Closed JohanBertrand closed 1 month ago
I am glad you remembered this!
@LeStarch Is there anything else needed in this PR? The CI did not pass, but I don't think it's because of the current changes
The RPI tests seem to be failing consistently though... 🤔
Interesting. Is there any link between this bash script and the RPI integration tests?
Otherwise, are there any artefacts from previous jobs that could have been impacted?
It used to pass without any problem in a previous PR: https://github.com/nasa/fprime/pull/2709/commits/c98a07056ed0cd497089c5cf341a4c3ebce21b61
It seems like the file descriptor stays at -1 during the test. I don't expect it to come from this bash script, and I don't think that's expected in the integration test?
[GDS] Received EVR: 2024-05-28 13:18:20: fatalAdapter.AF_ASSERT_0 (16896) (2(0)-1716927500:896756) EventSeverity.FATAL : Assert in file /home/runner/work/fprime/fprime/fprime/Drv/LinuxGpioDriver/LinuxGpioDriverCompon, line 276
From Drv/LinuxGpioDriver/LinuxGpioDriverComponentImpl.cpp
:
void LinuxGpioDriverComponentImpl ::
gpioWrite_handler(
const NATIVE_INT_TYPE portNum,
const Fw::Logic& state
)
{
FW_ASSERT(this->m_fd != -1);
NATIVE_INT_TYPE stat;
stat = gpio_set_value(this->m_fd,(state == Fw::Logic::HIGH) ? 1 : 0);
if (0 != stat) {
this->log_WARNING_HI_GP_WriteError(this->m_gpio,stat);
return;
}
}
This error means the RPIs need to be rebooted.....some orphaned process is holding onto the GPIO pins. They are supposed to reboot after each test, but this seems to be breaking.
Regardless this PR can be merged.
Change Description
Small change of the bash script to display errors happening on the CI. The bash version used on the Mac OS CI does not support negative values (See https://github.com/nasa/fprime/actions/runs/8902005968/job/24448082623 for the error, and https://github.com/nasa/fprime/actions/runs/8903107367/job/24450396620 for the fix)