tests: kernel: timer_api fails with hard fault in CONFIG_TICKLESS_KERNEL

[FRASTM]

Describe the bug
run the tests/kernel/timer/timer_api with the prj_tickless.conf on the nucleo_wb55rg
This board and soc device have PM with their own ./soc/arm/st_stm32/stm32wb/power.c

To Reproduce
Steps to reproduce the behavior:
set prj.conf with

CONFIG_ZTEST=y
CONFIG_PM=y
CONFIG_TICKLESS_KERNEL=y
CONFIG_MP_NUM_CPUS=1
CONFIG_TEST_RANDOM_GENERATOR=y

CONFIG_TEST_USERSPACE=y

1. west build -p auto -b nucleo_wb55rg tests/kernel/timer/timer_api/
   or
   ./scripts/twister -p nucleo_wb55rg -T tests/kernel/timer/timer_api

Impact

Running test suite timer_api                                                    
===================================================================             
START - test_time_conversions                                                   
 PASS - test_time_conversions in 0.113 seconds                                  
===================================================================             
START - test_timer_duration_period                                              
ASSERTION FAIL [!arch_is_in_isr()] @ WEST_TOPDIR/zephyr/kernel/sched.c:841      
                                                                                
E: ***** HARD FAULT *****                                                       
E:   Fault escalation (see below)                                               
E: r0/a1:  0x00000004  r1/a2:  0x00000349  r2/a3:  0x00000040                   
E: r3/a4:  0x0000003f r12/ip:  0xa0000000 r14/lr:  0x08007637                   
E:  xpsr:  0x4100003f                                                           
E: Faulting instruction address (r15/pc): 0x0800d676                            
E: >>> ZEPHYR FATAL ERROR 0: CPU exception on CPU 0                             
E: Fault during interrupt handling                                              
                                                                                
E: Current thread: 0x200004b0 (idle 00)                                         
E: Halting system 

Environment (please complete the following information):*
master : SHA1 = 7397e3e

Additional context
Also failing test : tests/kernel/tickless/tickless_concept/

Running test suite tickless_concept
===================================================================
START - test_tickless_sysclock
ASSERTION FAIL [!arch_is_in_isr()] @ WEST_TOPDIR/zephyr/kernel/sched.c:841
...

[FRASTM]

There is a PR #33222, where the commit bfc11b5 "power: Send exit notification after post_ops callback" is causing the kernel crash.

[FRASTM]

I propose to unlock the scheduler NOT in the pm_system_resume() itself but in the pm_system_suspend() just after the pm_system_resume() function call, as follows:

	pm_log_debug_info(z_power_state.state);
	pm_system_resume();
	k_sched_unlock();

	return z_power_state.state;
}

[FRASTM]

The tests/subsys/power/power_mgmt_soc also fails on nucleo_wb55 and nucleo_l552 boards

[FRASTM]

closing as the PR #33666 is now merged