLink
signing.
Drop the boot-hook deny block. The symlink/canonicalization
hardening (\_validatePathIsInRoot, O_NOFOLLOW, parent-resolved
realpath fallback) for GHSA-93v7-287q-qx4g stays.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Address review feedback on the FTP path validation hardening:
- Extract canonicalization into a static helper canonicalize_path() that
resolves either the full path or the parent directory plus the
reattached leaf, and reuse it from both the in-root check and the
boot-hook deny check. The previous deny check matched the raw input
string against the literal "PX4_STORAGEDIR/etc/" prefix, which could
be bypassed by paths such as "./etc/x" or "//etc/x" when the leaf did
not yet exist. The new check always compares the canonical absolute
path against the canonical "PX4_STORAGEDIR/etc/" prefix.
- Update the canonicalize_path() doc comment to state that the result
is always an absolute path produced by realpath(), so future
maintainers do not have to re-derive that property.
- Add the missing blank line between the realpath success branch and
the NuttX fallback so the file matches the project astyle rules.
Refs: GHSA-c6f8-f7w2-785x, GHSA-93v7-287q-qx4g
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Tighten path validation in MavlinkFTP so that FTP write operations
consistently reject unsafe targets and resolve symlinks before checking
the result against the FTP root.
Add a new helper _validatePathIsInRoot() that canonicalizes the
requested path with realpath() and verifies the result is contained
inside PX4_STORAGEDIR. For paths that do not yet exist (CreateFile,
CreateDirectory) the parent directory is canonicalized and the leaf
name reattached so the check still produces a meaningful result. NuttX
falls back to the previous string based prefix and traversal check
because realpath() is not available there.
Rewrite _validatePathIsWritable() to call the new in-root check on all
platforms, and to deny writes that target the boot executed startup
hook subtree (PX4_STORAGEDIR /etc/). Drop the NuttX only ifdef around
the existing prefix check so the same logic runs on POSIX builds too.
Apply _validatePathIsWritable() in _workOpen() for write mode opens
(O_WRONLY or O_RDWR), and add the missing _validatePath() call to
_workWrite() so the WriteFile opcode now matches the pattern used by
the other write capable opcodes.
Open the leaf file with O_NOFOLLOW where the platform provides it so a
TOCTOU race cannot redirect the leaf through a symlink between
validation and open(2).
Refs: GHSA-c6f8-f7w2-785x, GHSA-93v7-287q-qx4g
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* Remove unused parameters from function signature and make the parameter accessors consistent
* Update the caller function signature
* Update src/modules/navigator/rtl.cpp
---------
Co-authored-by: Jacob Dahl <37091262+dakejahl@users.noreply.github.com>
Use timestamp_sample instead of time_now_us for the rate limiter check
to sync to the sensor clock rather than the wall clock.
Switch from direct timestamp assignment to epoch-advance
(_last_publication_timestamp += interval_us) with a catch-up guard to
prevent aliasing artifacts when the sensor sample rate is close to the
configured publication rate.
getPreviousPositionItems() already decrements the start index
internally before searching. The call in on_activation() at line 227
passed _inactivation_index - 1, causing a double-decrement that made
the vehicle resume at waypoint n-2 instead of n-1.
All other call sites (rtl_mission_fast_reverse.cpp:81,
rtl_mission_fast_reverse.cpp:133, mission_base.cpp:1149) pass the
index directly without pre-decrementing.
The bug has been present since commit 007ed11bbe (June 2023).
Closes#26795
Signed-off-by: Pavel Guzenfeld <pavelgu@gmail.com>
The FC-side DroneCAN sensor bridges (accel, gyro, rangefinder) used
hrt_absolute_time() in the receive callback as timestamp_sample,
adding ~3-16ms of systematic CAN transport delay.
For messages with a uavcan.Timestamp field, the cannode can publish
the actual sample time via UAVCAN GlobalTimeSync. The RawIMU publisher
already did this for IMU data; apply the same pattern to the range
sensor publisher, and update all three FC bridges to prefer the
message timestamp with a fallback to hrt_absolute_time() for nodes
that don't set it.
* fix: added comment explaining why dev id address can only be 3 or 4
* fix: change link to point to main px4 repo
* fix: typo
* formatted
* chore: formatting
The packaging script only placed all_events.json.xz in an events/
subdirectory, but the firmware advertises the metadata URI at the
board directory top level. New build targets added after the
Jenkins-to-GHA migration had no legacy top-level copy, causing
QGC to get a 404 when fetching component metadata.
Fixes#26963
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
The asset file was renamed from X25-EVO.jpg to x25_evo.jpg in git but
all four locale files (en, ko, uk, zh) still referenced the old name.
macOS hid this because its filesystem is case-insensitive, but Linux CI
(case-sensitive) intermittently failed to resolve the reference during
Rollup bundling.
Fixes#26958
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
When EKF2_HGT_REF=2 (range sensor) with no GPS, optical flow could
never start. The starting condition required isTerrainEstimateValid()
or isHorizontalAidingActive(), but terrain is never "estimated" when
range is the height reference (ground is the datum, terrain state is
fixed at 0), and there's no horizontal aiding without GPS.
HAGL is directly known from the range measurement in this case, so
optical flow has everything it needs to fuse. Add the range height
reference check to the optical flow starting conditions.
Fixes: https://github.com/PX4/PX4-Autopilot/issues/25248
Add :latest tag alongside version tags for per-arch images and
multi-arch manifests on both Docker Hub and GHCR.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Add cmake/cpack infrastructure for building .deb packages from
px4_sitl_sih and px4_sitl_default targets. Includes install rules,
package scripts, Gazebo wrapper, and CI workflow.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Set BOARD_FLASH_SECTORS to 13 so the bootloader does not erase the
parameter sectors (14 and 15) during firmware updates. Previously set
to 14 which allowed the bootloader to erase sector 14, potentially
wiping stored parameters.
