Re: [PATCH v2 1/2] mfd: nct6694: Switch to devm_mfd_add_devices() and drop IDA

[Ming Yu]

Hi Bart, all,

Thanks for the review.

Bartosz Golaszewski <brgl@xxxxxxxxxx> 於 2026年4月8日週三 下午3:25寫道：
>
> On Wed, Apr 8, 2026 at 7:31 AM <a0282524688@xxxxxxxxx> wrote:
> >
> > From: Ming Yu <a0282524688@xxxxxxxxx>
> >
> > Currently, the nct6694 core driver uses mfd_add_hotplug_devices()
> > and an IDA to manage subdevice IDs.
> >
> > Switch the core implementation to use the managed
> > devm_mfd_add_devices() API, which simplifies the error handling and
> > device lifecycle management. Concurrently, drop the custom IDA
> > implementation and transition to using pdev->id.
> >
> > Signed-off-by: Ming Yu <a0282524688@xxxxxxxxx>
> > ---
>
> This does result in a nice code shrink but I'd split this commit into
> two: one switching to using MFD_CELL_BASIC() with hard-coded devices
> IDs and one completing the transition to devres.
>


You are right that this change is trying to do too much at once, and
splitting it as you suggested would make the series much cleaner.

After looking more closely at the ID handling and hotplug
implications, I realized that switching to devm_mfd_add_devices() and
dropping the IDA is not a good fit for this driver. The current
mfd_add_hotplug_devices() path uses PLATFORM_DEVID_AUTO, which gives
globally unique device IDs and avoids sysfs name collisions. If we
switch to devm_mfd_add_devices() with fixed IDs, multiple identical
NCT6694 devices can end up registering subdevices with the same
platform device names, which would break hotplug support when more
than one device is present.

So I think it is better not to pursue this direction further.

For the next revision, I will drop this part of the change and keep
the existing MFD core logic, including the IDA usage. The series will
focus on adding the nct6694-hif MFD driver only, and I will add the
IDA initialization there as needed.

Thanks again for the suggestion and review.


Best regards,
Ming