RE: [PATCH 3/3] hwmon: (pmbus/fd5121): Add support FD5121, FD5123 and FD5125
From: Selvamani Rajagopal
Date: Tue Jun 23 2026 - 19:13:03 EST
> -----Original Message-----
> From: Guenter Roeck <groeck7@xxxxxxxxx> On Behalf Of Guenter Roeck
> Subject: Re: [PATCH 3/3] hwmon: (pmbus/fd5121): Add support FD5121, FD5123 and
> FD5125
>
>
> > + &sensor_dev_attr_vout_transition_rate.dev_attr.attr,
> > + NULL
>
> This is a complete no-go. We do not explose raw register data as sysfs
> attributes. You may expose essential register data as debugfs files,
> but only those deemed necessary. The above is just "let's blindly
> expose everything". Most of the above should be programmed in manufacturing
> and not be touched subsequently, much less as writeable attributes.
> Writing bad/unexpected values into many of those attributes can turn
> a board into a brick. It is bad enough that/if this is even possible,
> but exposing it as sysfs attribute would be a terrible idea.
>
> I am not going to review this driver any further at this point.
>
> Guenter
>
Thanks for your comment. Understood and agree on comment on exposing everything without enforcing what is being written/read.
We will discuss internally to see how to go about it. We need some custom registers for customers. We will identify those and
expose under debugfs, if needed.
> >