Three issues caused the monthly audit to report already-resolved submodules:
1. The audit workflow grepped for "NOASSERTION" anywhere in the output,
matching the Detected column even when the Final column had a valid
override (e.g. libtomcrypt detected as NOASSERTION but overridden to
Unlicense). Changed to grep for "<-- UNRESOLVED" marker instead.
2. Submodules with an explicit NOASSERTION override in license-overrides.yaml
(like libfc-sensor-api, which is proprietary) were still counted as
failures. Now treated as "acknowledged" since someone intentionally
added the override entry.
3. Added missing BSD-3-Clause override for sitl_gazebo-classic (PX4 org
project with no LICENSE file in repo).
Fixes#26932
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Add MAVLink stream that maps EstimatorFusionControl uORB message to
ESTIMATOR_SENSOR_FUSION_STATUS, exposing per-sensor intended/active
bitmasks to the GCS.
Split FusionSensor into available (CTRL param != disabled) and enabled
(runtime-toggleable). intended() = enabled && available. EKF core aid
sources now set available themselves and use intended() or _params
directly for CTRL-level checks. Remove drag/imu from FusionControl,
add aspd/rngbcn. Add AGP sourceFusingBitmask() for active-status.
* ci(claude): add review-pr skill for domain-aware PR reviews
Add a Claude Code skill that reviews pull requests with checks
tailored to the domains touched (estimation, control, drivers,
simulation, system, CI/build, messages, board additions).
Built from analysis of 800+ PR reviews across 8 PX4 maintainers.
Includes merge strategy recommendation, interactive dialog for
submitting reviews, and human-sounding PR comment formatting.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* ci(copilot): add domain-scoped review instructions for GitHub Copilot
Add .github/instructions/ files that give GitHub Copilot PR reviews
the same domain-aware context as the Claude Code review-pr skill.
Each file is scoped via applyTo to the relevant source paths:
core review, estimation, control, drivers/CAN, simulation, system,
CI/build, messages/protocol, and board additions.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* fix(claude): address Copilot review feedback
- Fix step reference in review-pr skill (step 8 -> step 9)
- Capitalize CMake consistently in skill and Copilot instructions
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
---------
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Several helper scripts assumes bash is available at /bin/bash. That breaks on systems
such as NixOS, where bash is resolved from PATH instead of a fixed /bin location and
causes failures like `bad interpreter` during `make format`, e.g., on my host machine:
```sh
$ make format
/PX4-Autopilot/Tools/astyle/check_code_style.sh: /PX4-Autopilot/Tools/astyle/fix_code_style.sh: /bin/bash: bad interpreter: No such file or directory
```
This change switches these entrypoints to `#!/usr/bin/env bash` so they locate bash properly.
No functional changes intended.
Signed-off-by: Onur Özkan <work@onurozkan.dev>
Disable EKF2 fusion features with no corresponding hardware:
- px4_fmu-v2: optical flow, range finder (~17 KB saved)
- mamba-f405-mk2: optical flow, range finder, external vision,
aux global position, aux velocity, baro compensation,
drag fusion (~42 KB saved)
Before this change, filtered test runs still built every gtest target
because `test_results` depended on all unit and functional gtest targets.
This updates both `px4_add_unit_gtest()` and `px4_add_functional_gtest()`
to use the filtered dependency helper so filtered runs only build the
selected targets.
Signed-off-by: Onur Özkan <work@onurozkan.dev>
The MPL3115A2 ADC conversion at OSR 2 (ratio 4) takes ~18ms. The
driver polls until the conversion completes, so the read time is at
the end of the integration window. Correct timestamp_sample to the
midpoint by subtracting CONVERSION_TIME / 2.
The LPS25H one-shot measurement is integrated over a ~40ms window
(at 25Hz-equivalent internal averaging). The read time corresponds to
the end of the integration window. Correct timestamp_sample to the
midpoint by subtracting CONVERSION_INTERVAL / 2.
The LPS22HB one-shot measurement is integrated over a ~40ms window
(at 25Hz-equivalent internal averaging). The read time corresponds to
the end of the integration window. Correct timestamp_sample to the
midpoint by subtracting CONVERSION_INTERVAL / 2.
Same fix as MS5611: the MS5837 ADC conversion at OSR 1024 takes
~2.28ms, but the data is read after a 10ms scheduling delay. Correct
timestamp_sample by subtracting (CONVERSION_INTERVAL - CONVERSION_TIME/2)
from the read time.
The MS5611 ADC conversion at OSR 1024 takes ~2.28ms, but the data is
read after a 10ms scheduling delay. The current code timestamps the
read time, which is ~8.9ms after the true integration midpoint.
Correct timestamp_sample by subtracting the full offset
(CONVERSION_INTERVAL - CONVERSION_TIME/2) from the read time.
The BMP581 pressure measurement is integrated over a configurable
window (~23ms at 32x pressure / 2x temperature oversampling). The
read time corresponds to the end of the integration window, introducing
a systematic timing bias. Correct timestamp_sample to the midpoint by
subtracting measurement_time / 2.
The BMP280 pressure measurement is integrated over _measure_interval
(~43ms at 16x pressure / 2x temperature oversampling). The read time
corresponds to the end of the integration window, introducing a
systematic timing bias. Correct timestamp_sample to the midpoint by
subtracting measurement_time / 2.
Set all high_rate_sensors_topics to 100hz (10ms interval) and add
vehicle_air_data, vehicle_thrust_setpoint, estimator_aid_src_baro_hgt,
and vehicle_magnetometer.
