tests/drivers/uart/uart_async_api/drivers.uart.async_api fails at nrf54h20

[nordic-piks]

• Describe the bug
  The tests:
• tests/drivers/uart/uart_async_api/drivers.uart.async_api

Observed for

• [email protected]/nrf54h20/cpuapp
• [email protected]/nrf54h20/cpurad

To Reproduce
Steps to reproduce the behavior:

1. have [email protected]/nrf54h20/cpuapp connected
2. go to your zephyr dir
3. call ./scripts/twister -T tests/drivers/uart/uart_async_api -p [email protected]/nrf54h20/cpuapp --device-testing --device-serial /dev/ttyACM0 -v --inline-logs
4. See console output with error

Expected behavior
Valid console output

Impact
Not clear

Logs and console output

*** Booting Zephyr OS build v4.1.0-rc2-57-g05fb46cb9177 ***
W: Init not done!
Running TESTSUITE uart_async_chain_read
===================================================================
UART instance:uart@9d6000
START - test_chained_read
E: ***** MPU FAULT *****
E:   Data Access Violation
E:   MMFAR Address: 0x2f0146c0
E: r0/a1:  0x0000004d  r1/a2:  0x2f01713c  r2/a3:  0x2f0146c1
E: r3/a4:  0x2f0146c0 r12/ip:  0x0e0a7c41 r14/lr:  0x0e0a7727
E:  xpsr:  0x81001800
E: Faulting instruction address (r15/pc): 0x0e0aa362
E: >>> ZEPHYR FATAL ERROR 0: CPU exception on CPU 0
E: Current thread: 0x2f0148a0 (test_chained_read)
E: Halting system

Environment (please complete the following information):

• OS: Ubuntu 20.04.1 LTS
• Toolchain Zephyr SDK 0.17.0
• Commit SHA or Version used: v4.1.0-rc2-57-g05fb46cb9177

[nordic-piks]

Guessing that may be related to #85548
@sswetha18

[sswetha18]

Could you please try by disabling Userspace
CONFIG_TEST_USERSPACE = n

[nordic-krch]

@sswetha18 you made all those buffers not only aligned but also static and because of that test is failing when userspace is enabled. Can you remove static from those variables? Test was supporting CONFIG_TEST_USERSPACE=y and it should stay that way.

[sswetha18]

sure will remove

[djiatsaf-st]

Hello @sswetha18, I have also encountered the same issue on many STM32 platforms.

The suggestion made by @nordic-krch should normally correct the regression.

[sswetha18]

#86445