EFR32 Unit Test crashes

[rgoliver]

Problem

EFR32 Unit tests crashes while running:
python -m nl_test_runner.nl_test_runner -d /dev/ttyACM0 -f out/debug/chip-efr32-device_tests.s37 -o out.log

...
<detail> [IN] New secure session created for device 0x0000000006A11E3D, key 2!!
<detail> [IN] New secure session created for device 0x000000000001E306, key 1!!
<info  > [IN] Prepared encrypted message 0x20020518 to 0x0000000006A11E3D of type 0xhx and protocolId (1, 0) on exchange 4� with
 MessageCounter:105.
<info  > [IN] Sending encrypted msg 0x20020518 with MessageCounter:1 to 0x0000000006A11E3D at monotonic time: ld msec
<info  > [-] 
FREERTOS ASSERT ( (pxLink->xBlockSize & xBlockAllocatedBit) != 0 )

<info  > [-] 
FREERTOS ASSERT ( ( portNVIC_INT_CTRL_REG & portVECTACTIVE_MASK ) == 0 )

[rgoliver]

After doing a bisect I think this it was broken by this CL: d48d754

[rgoliver]

App tests have a harfault during the TestReadInteraction:
HardFault:

SCB->CFSR 0x00000400
SCB->HFSR 0x40000000
SCB->MMFAR 0xe000ed34
SCB->BFAR 0xe000ed38
SP 0x20029190
R0 0x00000f29

R1 0x30bd08fc

R2 0x01200000

R3 0x4fb50847

R12 0xa5a5a5a5

LR 0x000105ad

PC 0x0000f88e

PSR 0x01000000

[rgoliver]

src/controller/tests/TestDevice.cpp test has hardfault:
HardFault:

SCB->CFSR 0x00000400
SCB->HFSR 0x40000000
SCB->MMFAR 0xe000ed34
SCB->BFAR 0xe000ed38
SP 0x20028c28
R0 0x00001fe5

R1 0x002b1080

R2 0x1180b0f0

R3 0x002b1280

R12 0xa5a5a5a5

LR 0x000126b1

PC 0x00011994

PSR 0x01000000

[rgoliver]

Test-CHIP-ReliableMessageProtocol test has hardfault:

HardFault:

SCB->CFSR 0x00008200
SCB->HFSR 0x40000000
SCB->MMFAR 0xfa0f02f4
SCB->BFAR 0xfa0f02f4
SP 0x20028fa0
R0 0x00001d7d

R1 0x046803b9

R2 0x0360da22

R3 0xfa0f02f0

R12 0xa5a5a5a5

LR 0x000105ad

PC 0x0000f890

PSR 0x01000000

[andy31415]

Still happening, @rgoliver asked to reopen

[rgoliver]

#14238 disables TestExchangeMgr which hands on EFR32.

[bzbarsky-apple]

@rgoliver Is this still an issue?

[rgoliver]

@jmartinez-silabs I know you were looking at fixing and re-enabling the broken EFR unit tests, can this be closed?

[jmartinez-silabs]

If fixed two test group if i recall. The others i could not even try to run them or debug because enabling them overdflows the flash space of our devices

[rgoliver]

That is an issue with running the tests on the device, one option would be to group the tests and build multiple test binaries instead of a single combined one.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs.

[stale]

This stale issue has been automatically closed. Thank you for your contributions.

[andy31415]

Likely still an issue

[jmartinez-silabs]

I just tested the tests. They all link but there is an issue with the event loop. The queue gets full and it causes other issues in the exchanges manager from what I can see. I also opened #33373

I should have a few cycles to investigate further.