The BMP388 pressure measurement is integrated over a configurable
window (e.g. 37ms at 16x oversampling). The previous code used the
read time as timestamp_sample, which is the end of the integration
window. Correct to the midpoint by subtracting half the measurement
time, with a guard against unsigned underflow.
rc.mc_defaults sets MAV_TYPE=2 (quadrotor) which the hex airframe
never overrides. Set MAV_TYPE=13 (hexarotor) so the heartbeat
correctly identifies the vehicle type.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* Swap joystick surge/heave mapping in manual, stabilized and acro modes to make it similar to position modes
* docs: update UUV/BlueROV2 modes and joystick mapping
* Document basic control axes and joystick mapping
Added basic control axes and stick mapping for BlueROV2.
* Fixed formatting issue
* Enhance clarity of control axes and stick mapping
Clarified descriptions of motion axes and joystick controls for BlueROV2.
The MAVLink standard defines ACTUATOR_OUTPUT_FUNCTION_MOTOR1=1..MOTOR16=16,
but PX4 internally uses OutputFunction::Motor1=101..Motor12=112. The DShot
driver only handled PX4 internal values (101+) and QGC legacy values (1101+),
so any standards-compliant GCS sending the MAVLink enum values would get
VEHICLE_CMD_RESULT_UNSUPPORTED back from MAV_CMD_CONFIGURE_ACTUATOR.
Add a mapping from MAVLink standard values (1-16) to PX4 internal values
(101-116) by adding 100, matching the existing QGC backwards-compat pattern.
* fix(commander): add tab character to critical system loss messages
* fix(commander): extend timeout for traffic avoidance system heartbeat check
* Commander: Only Warn the user about traffic avoidance system loss if COM_ARM_TRAFF is set
Signed-off-by: Claudio Micheli <claudio@auterion.com>
---------
Signed-off-by: Claudio Micheli <claudio@auterion.com>
Co-authored-by: Claudio Micheli <claudio@auterion.com>
GCC 14.3.0 emits `-Wstringop-overflow` when `RtcmTest::buildRawFrame()`
is optimized and inlined.
This change marks the helper `noinline` to keep it out of that optimization path.
Preserves the existing logic and only changes how the compiler emits the test helper.
Fixes https://github.com/PX4/PX4-Autopilot/issues/26875
Signed-off-by: Onur Özkan <work@onurozkan.dev>
When all subscription topics are commented out in dds_topics.yaml,
the build failed in two ways:
1. KeyError in generate_dds_topics.py when the subscriptions key is
absent from YAML — fixed by using dict.get() with fallback to
empty list, consistent with how subscriptions_multi is handled.
2. Unused variable errors (-Werror) in the generated dds_topics.h
when no subscriptions exist — fixed by guarding on_topic_update(),
time_offset_us, and uxr_set_topic_callback() with conditional
template blocks. Also marked create_data_reader() as
__attribute__((unused)) since it is only called from generated
subscription code.
Closes#26799
Signed-off-by: Pavel Guzenfeld <pavelgu@gmail.com>
* [feat] allowed to assign up to 16 ESC CAN
* Update EscStatus.msg
lowered down to 12 motors, hardware tested
* Update module.yaml
lowered down to 12 motors, hardware tested
---------
Co-authored-by: klelkov <kon.lelkov@yandex.ru>
Co-authored-by: Jacob Dahl <37091262+dakejahl@users.noreply.github.com>
Add INA226 and INA228 power monitor drivers to the voxl2 SLPI board
config and add startup options in voxl-px4-start to select them via
the POWER_MANAGER environment variable.
Give the operator the optiont to configure a "Hold at position where
the data link was still coming through" by setting NAV_DLL_ACT to Hold
and the new param NAV_LTR_LAST_DL to 1.
Signed-off-by: Silvan <silvan@auterion.com>
Existing saved parameters store RC*_REV as float. The parameter
import system does strict type checking and would silently skip
these on firmware update. Add migration to preserve user settings.
RC*_REV parameters are binary toggles (-1 or 1) immediately converted
to bool. Using int32 allows reverting the module schema enum key type
from number back to integer, keeping validation strict.
All parameters are now defined in YAML module configuration files.
Remove the cmake infrastructure that discovered and processed
legacy params.c files:
- Remove GLOB_RECURSE for *params.c/*parameters.c
- Remove .c file scanning from DISABLE_PARAMS_MODULE_SCOPING
- Remove module_list from px_process_params.py --src-path
- Remove PX4_MODULE_PATHS usage (no longer needed for param scanning)
- migrate_c_params.py: preserve newlines and paragraph breaks in long
descriptions, use YAML block scalars for multi-line strings
- generate_params.py: support @value tags on float type parameters
(fixes RC*_REV enum values being lost during yaml generation)
Reject pressure readings outside the sensor's operating range
(30-125 kPa) to detect I2C data corruption. When I2C transfers
complete successfully but return corrupted data, this check
prevents invalid samples from being published.
Co-authored-by: Jacob Dahl <37091262+dakejahl@users.noreply.github.com>
The failure_detector_status bitmask in vehicle_status duplicates the
separate FailureDetectorStatus topic. Remove it and read directly from
the dedicated topic in failureDetectorCheck and HIGH_LATENCY2.
* feat(pps_capture): allow selecting GPS receiver by device ID
Add PPS_CAP_GPS_ID parameter to select which GPS receiver's data
is used for PPS timestamp correlation. Matches by device ID rather
than uORB instance index, which avoids dependence on instance ordering.
When set to 0 (default), uses the first available instance for
backward compatibility.
* docs(pps_capture): document PPS_CAP_GPS_ID for multi-GPS setups
* fix(pps_capture): use GPS_MAX_RECEIVERS constant and mark PPS_CAP_GPS_ID reboot-required
Extend the arm throttle safety check to all vehicle types including
rovers, which were previously excluded. Unify the two separate throttle
checks into a single evaluation at arm-time that accounts for vehicle
type and control mode: rovers require centered stick, climb-rate modes
require stick at or below center, and manual/stab/acro modes require
stick at bottom.
