px4_fmu-v5 ostest priority_inheritance hard fault

[dagar]

On px4_fmu-v5_test (px4_fmu-v5_default + test modules, nuttx ostest, etc) there's a hard fault when running the NuttX ostest.

Examples

• http://px4-jenkins.dagar.ca:8080/blue/organizations/jenkins/PX4-Autopilot/detail/PR-20880/3/pipeline

[dagar]

The ostest is failing intermittently on px4_fmu-v5_test (maybe 10-20% of the runs), but seems to be solid on the other architectures (stm32f4, stm32h7, kinetis).

[dagar]

http://px4-jenkins.dagar.ca:8080/blue/organizations/jenkins/PX4-Autopilot/detail/pr-boards_fmu-v6x_ext2_fix/1/pipeline/530 (cuav_v5_plus)

[dagar]

http://px4-jenkins.dagar.ca:8080/blue/organizations/jenkins/PX4-Autopilot/detail/pr-simulation_airframes_contd/7/pipeline/454 (cuav_v5_plus)

[davids5]

@dagar i am Oot till 17th.

[ThomasDebrunner]

Can't reproduce it on my fmu_v5 with (building px4_fmu-v5_test from main, running ostest). Will test some more

[dagar]

Can't reproduce it on my fmu_v5 with (building px4_fmu-v5_test from main, running ostest). Will test some more

Which FMUv5 variant and is the airframe configured? Might just be a coincidence, but so far all of the ostest failures are happening on either the CUAV v5plus (mostly) and Holybro Pixhawk 4. There are 4 variants of the FMUv5 on the hardware test rack (Holybro Pixhawk 4, 4 mini with no px4io, CUAV v5 plus, CUAV v5 nano with no px4io) and Jenkins will schedule any of them to run these jobs.

[dagar]

Anecdata, I don't think I've seen this failure on the test rack since bumping the interrupt stack last week. https://github.com/PX4/PX4-Autopilot/commit/dc5ce9b0cecc1a165af0b639a903ec1d694e0e80