Re: [PATCH v24 08/12] mfd: psci-mfd: Add psci-reboot-mode child cell

[Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>]

On 8/3/2026 3:37 PM, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Medium] The MFD probe function swallows errors from devm_mfd_add_devices() instead of propagating them to abort the probe.
> --
> 
> --- Patch [8]: [PATCH v24 08/12] mfd: psci-mfd: Add psci-reboot-mode child cell ---
> commit 88cfe078ed8299db3dd71fd877f29ff443d5037e
> Author: Shivendra Pratap <shivendra.pratap@oss.qualcomm.com>
> 
> mfd: psci-mfd: Add psci-reboot-mode child cell
>      
> This commit exposes the PSCI reboot-mode node as a proper child cell using
> a named_fwnode. It also registers the psci-cpuidle-domain cell first to
> isolate it from reboot-mode failures.
> 
>> diff --git a/drivers/mfd/psci-mfd.c b/drivers/mfd/psci-mfd.c
>> index 7affd6bb09dd8..7befc4fa86f5d 100644
>> --- a/drivers/mfd/psci-mfd.c
>> +++ b/drivers/mfd/psci-mfd.c
> 
> [ ... ]
> 
>> @@ -14,10 +14,35 @@ static const struct mfd_cell psci_cells[] = {
>>   static int psci_mfd_probe(struct platform_device *pdev)
>>   {
>> -	return devm_mfd_add_devices(&pdev->dev, PLATFORM_DEVID_AUTO, psci_cells,
>> +	int ret;
>> +
>> +	ret = devm_mfd_add_devices(&pdev->dev, PLATFORM_DEVID_AUTO, psci_cells,
>>   				   ARRAY_SIZE(psci_cells), NULL, 0, NULL);
>> +	if (ret)
>> +		goto out;
>> +
>> +	ret = devm_mfd_add_devices(&pdev->dev, PLATFORM_DEVID_AUTO,
>> +				   psci_reboot_mode_cell,
>> +				   ARRAY_SIZE(psci_reboot_mode_cell),
>> +				   NULL, 0, NULL);
>> +	if (ret) {
>> +		if (ret == -ENOMEM)
>> +			goto out;
>> +		dev_warn(&pdev->dev, "reboot-mode child cell failed to add: %d\n", ret);
>> +		ret = 0;
>> +	}
> 
> [Severity: Medium]
> Should this propagate the error instead of returning 0?
> 
> According to the MFD subsystem guidelines, if a child device fails to
> register during an MFD probe, the entire probe must fail and unwind
> previously registered children. By setting ret = 0 here, the driver reports
> success for the parent probe even if the reboot-mode child cell fails to
> register. Does this leave the device in an incomplete state and violate the
> requirement to propagate devm_mfd_add_devices() errors?

This is done on purpose.
psci-mfd adds two child cells, psci-cpuidle-domain and psci-reboot-mode.
psci-cpuidle-domain is a critical driver for device device bootup on 
supported systems. psci-reboot-mode child is hence added separately as 
optional child which will not affect the probe of psci-mfd.

thanks,
Shivendra