This is not only needed for engine warmup but in general, when the
vehicle is static on the ground, relying on fixed position to maintain a
valid global position estimate before takeoff, even when bumping it or
starting the engine.
* boards: corvon 743v1 support (Docs and LED alignment)
This PR addresses #24769 by providing the required official documentation, while simultaneously aligning the board's LED semantics entirely with the PX4 standard.
Key Changes:
- Add complete corvon 743v1 hardware documentation and manufacturer link.
- Fix LED out-of-bounds bug and strictly align RGB states to Pixhawk standard (LED_BLUE=0, LED_RED=1, LED_GREEN=3).
- Update bootloader pin config (hw_config.h) to use red LED for boot/error, and update pre-built bootloader.bin.
* Prettier and file reduce
* docs: address reviewer feedback & board ID fix
* Apply suggestion from @hamishwillee
Co-authored-by: Hamish Willee <hamishwillee@gmail.com>
* docs: resolve final reviewer feedback (PPM, Debug Port, Manufacturer List)
---------
Co-authored-by: Hamish Willee <hamishwillee@gmail.com>
Co-authored-by: Ramon Roche <mrpollo@gmail.com>
When a CRSF frame arrives with packet_size < 2 (PACKET_SIZE_TYPE_SIZE),
the subtraction `packet_size - PACKET_SIZE_TYPE_SIZE` underflows the
uint32_t working_segment_size to 0xFFFFFFFF. The subsequent overflow
check also wraps and fails to catch it. Since working_segment_size is
static, the parser is permanently stalled — no further CRSF messages
can be processed until reboot.
Validate packet_size >= PACKET_SIZE_TYPE_SIZE early, before any
subtraction, protecting both the known variable-length and unknown
packet branches in a single check.
Supersedes #26782 which only guarded the unknown-packet branch.
The flow output table shows forward movement producing +Y flow and
rightward movement producing -X flow, which confuses users whose sensors
have X-forward/Y-right coordinate systems. Add an info note explaining
that integrated flow values are angular rotations (radians) about the
body axes using the right-hand convention, which is why the axes are
cross-coupled with translational motion.
Add reusable skill definitions for common contributor workflows:
- commit: creates conventional commits with proper type(scope) format
- pr: creates PRs with conventional commit titles
- rebase-onto-main: handles rebasing onto main when parent branches
were squash-merged
Add a scalable .deb packaging framework for VOXL2, built on the
existing cmake/package.cmake CPack infrastructure. The framework
handles multi-processor boards by having the POSIX (_default) build
own the .deb and pull in the companion SLPI build's artifacts.
Board-specific files:
- cmake/package.cmake: CPack variable overrides (name, deps, version)
- cmake/install.cmake: install() rules for all .deb contents
- debian/postinst: px4-* symlinks, DSP signature, directory setup
- debian/prerm: service stop, symlink cleanup
- debian/voxl-px4.service: systemd unit (after sscrpcd)
Infrastructure changes:
- cmake/package.cmake: hook for board-specific CPack overrides
- platforms/posix/CMakeLists.txt: hook for board install.cmake
- Makefile: %_deb pattern rule (build _default, then cpack -G DEB)
- CI: auto-discover _deb targets, collect .deb artifacts, upload
to GitHub Releases
Future boards: add cmake/package.cmake + cmake/install.cmake and
CI discovers it automatically. No new file formats or tools needed.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Read companion_targets files from board directories and exclude those
targets from CI grouped builds. The parent target builds them via
Make prerequisite, avoiding redundant CI jobs.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Add Makefile rules so that both `make modalai_voxl2` and
`make modalai_voxl2_default` build the SLPI DSP firmware first.
Add companion_targets file listing modalai_voxl2_slpi so CI knows
to exclude it from independent build groups.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
Consolidate the VOXL2 SLPI (DSP) board from a separate directory into
the existing voxl2 board as a variant. Multi-processor SoCs like the
QRB5165 should use one board directory with multiple .px4board files
rather than separate directories per processor.
Changes:
- Add slpi.px4board (QURT/DSP) alongside default.px4board (POSIX/apps)
- Merge board_config.h with __PX4_QURT / __PX4_POSIX preprocessor guards
- Merge CMakeLists.txt with PX4_PLATFORM conditionals
- Split bus definitions into platform-specific files (i2c/spi_posix/qurt)
- Reorganize drivers into drivers/posix/ and drivers/qurt/ subdirectories
- Guard cmake/init.cmake and cmake/link_libraries.cmake for posix-only
- Update build and install scripts for new target names
- Delete boards/modalai/voxl2-slpi/ entirely
Build targets change:
- modalai_voxl2-slpi_default -> modalai_voxl2_slpi
- modalai_voxl2_default (unchanged)
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* ci(ros): use matching branch for px4-ros2-interface-lib
When running on release branches, the ROS integration tests now
check if a matching branch exists in px4-ros2-interface-lib and
clone it instead of always using main. This prevents build failures
caused by uORB message divergence between main and release branches.
Fixes https://github.com/Auterion/px4-ros2-interface-lib/issues/184
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* ci(ros): dispatch release branch creation to px4-ros2-interface-lib
Add a standalone workflow triggered by the create event that fires a
repository_dispatch to Auterion/px4-ros2-interface-lib when a
release/X.Y branch is created. Also supports manual workflow_dispatch.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
* ci(ros): add empty permissions block to dispatch workflow
Fixes code scanning alert about missing GITHUB_TOKEN permissions.
This workflow only uses a PAT secret, not GITHUB_TOKEN, so no
permissions are needed.
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
---------
Signed-off-by: Ramon Roche <mrpollo@gmail.com>
There are many settings falling into the RC_* category
that definitely should be reset when e.g. placing the autopilot
into a new airframe.
And even for RC calibration values: it's not the worst
if those are newly calibrated after a reset. Or if they
are not expected to change one can bake them into the
airframe file.
