Re: [PATCH 1/4] firmware: psci: switch SYSTEM_OFF to sys-off handler API

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Premature memory allocation during early boot causing kernel panic
--

commit ef783d7ffab1683b74c28865d23fcb00064437a6
Author: Diogo Ivo <diogo.ivo@tecnico.ulisboa.pt>

firmware: psci: switch SYSTEM_OFF to sys-off handler API

This commit replaces the legacy pm_power_off hook with the generic sys-off
handler infrastructure. It converts psci_sys_poweroff to the sys-off callback
prototype and registers it through register_sys_off_handler.

> diff --git a/drivers/firmware/psci/psci.c b/drivers/firmware/psci/psci.c
> index 38ca190d4a22d..d6e9721d11e54 100644
> --- a/drivers/firmware/psci/psci.c
> +++ b/drivers/firmware/psci/psci.c

[ ... ]

> @@ -671,7 +672,8 @@ static void __init psci_0_2_set_functions(void)
>  
>  	register_restart_handler(&psci_sys_reset_nb);
>  
> -	pm_power_off = psci_sys_poweroff;
> +	register_sys_off_handler(SYS_OFF_MODE_POWER_OFF, SYS_OFF_PRIO_FIRMWARE,
> +				 psci_sys_poweroff, NULL);
>  }

Can this cause a regression during early boot?

The function psci_0_2_set_functions() is called extremely early via
psci_dt_init() or psci_acpi_init() from setup_arch(). This happens before
mm_core_init() brings up the slab memory allocator.

Because the priority passed is SYS_OFF_PRIO_FIRMWARE, alloc_sys_off_handler()
falls back to dynamically allocating the handler using kzalloc().

Could calling kzalloc() before the slab allocator is initialized result in a
kernel panic?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260514-smaug-poweroff-v1-0-30f9a4688966@tecnico.ulisboa.pt?part=1