stm32h7: systematic crash at each second boot with NETWORKING=y

[erwango]

Describe the bug On STM32H7 platform, a systematic error happens at each second boot. git bisect leads to following PR: https://github.com/zephyrproject-rtos/zephyr/pull/34086

To Reproduce Use any sample that uses NETWORKING=y on any h7 based board. For instance: west build -b nucleo_h745zi_q_m7 samples/posix/eventfd/

Expected behavior No crash at boot

Impact Blocker

Logs and console output

(gdb) c
Continuing.
^C
Program received signal SIGINT, Interrupt.
arch_system_halt (reason=reason@entry=0) at /local/mcu/zephyrproject/zephyr/kernel/fatal.c:32
32      }
(gdb) bt
#0  arch_system_halt (reason=reason@entry=0) at /local/mcu/zephyrproject/zephyr/kernel/fatal.c:32
#1  0x08012204 in k_sys_fatal_error_handler (reason=reason@entry=0, esf=esf@entry=0x24003a80 <z_interrupt_stacks+2016>) at /local/mcu/zephyrproject/zephyr/kernel/fatal.c:44
#2  0x08012228 in z_fatal_error (reason=0, esf=esf@entry=0x24003a80 <z_interrupt_stacks+2016>) at /local/mcu/zephyrproject/zephyr/kernel/fatal.c:127
#3  0x0800e842 in z_arm_fatal_error (reason=<optimized out>, esf=esf@entry=0x24003a80 <z_interrupt_stacks+2016>) at /local/mcu/zephyrproject/zephyr/arch/arm/core/aarch32/fatal.c:62
#4  0x080043c2 in z_arm_fault (msp=<optimized out>, psp=<optimized out>, exc_return=<optimized out>, callee_regs=<optimized out>)
    at /local/mcu/zephyrproject/zephyr/arch/arm/core/aarch32/cortex_m/fault.c:1034
#5  0x0800423c in z_arm_usage_fault () at /local/mcu/zephyrproject/zephyr/arch/arm/core/aarch32/cortex_m/fault_s.S:102

[erwango]

@ioannisg, @mmahadevan108, would you mind having a look ?

[mnkp]

This seems to be related to #35107. STM32H7 platform also requires CONFIG_NOCACHE_MEMORY=y for the Ethernet driver.

[galak]

Fixed by #35258