Re: [PATCH v2 1/9] pmdomain: arm: Fix -EINVAL from scmi_pd_set_perf_state() on state 0

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] Returning success for performance state 0 without notifying the SCMI hardware causes a power leak and genpd state desynchronization.
--

commit 2b6c1834f31c5a0d438704d927e8d116c0dd29ad
Author: Praveen Talari <[email protected]>

pmdomain: arm: Fix -EINVAL from scmi_pd_set_perf_state() on state 0

The commit modifies scmi_pd_set_perf_state() to return 0 (success) instead
of -EINVAL when the requested performance state is 0. This change aims to
accommodate the OPP framework which uses state 0 when a device is runtime
suspended.

> diff --git a/drivers/pmdomain/arm/scmi_perf_domain.c b/drivers/pmdomain/arm/scmi_perf_domain.c
> index 3693423459c9c..e390f902a4442 100644
> --- a/drivers/pmdomain/arm/scmi_perf_domain.c
> +++ b/drivers/pmdomain/arm/scmi_perf_domain.c
> @@ -33,7 +33,7 @@ scmi_pd_set_perf_state(struct generic_pm_domain *genpd, unsigned int state)
>  		return 0;
>  
>  	if (!state)
> -		return -EINVAL;
> +		return 0;

[Severity: Medium]
Will this return 0 without notifying the SCMI hardware of the state change?

Since SCMI performance domains are marked with GENPD_FLAG_ALWAYS_ON, they are
never actually powered off. When all devices in the domain suspend and genpd
aggregates the state to 0, it invokes scmi_pd_set_perf_state() with state 0.

If we immediately return 0 here without calling pd->perf_ops->level_set(),
no request is sent to the firmware.

Does this cause the hardware to get stuck at its last requested
high-performance level instead of scaling down, leading to a silent power
leak and a desynchronization between genpd software state and the hardware?

>  
>  	ret = pd->perf_ops->level_set(pd->ph, pd->domain_id, state, false);
>  	if (ret)

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260827-derive_clk_perf_tbl_from_perf_domain_opp_table-v2-0-091697dbeb02@oss.qualcomm.com?part=1