[PATCH 9/9] soc/qcom: Add REVID driver

[Greg Kroah-Hartman]

On Sun, Jul 26, 2020 at 01:40:46PM +0200, Konrad Dybcio wrote:
> >Horrible global symbol name.  Who calls this?
> 
> Welcome to development on qcom platforms :D

Yeah, I have seen :(

> >This is the last patch in
> >the series, so if there is no user for this, please don't export it.
> 
> Other downstream drivers make use of it.. need to get this up first, sorry :V

Then use a proper name for it, with a correct prefix, and export it with
EXPORT_SYMBOL_GPL() please.

> >Why do you need a .h file in the include directory if only a single .c
> >file needs it?  Just put that info in the .c file itself.
> 
> Again, other downstream drivers which some people and I intend to
> bring to upstream standards use that to access the PMIC model/hw revision.

But all of those defines are not needed, just the function name, right?

> >But again, who uses this module?  If it's only good for a single line in
> >the kernel log, that feels like a huge waste to me.
> 
> downstream-kernel-dir$ rg -l qpnp-revid.h | wc -l
> 25
> 
> So yeah, quite a bunch of other qcom-specific drivers.
> 
> I'll try to fix these and send a v2.

Hold off on this and wait until there is a real user for it, as we can
not take api additions without a user, otherwise we will just come along
and delete the code.

Submit it as part of a series that actually uses the function, so we can
at least see if the function makes sense to be used in that way as well,
right now we have no clue.

thanks,

greg k-h