tests/drivers/watchdog/wdt_basic_api/testcase.yaml: test_wdt_no_callback() failed at "Waiting to restart MCU"

[cinlyooi-intel]

Describe the bug
After restart, test_wdt_no_callback() output stop at "Waiting to restart MCU"

To Reproduce
Steps to reproduce the behavior:

1. mkdir build; cd build
2. cmake -DBOARD=arduino2 ..
3. make
4. make run
   for sam_e70: Replace (2) with "cmake -DBOARD=sam_e70_xplained -DARCH=arm"

Screenshots or console output
For both arduino2 and same_e70_xplained:

***** delaying boot 1000ms (per build configuration) *****
***** Booting Zephyr OS v1.14.0-rc1-211-g95e8a582b4 (delayed boot 1000ms) *****
Running test suite wdt_basic_test
===================================================================
starting test - test_wdt
Testcase: test_wdt_no_callback
Watchdog setup error
Waiting to restart MCU

Environment (please complete the following information):

• OS: Fedora29
• Toolchain: zephyr
• Commit SHA or Version used: 9072d34 (andruino2)
  6aa8d94 (sam_370_xplained)

Additional context
Add any other context about the problem here.

[cinlyooi-intel]

Repeated in Commit ID fd7c443 for sam_e70_xplained, arduino2

***** delaying boot 1000ms (per build configuration) *****
***** Booting Zephyr OS v1.14.0-rc1-704-gfd7c4430e139 (delayed boot 1000ms) *****
Running test suite wdt_basic_test
===================================================================
starting test - test_wdt
Testcase: test_wdt_no_callback
Watchdog setup error
Waiting to restart MCU

[aurel32]

The problem is that on the SAM WDT can only be enabled or disabled once (the mode register is write once). By default the CONFIG_WDT_SAM_DISABLE_AT_BOOT option is enabled and disables the watchdog. Then it's not possible to enable it again.

After disabling this option, the test passes. I wonder how this option is different from the generic CONFIG_WDT_DISABLE_AT_BOOT option which is NOT enabled by default.

[aurel32]

After disabling this option, the test passes.

Well I am wrong there. The MCU reboots correctly when the WDT triggers, however it also reset the RAM, while the test assumes that the RAM is left untouched by the WDT reset.

[overheat]

Fixed through #14968

[ioannisg]

After disabling this option, the test passes.

Well I am wrong there. The MCU reboots correctly when the WDT triggers, however it also reset the RAM, while the test assumes that the RAM is left untouched by the WDT reset.

So this test, as is, cannot pass on the SAM WDT @aurel32 and we should just exclude the platform from the test for now, until we extend the test to cover the SAM watchdog.

@nashif any thoughts against?