[stm32] PM restore console after sleep mode

[FRASTM]

When testing the PowerManagement on stm32L4 or stm32WB55 through the tests/subsys/power/power_mgmt
the console is no more functional and prints spurious characters.

Describe the solution you'd like
To define a function in the uart_stm32.c driver (replace device_pm_control_nop)
to flush the console before going to sleep
To restore the correct clock for the console after sleep mode

**Console: **

Running test suite power_management_test                                        
===================================================================             
START - test_power_idle                                                         
give way to idle thread                                                         
 PASS - test_power_idle                                                         
===================================================================             
START - test_power_state_trans                                                  
�������������������������������������������������������������������������������

[FRASTM]

When considering STOP 0/1/2 low power modes as DEVICE_PM_LOW_POWER_STATE for the stm32l4xx soc:

1. before going to sleep mode, wait for USART Rx and Tx empty
2. USART is clocked by the PLL. During STOP mode the all clocks are OFF except LSE and LSI. After STOP mode HSI16 or MSI are wakeup clock, so the USART needs one of them or system must restore the PLL
3. in the test application tests/subsys/power/power_mgmt, the dummy_driver is put in DEVICE_PM_SUSPEND_STATE
   not in DEVICE_PM_LOW_POWER_STATE
4. LPUART is not available on this nucleo board

[FRASTM]

In the test application tests/subsys/power/power_mgmt, the dummy_driver is put in DEVICE_PM_SUSPEND_STATE (=3)
Going to sleep mode the dummy driver is still in DEVICE_PM_SUSPEND_STATE (=3) when the irq (lptim) occurs and cause the system to wakeUp.
On wakeUp, the pm_notify_power_state_exit() is called in the pm_system_resume(), "from the ISR of the event that caused exit from kernel idling after PM operations" before the pm_resume_devices() function.
The pm_notify_power_state_exit, implemented in the main.c, is checking the device_power_state of the dummy driver.
At this moment the device_power_state of the dummy driver is still DEVICE_PM_SUSPEND_STATE.
--> zassert_equal(device_power_state, DEVICE_PM_ACTIVE_STATE, NULL); fails
After that, the pm_resume_devices(), especially for the dummy_driver is setting back the device_power_state from DEVICE_PM_SUSPEND_STATE to DEVICE_PM_ACTIVE_STATE.

[FRASTM]

add the CONFIG_PM_DEVICE definition from the CONFIG_PM for the soc that supports the low power modes.

[FRASTM]

Note that the tests/subsys/power/power_mgmt_soc might requires a stack size with CONFIG_IDLE_STACK_SIZE=640 in its prj.conf