[bmp388] Try to initialize multiple times

[niklaut]

Solved Problem

The BMP388 sometimes boots slowly on first power-on, so we should try to ping it multiple times.

The issue happens only on first power-on, if you then software reboot, the issue goes away.

Solution

Ping it more often?
The init function is only called once, which is a bit… risky. Unfortunately almost all drivers are implemented like this.

Changelog Entry

For release notes:

Bugfix BMP388: initialize multiple times

Alternatives

Have the workqueue continuously initialize the sensor, this would also guard against communication issues mid-flight. But that requires a rewrite of the driver.

Test coverage

• Tested in hardware

[dakejahl]

Power-on delay is a common problem across all the drivers. The issue here is that init() is being called from the context of the startup script, which could be really early during boot depending on the order of drivers. In the future maybe we consider implementing a startup queue... or rewriting all of the drivers to init() after they've been given the proper ScheduleDelayed(power_on_delay)

[github-actions]

🔎 FLASH Analysis

px4_fmu-v5x [Total VM Diff: 40 byte (0 %)]

    FILE SIZE        VM SIZE    
--------------  -------------- 
+0.0%     +40  +0.0%     +40    .text
  +127%     +28  +127%     +28    BMP388_I2C::init()
   +17%      +8   +17%      +8    BMP388_I2C::BMP388_I2C()
   +33%      +8   +33%      +8    IIS2MDC_I2C::IIS2MDC_I2C()
  -5.9%      -4  -5.9%      -4    CameraCapture::stop()
+0.0%     +70  [ = ]       0    .debug_abbrev
+0.0%     +12  [ = ]       0    .debug_frame
+0.0%     +71  [ = ]       0    .debug_info
+0.0%     +75  [ = ]       0    .debug_line
 -40.0%      -2  [ = ]       0    [Unmapped]
  +0.0%     +77  [ = ]       0    [section .debug_line]
+0.0%     +38  [ = ]       0    .debug_loclists
+0.0%     +10  [ = ]       0    .debug_rnglists
  [DEL]      -2  [ = ]       0    [Unmapped]
  +0.0%     +12  [ = ]       0    [section .debug_rnglists]
-0.4%     -40  [ = ]       0    [Unmapped]
+0.0%    +276  +0.0%     +40    TOTAL

px4_fmu-v6x [Total VM Diff: 40 byte (0 %)]

    FILE SIZE        VM SIZE    
--------------  -------------- 
+0.0%     +40  +0.0%     +40    .text
  +127%     +28  +127%     +28    BMP388_I2C::init()
   +17%      +8   +17%      +8    BMP388_I2C::BMP388_I2C()
   +33%      +8   +33%      +8    IIS2MDC_I2C::IIS2MDC_I2C()
  -5.9%      -4  -5.9%      -4    CameraCapture::stop()
+0.0%     +70  [ = ]       0    .debug_abbrev
+0.0%     +12  [ = ]       0    .debug_frame
+0.0%     +71  [ = ]       0    .debug_info
+0.0%     +75  [ = ]       0    .debug_line
 -66.7%      -2  [ = ]       0    [Unmapped]
  +0.0%     +77  [ = ]       0    [section .debug_line]
+0.0%     +38  [ = ]       0    .debug_loclists
+0.0%     +14  [ = ]       0    .debug_rnglists
  +200%      +2  [ = ]       0    [Unmapped]
  +0.0%     +12  [ = ]       0    [section .debug_rnglists]
-0.8%     -40  [ = ]       0    [Unmapped]
+0.0%    +280  +0.0%     +40    TOTAL

Updated: 2025-06-02T14:13:51

[niklaut]

I've tested this on real hardware with wonky wiring and this fixes issues with the on-board BMP390 sensor on the auterion/FMUv6s.