Signed-off-by: Silvan <silvan@auterion.com>
PR #25799 added 'servo-launch-lock' to apply_identifiers (6 items) but
did not add a corresponding 6th entry to the 19 rule item arrays, causing
QGC to reject all servo-type rules with "unexpected num items expected: 6".
Remove the @volatile flag from ASPD_SCALE_1/2/3 so the estimated
airspeed scale persists across reboots and can be transferred between
vehicles of the same model. The scale is primarily determined by pitot
position on the airframe, not the individual sensor.
To avoid corrupting the param transfer hash with negligible changes
every flight, raise the save threshold from FLT_EPSILON to 3% relative
change, per dev-call consensus.
Supersedes #22760
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Update the opencyphal/public_regulated_data_types submodule to latest
master which renames all .uavcan files to .dsdl, eliminating ~456
Nunavut deprecation warnings per build.
The N after MB was parsed as "play note 0" (rest) rather than "Music Normal" mode, since M already consumed B. Replacing it with a trailing P matches the ERROR_TUNE pattern and provides an intentional inter-repetition pause at the correct tempo.
After #25648, when performing RTL mission fast reverse, the vehicle would go to the waypoint before the previous one (i.e., two waypoints back). If the drone was already on its way to the first waypoint at that moment, it would even fly to the second waypoint first, and only then reverse back toward the takeoff point. This PR fixes that bug. #25648 was intended to address issues with NAV_CMD_CONDITION_GATE. The modification proposed in this PR also correctly bypasses NAV_CMD_CONDITION_GATE waypoints.
Line 319 used nested quotes inside f-strings, a feature only available
in Python 3.12+. The CI Docker image (px4-dev-base-focal:2021-08-18)
runs Python 3.8, causing the "msg file docs" Jenkins stage to fail on
every main build since 6bf73d9d89.
Extract the join expressions into local variables to restore
compatibility with Python 3.8+.
* sensors: add per-receiver GPS delay parameters
Add SENS_GPS{0,1}_DELAY params to vehicle_gps_position, following the
same device-ID matching pattern used for antenna offsets. Each receiver
can now have its own measurement delay relative to the IMU.
The delay is applied to timestamp_sample before blending. When PPS time
correction is active it takes priority over the parameter-based delay.
When a GPS driver already provides its own timestamp_sample the
per-receiver delay is not applied on top of it.
* fix(ekf2): remove EKF2_GPS_DELAY and perform param transation
* fix(param_translation): fix GPS param migration return values
Add missing return for EKF2_GPS_POS_Z and remove incorrect return for
EKF2_GPS_DELAY (1-to-many migration should not return PARAM_MODIFIED).
* fix(sensors,ekf2): rename pps_compensation and clarify delay default
* fix(ekf2): account for SENS_GPS*_DELAY in observation buffer sizing
* fix(docs): migrate EKF2_GPS_DELAY param
* sensors: move GPS antenna offsets to per-receiver parameters
Move antenna position configuration from the single EKF2_GPS_POS_X/Y/Z
parameter set into per-receiver SENS_GPS{0,1}_OFF{X,Y,Z} parameters in
the sensors module. Each offset slot is matched to a physical receiver
by device ID (SENS_GPS{0,1}_ID), falling back to uORB instance index
when no IDs are configured.
The antenna offset is now carried through the SensorGps uORB message
and blended alongside other GPS states when multi-receiver blending is
active, so EKF2 receives the correct lever arm for whichever receiver
(or weighted combination) is selected.
- Add antenna_offset_{x,y,z} fields to SensorGps.msg
- Remove EKF2_GPS_POS_X/Y/Z params; EKF2 reads offset from gnssSample
- Add SENS_GPS{0,1}_ID and SENS_GPS{0,1}_OFF{X,Y,Z} params (module.yaml)
- Blend antenna offsets in GpsBlending (weighted average)
- Add unit tests for single, blended, and failover antenna offset cases
- Migrate params.c to module.yaml for the vehicle_gps_position module
* sensors: gps_blending: add asymmetric weight and fallthrough offset tests
Add two additional antenna offset test cases:
- dualReceiverAsymmetricWeightAntennaOffset: verify that unequal eph
values produce correctly skewed blend weights (0.8/0.2) and that the
output antenna offset reflects the weighted average
- blendingFallthroughAntennaOffset: verify that when blending is enabled
but can_do_blending evaluates false (eph=0), the non-blending path
correctly assigns the selected receiver's antenna offset
* feat(param_translation): translate EKF2_GPS_POS_ to SENS_GPS0_OFF_
* fix(msgs): proper formatting
* chore(msg): 0 if invalid/unknown
* fix(ROMFS): migrate EKF2_GPS_POS_ params
* fix(docs): migrate EKF2_GPS_POS_ params
* fix(blending): unsigned param
* Update msg/SensorGps.msg
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* fix(sensors/gps): remove 'values:' tag in module.yaml
* fix(sensors/gps): unsigned instance index
* fix(blending): restore const on gps_blend_states()
Move antenna offset blending into blend_gps_data() where the
weights are computed, keeping gps_blend_states() const.
* fix(sensors/gps): fix msg annotation and restore SENS_GPS_PRIME values
Remove incorrect @invalid NaN annotation from antenna offset fields
(0.0 default is correct, not a sentinel). Restore values tag for
SENS_GPS_PRIME so QGC shows a dropdown.
* fix(gps_blending): fix pre-existing bug to clear _gps_updated flags
The loop iterates over i but always clears gps_select_index. The intent is to clear
all updated flags, but only the selected one gets cleared (N times)
* test(gps_blending): add stale update flag regression test
Overhauls the DShot driver with per-timer BDShot selection, multi-timer
sequential capture, Extended DShot Telemetry (EDT), and AM32 ESC EEPROM
read/write via MAVLink. Expands ESC support from 8 to 12 channels.
