AP_Compass: remove IIS2MDC driver and migrate boards to LIS2MDL#33121
Open
dakejahl wants to merge 3 commits into
Open
AP_Compass: remove IIS2MDC driver and migrate boards to LIS2MDL#33121dakejahl wants to merge 3 commits into
dakejahl wants to merge 3 commits into
Conversation
LIS2MDL and IIS2MDC are the same die and share WHO_AM_I (0x40 at reg 0x4F), the same register layout, the same axis convention, and the same sensitivity. ST sells them as separate part numbers (consumer vs industrial grade) but they are electrically and software-identical, so having two ArduPilot drivers for them is redundant. The LIS2MDL driver is default-enabled on boards with >1 MB flash, and its internal-bus auto-probe lacked the `if (all_external)` gate used by HMC5843, QMC5883L, QMC5883P, and IIS2MDC. The LIS2MDL auto-probe therefore claimed every IIS2MDC chip first with ROTATION_NONE hard-coded, and any explicit `COMPASS IIS2MDC ...` line was then blocked by `_i2c_sensor_is_registered()`. The rotation declared in the board hwdef was silently dropped. Delete the AP_Compass_IIS2MDC driver, the AP_COMPASS_IIS2MDC_ENABLED config, DRIVER_IIS2MDC, DEVTYPE_IIS2MDC, the IIS2MDC probe block in `_probe_external_i2c_compasses()`, and the corresponding entries in Tools/scripts/build_options.py and Tools/scripts/decode_devid.py. Gate the LIS2MDL internal-bus auto-probe with `if (all_external)` so it stops shadowing explicit COMPASS lines on internal buses. Stored COMPASS_DEV_ID values with DEVTYPE_IIS2MDC (0x18) will be replaced by DEVTYPE_LIS2MDL (0x19) on next boot. In practice no real unit should have 0x18 stored - the LIS2MDL auto-probe was already shadowing every IIS2MDC chip - so anyone with a working calibration on these boards already has 0x19.
The IIS2MDC driver has been removed; LIS2MDL handles the identical silicon. Switch the COMPASS line on each affected board to LIS2MDL, preserving the existing rotation: ARK_FPV ROTATION_NONE ARK_PI6X ROTATION_YAW_180 CUAV-7-Nano ROTATION_ROLL_90_YAW_270 Morakot ROTATION_NONE PrincipIoTH7Pi ROTATION_YAW_270 navigator ROTATION_ROLL_180_YAW_270 Also drop the no-longer-valid `define AP_COMPASS_IIS2MDC_ENABLED 1` lines from each hwdef.
Some boards that referenced the IIS2MDC driver did not enable `AP_COMPASS_PROBING_ENABLED` (e.g. PrincipIoTH7Pi), so only the explicit COMPASS line ran on those units and `COMPASS*_DEV_ID` was stored with the old DEVTYPE_IIS2MDC (0x18) byte. Without this migration, removing the driver would force every such unit to re-calibrate the magnetometer, which is painful for users managing fleets. At Compass::init(), rewrite the devtype byte of every stored `_state[i].dev_id` and `_priority_did_stored_list[i]` whose devtype is 0x18 (the now-removed DEVTYPE_IIS2MDC) to 0x19 (DEVTYPE_LIS2MDL), preserving the bus type, bus number, and I2C address. Existing calibration data, which is keyed off the saved dev IDs, is then automatically associated with the LIS2MDL driver after upgrade and the user is not forced to re-calibrate. The literal 0x18 is used rather than a named constant because the DEVTYPE_IIS2MDC enum value was removed in the prior commit.
Contributor
Author
|
@andyp1per this one is for you |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Delete the redundant
AP_Compass_IIS2MDCdriver and switch every board that referenced it to theLIS2MDLdriver, which handles the identical silicon. Also gate the LIS2MDL internal-bus auto-probe withif (all_external)so it stops shadowing explicitCOMPASSlines on internal buses.Problem
LIS2MDL and IIS2MDC are the same die — same WHO_AM_I (0x40 at reg 0x4F), same register layout, same axis convention, same sensitivity. ST sells them under separate part numbers (consumer vs industrial grade) but they're electrically and software-identical.
LIS2MDL is default-enabled on boards with >1 MB flash and its internal-bus auto-probe in
AP_Compass.cpplacks theif (all_external)gate that HMC5843/QMC5883L/QMC5883P/IIS2MDC use. So on every board withAP_COMPASS_PROBING_ENABLEDand an internal IIS2MDC at I2C 0x1E, the LIS2MDL auto-probe registered the chip first with hard-codedROTATION_NONE, and the explicitCOMPASS IIS2MDC ... ROTATION_XYZline was blocked by_i2c_sensor_is_registered(). The rotation in the hwdef was silently dropped.On boards that already used
ROTATION_NONE(ARK_FPV, Morakot, navigator) this was invisible. On boards that needed a real rotation (ARK_PI6XYAW_180, CUAV-7-NanoROLL_90_YAW_270, PrincipIoTH7PiYAW_270) mag calibration could not converge because the IMU and the compass were misaligned in yaw. Discovered during ARK FMU V6S board bring-up.Solution
Remove all IIS2MDC driver code, config, enum values, and tooling entries. Gate the LIS2MDL internal-bus auto-probe with
if (all_external). Switch the six affected boards fromCOMPASS IIS2MDC ...toCOMPASS LIS2MDL ..., preserving each board's existing rotation: ARK_FPV, ARK_PI6X, CUAV-7-Nano, Morakot, PrincipIoTH7Pi, navigator (Linux).On boards that enabled
AP_COMPASS_PROBING_ENABLED, the LIS2MDL auto-probe was already shadowing the IIS2MDC, so storedCOMPASS*_DEV_IDvalues already carryDEVTYPE_LIS2MDL(0x19). On boards that disabled probing (e.g. PrincipIoTH7Pi), only the explicitCOMPASS IIS2MDC ...line ran and 0x18 was actually stored. APARAMETER_CONVERSIONblock inCompass::init()rewrites any stored 0x18 to 0x19, preserving bus type/number/address so existing mag calibrations survive the upgrade without forcing users to re-calibrate.