BDShot:
- Per-timer BDShot protocol selection via actuator config UI
- Multi-timer sequential burst/capture on any DMA-capable timer
- Adaptive per-channel GCR bitstream decoding
- Per-channel online/offline detection with hysteresis
Extended DShot Telemetry (EDT):
- Temperature, voltage, current from BDShot frames (no serial wire)
- New DSHOT_BIDIR_EDT parameter
- EDT data merged with serial telemetry when both available
AM32 EEPROM:
- Read/write AM32 ESC settings via MAVLink ESC_EEPROM message
- ESCSettingsInterface abstraction for future ESC firmware types
- New DSHOT_ESC_TYPE parameter
Other changes:
- Per-motor pole count params DSHOT_MOT_POL1–12 (replaces MOT_POLE_COUNT)
- EscStatus/EscReport expanded to 12 ESCs with uint16 bitmasks
- Numerous bounds-check, overflow, and concurrency fixes
- Updated DShot documentation
2026-03-17 16:38:33 -08:00
1260 changed files with 39174 additions and 33683 deletions
description: Create a conventional commit for PX4 changes
disable-model-invocation: true
argument-hint: "[optional: description of changes]"
allowed-tools: Bash, Read, Glob, Grep
---
# PX4 Conventional Commit
Create a git commit: `type(scope): description`
**NEVER add Co-Authored-By lines. No Claude attribution in commits.**
Follow [CONTRIBUTING.md](../../CONTRIBUTING.md) for full project conventions.
## Steps
1.**Read [CONTRIBUTING.md](../../CONTRIBUTING.md)** for commit message format, types, scopes, and conventions.
2. Check branch (`git branch --show-current`). If on `main`, create a feature branch. Use `<username>/<description>` format where `<username>` comes from `gh api user --jq .login`. If unavailable, just use `<description>`.
3. Run `git status` and `git diff --staged`. If nothing staged, ask what to stage.
4. Follow the commit message convention from CONTRIBUTING.md: pick the correct **type** and **scope**, write a concise imperative description.
5. Body (if needed): explain **why**, not what.
6. Run `make format` or `./Tools/astyle/fix_code_style.sh <file>` on changed C/C++ files before committing.
7. Check if GPG signing is available: `git config --get user.signingkey`. If set, use `git commit -S -s`. Otherwise, use `git commit -s`.
8. Stage and commit. No `Co-Authored-By`.
If the user provided arguments, use them as context: $ARGUMENTS
description: Create a pull request with conventional commit title and description
disable-model-invocation: true
argument-hint: "[optional: target branch or description]"
allowed-tools: Bash, Read, Glob, Grep
---
# PX4 Pull Request
**No Claude attribution anywhere (no Co-Authored-By, no "Generated with Claude").**
Follow [CONTRIBUTING.md](../../CONTRIBUTING.md) for full project conventions.
## Steps
1. Check branch. If on `main`, create a feature branch. Use `<username>/<description>` format where `<username>` comes from `gh api user --jq .login`. If unavailable, just use `<description>`.
3. PR **title**: `type(scope): description` — under 72 chars, describes the overall change across all commits. This becomes the squash-merge commit message.
4. PR **body**: brief summary + bullet points for key changes. No filler.
5. Push with `-u` if needed, then `gh pr create`. Default base is `main` unless user says otherwise.
6. Return the PR URL.
If the user provided arguments, use them as context: $ARGUMENTS
description: Rebase a branch onto main, handling squash-merged parent branches cleanly
argument-hint: "[optional: branch name, defaults to current branch]"
allowed-tools: Bash, Read, Glob, Grep, Agent
---
# Rebase Branch onto Main
Rebase the current (or specified) branch onto `main`, correctly handling the case where the branch was built on top of another branch that has since been squash-merged into `main`.
## Background
When a parent branch is squash-merged, its individual commits become a single new commit on `main` with a different hash. A normal `git rebase main` will try to replay the parent's original commits, causing messy conflicts. The fix is to **cherry-pick only the commits unique to this branch** onto a fresh branch from `main`.
## Steps
1.**Identify the branch.** Use `$ARGUMENTS` if provided, otherwise use the current branch.
2.**Fetch and update main:**
```
git fetch origin main:main
```
3. **Find the merge base** between the branch and `main`:
```
git merge-base <branch> main
```
4. **List all commits** on the branch since the merge base:
```
git log --oneline <merge-base>..<branch>
```
5. **Identify which commits are unique to this branch** vs. inherited from a parent branch. Look for:
- Squash-merged commits on `main` that correspond to a group of commits at the bottom of the branch's history (check PR titles, commit message keywords).
- The boundary commit: the first commit that belongs to *this* branch's work, not the parent's.
- If ALL commits are unique (no parent branch), just do a normal `git rebase main` and skip the rest.
6. **Create a fresh branch from `main`:**
```
git checkout -b <branch>-rebase main
```
7. **Cherry-pick only the unique commits** (oldest first):
```
git cherry-pick <first-unique-commit>^..<branch>
```
The `A^..B` range means "from the parent of A through B inclusive."
8. **Handle conflicts** if any arise during cherry-pick. Resolve and `git cherry-pick --continue`.
9. **Replace the old branch:**
```
git branch -m <branch> <branch>-old
git branch -m <branch>-rebase <branch>
```
10. **Verify** the result:
```
git log --oneline main..<branch>
```
Confirm only the expected commits are present.
11. **Ask the user** before force-pushing. When approved:
-`gh pr checks <PR>` (exit code 8 means some checks are pending, this is normal, not an error)
-`gh pr diff <PR>` -- if this fails with HTTP 406 (300+ files), do NOT retry. Instead use `gh api repos/OWNER/REPO/pulls/NUMBER/files --paginate` to get the full file list in one call, then fetch patches for key infrastructure files individually and sample representative changes from each domain touched.
-`gh api repos/OWNER/REPO/pulls/NUMBER/comments --paginate --jq '.[] | {user: .user.login, body: .body, path: .path, created_at: .created_at}'` to get inline review comments
-`gh api repos/OWNER/REPO/issues/NUMBER/comments --paginate --jq '.[] | {user: .user.login, body: .body, created_at: .created_at}'` to get PR conversation comments
From the PR metadata, note:
- **Assigned reviewers**: who has been requested to review (from `reviewRequests`)
- **Existing reviews**: who has already reviewed and their verdict (from `reviews` -- approved, changes_requested, commented, dismissed)
- **PR comments and inline comments**: read all existing feedback to avoid duplicating points already raised by other reviewers, and to build on their discussion rather than ignoring it
2.**Check CI status.** From the `gh pr checks` output in step 1, summarize pass/fail/pending. If there are failures, fetch logs with `gh run view <run-id> --log-failed`. Include CI status in the output.
3.**Recommend merge strategy.** Analyze the commit history and recommend squash or rebase merge. This decision informs all subsequent commit hygiene feedback.
**Recommend rebase merge** when:
- Commits are atomic, each builds/works independently
- Each commit has a proper `type(scope): description` message
- The PR intentionally separates logical changes (e.g., refactor + feature, or one commit per module)
- The commit history tells a useful story that would be lost by squashing
**Recommend squash merge** when:
- There are WIP, fixup, or review-response commits
- Commit messages are messy or inconsistent
- The PR is a single logical change spread across multiple commits
- There are "oops" or "make format" commits mixed in
Include the recommendation in the output. If recommending rebase, flag any commits that break atomicity or have bad messages. If recommending squash, don't bother flagging individual commit messages (they'll be discarded) but ensure the PR title is correct since it becomes the squash commit message.
4.**Check conventional commit title.** Verify the PR title follows `type(scope): description` per CONTRIBUTING.md. The PR title becomes the commit message on squash-merge, so it must be accurate and descriptive. Verify the scope matches the primary area of changed files. If the PR introduces breaking changes, the title must include `!` before the colon. If rebase merge was recommended in step 3, also scan individual commit messages for anti-patterns: vague messages ("fix", "update"), missing type prefix, review-response noise ("apply suggestions from code review", "do make format"), or WIP markers. Flag these for rewording.
5.**Identify domains touched.** Classify changed files into domains based on paths (a PR may touch multiple):
- **Type safety**: int16 overflow, float/double promotion, unsigned subtraction, use `uint64_t` for absolute time
- **Initialization**: uninitialized variables, missing default construction
- **Buffer safety**: unchecked array access, stack allocation of large buffers, snprintf bounds
- **Magic numbers**: every numeric literal needs a named constant or justification
- **Framework reuse**: use PX4_ERR/WARN/INFO, existing libraries (AlphaFilter, SlewRate, RateControl), MAVLink constants from the library
- **Naming**: accurate, no unjustified abbreviations, current terminology (GPS -> GNSS for new code)
- **Unnecessary complexity**: can code be removed instead of added? Is there a simpler pattern?
- **Test coverage**: new features should include unit or integration tests; bug fixes should include regression tests where practical. When automated testing is infeasible (hardware-specific), require a flight log link from https://logs.px4.io or bench test evidence.
- **PR hygiene**: focused scope, no unrelated formatting, no stale submodule changes. Commits should be atomic and independently revertable. Multiple WIP or review-response commits should be squashed. Clean, logical commits will be preserved individually on main via rebase merge. **Do NOT assume PRs are squash-merged. Both squash and rebase merge are enabled; merge commits are disabled.** Verify the PR targets `main` unless it is a backport or release-specific fix.
- **Formatting**: `make format` / `make check_format` (astyle) for C/C++ files; `clang-tidy` clean. Python files checked with `mypy` and `flake8`. PRs failing CI format or lint checks will not be merged.
- **Coding style**: C/C++ must follow the [PX4 coding style](https://docs.px4.io/main/en/contribute/code.html)
- **Necessity**: challenge every addition with "Why?" Is this actually needed or just copied? Can we change a default instead of adding runtime detection?
- **Root cause vs symptom**: is this fixing the real problem or masking it?
- **Ecosystem impact**: what does this change mean for QGC users, log analysis tools, and third-party integrations?
- **Sustainability**: who will maintain this? Does it create long-term burden?
- **Architecture fit**: does the code live in the module that naturally owns the data? Are there unnecessary cross-module dependencies?
- **End user impact**: will parameters confuse less-technical users? Are error messages actionable in QGC?
7.**Apply domain checks** based on step 5:
**Estimation:**
- Singularities in aerospace math (euler angles near gimbal lock, sideslip at low airspeed)
- Aliasing from downsampling sensor data without filtering
- Ubuntu LTS support policy (latest + one prior only)
- Build time impact
- CMake preferred over Makefiles
**Messages/Protocol:**
- Backwards compatibility: will this break QGC, post-flight tools, or uLog parsers?
- uORB: `timestamp` for publication metadata, `timestamp_sample` close to physical sample, include `device_id`
- Don't version messages unless strictly needed
- Parameter UX: will this confuse users in a GCS? Every new param is a configuration burden
- MAVLink: use library constants, don't implement custom stream rates
**Board Addition:**
- **Flight logs**: require a link to https://logs.px4.io demonstrating basic operation for the vehicle type (hover for multicopters, stable flight for fixed-wing, driving for rovers, etc.); short bench-only logs are insufficient
- **Documentation**: require a docs page in `docs/en/flight_controller/` with pinout, where-to-buy, connector types, version badge, and manufacturer-supported notice block
- **USB VID/PID**: must not reuse another manufacturer's Vendor ID; manufacturer must use their own
- **Board naming**: directory is `boards/{manufacturer}/{board}/`, both lowercase, hyphens for board name
- **Unique board_id**: registered in `boards/boards.json`, no collisions
- **Copied code cleanup**: check for leftover files, configs, or comments from the template board; "Is this real or leftover?"
- **RC configuration**: prefer `CONFIG_DRIVERS_COMMON_RC` over legacy `CONFIG_DRIVERS_RC_INPUT`
- **No board-specific custom modules**: reject copy-pasted drivers (e.g., custom heater) when existing infrastructure works
- **Bootloader**: expect a bootloader defconfig (`nuttx-config/bootloader/defconfig`) or explanation of shared bootloader
- **CI integration**: board must be added to CI compile workflows so it builds on every PR
- **Flash constraints**: verify enabled modules fit in flash; we are running low across all board targets
- **Port labels**: serial port labels must match what is physically printed on the board
- **Hardware availability**: for unknown manufacturers, verify the product exists and is purchasable (no vaporware)
8.**Format output** as:
- **CI status**: pass/fail summary, link to failed runs if any
- **Merge strategy**: recommend squash or rebase merge with reasoning
- **Title check**: pass/fail with suggestion
- **Review status**: list assigned reviewers and any existing reviews (who approved, who requested changes, key points already raised). Note if your review would duplicate feedback already given.
- **Domains detected**: list which domain checks were applied
- **Summary**: one paragraph on what the PR does and whether the approach is sound
- **Issues**: numbered list, each with file:line, severity (blocker/warning/nit), and explanation. Skip issues already raised by other reviewers unless you have something to add.
- **Verdict**: approve, request changes, or needs discussion
After the structured output, also display a **draft PR comment** formatted using the PR comment formatting rules from step 9. This gives the user a preview of what would be posted.
9.**Interactive dialog.** After displaying the review, present the user with these options:
Present options based on the verdict:
If verdict is **approve**:
```
What would you like to do?
1. Chat about this PR (ask questions, explore code) [default]
2. Approve this PR and post the review comment
3. Adjust the review or draft (tell me what to change)
4. Done for now
```
If verdict is **request changes**:
```
What would you like to do?
1. Chat about this PR (ask questions, explore code) [default]
2. Request changes on this PR and post the review comment
3. Adjust the review or draft (tell me what to change)
4. Done for now
```
If verdict is **needs discussion**:
```
What would you like to do?
1. Chat about this PR (ask questions, explore code) [default]
2. Post the review as a comment (no approval or rejection)
3. Adjust the review or draft (tell me what to change)
4. Done for now
```
Wait for the user to choose before proceeding. If they pick:
- **1 (chat)**: enter a free-form conversation about the PR. The user can ask about specific files, code paths, or decisions. When done, loop back to the options. This is the default if the user just presses enter.
- **2 (submit)**: use the draft PR comment already shown. Before posting, check if you have review permissions: run `gh api repos/OWNER/REPO/collaborators/$(gh api user --jq .login)/permission --jq .permission` -- if `admin` or `write`, submit as a formal review with `gh pr review <PR> --approve --body "..."` or `gh pr review <PR> --request-changes --body "..."` based on the verdict. If no write access, fall back to `gh pr comment <PR> --body "..."`. Always confirm with the user before posting.
- **3 (adjust)**: ask what to change, update the review and draft, then loop back to the options.
- **4 (done)**: stop.
**PR comment formatting rules** (for the draft):
When writing the GitHub comment, rewrite the review to sound like a human reviewer, not a structured report. Do NOT include the full skill output. Instead:
- Drop most meta-sections (CI status, title check, domains detected, severity labels) but keep the merge strategy recommendation (e.g., "I'd suggest a rebase merge here since the commits are clean and atomic" or "This should be squash-merged, the commit history is messy")
- Write conversationally: "Nice work on this. A few things I noticed:" not "Issues: 1. file:line (warning):"
- Lead with a brief take on the overall change (1-2 sentences)
- List only actionable feedback as natural review comments, not numbered checklists
- Skip nits unless they are particularly useful
- End with a clear stance: looks good to merge, needs a few changes, or let's discuss X
- Post with `gh pr comment <PR> --body "$(cat <<'EOF' ... EOF)"`. Do not post without explicit confirmation.
If the user provided arguments, use them as context: $ARGUMENTS
In addition to the core code review guidelines, when reviewing new board additions:
- **Flight logs**: require a link to https://logs.px4.io demonstrating basic operation for the vehicle type (hover for multicopters, stable flight for fixed-wing, driving for rovers, etc.); short bench-only logs are insufficient
- **Documentation**: require a docs page in `docs/en/flight_controller/` with pinout, where-to-buy, connector types, version badge, and manufacturer-supported notice block
- **USB VID/PID**: must not reuse another manufacturer's Vendor ID; manufacturer must use their own
- **Board naming**: directory is `boards/{manufacturer}/{board}/`, both lowercase, hyphens for board name
- **Unique board_id**: registered in `boards/boards.json`, no collisions
- **Copied code cleanup**: check for leftover files, configs, or comments from the template board. Ask "Is this real or leftover?"
- **RC configuration**: prefer `CONFIG_DRIVERS_COMMON_RC` over legacy `CONFIG_DRIVERS_RC_INPUT`
- **No board-specific custom modules**: reject copy-pasted drivers (e.g., custom heater) when existing infrastructure works
- **Bootloader**: expect a bootloader defconfig (`nuttx-config/bootloader/defconfig`) or explanation of shared bootloader
- **CI integration**: board must be added to CI compile workflows so it builds on every PR
- **Flash constraints**: verify enabled modules fit in flash; we are running low across all board targets
- **Port labels**: serial port labels must match what is physically printed on the board
- **Hardware availability**: for unknown manufacturers, verify the product exists and is purchasable (no vaporware)
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.