[alsa-devel] [PATCH v2 3/3] dt/bindings: Add milivolt, milliamps, and nanohenry units

Brian Austin brian.austin at cirrus.com
Wed Oct 19 15:15:49 CEST 2016


On Tue, 18 Oct 2016, Rob Herring wrote:

> On Tue, Oct 18, 2016 at 12:43:48AM +0000, Handrigan, Paul wrote:
> > 
> > 
> > 
> > >>  Electricity
> > >>  ----------------------------------------
> > >>  -microamp      : micro amps
> > >> +-milliamp      : milli amps
> > >
> > >Please explain why you need need this. Is 2^32 uA not enough range for
> > >you? While we do have multiple units for some, that's because people
> > >picked arbitrary units for their properties before this doc existed.
> > >Ideally, we only have 1 unit to prevent that problem.
> > 
> > 2^32 would be enough, however, microamps would not be a good
> > representation for
> > this particular entry. For example, if the current limit is 2.1A the
> > entry will have to be 2100000.  Micro amps will not be the best
> > representation for this value from a scientific notation perspective.
> > Since the scale is in miiliamps as it is defined in the data sheet of
> > this device, it will be best to stay consistent since an error in
> > conversion
> > could cause serious problems.
> > 
> > Also, I really don¹t see why adding milliamps and millivolts as being a
> > major
> > problems since both of them are commonly used.
> 
> As I explained, we want to have some uniformity in properties especially 
> across different bindings. That's not something I want to check in every
> review. Telling people to add units is enough.
> 
> I don't see how adding '000' is a problem.
> 
> Rob
> 
Why is it such a big deal to not have a representation for all the common 
units of measurement? They are already defined so why not add them to your 
suffix file? I would think that if you are going to make us use this 
suffix devicetree thing, you would not restrict us to use only the units 
you think are appropriate.

I wouldn't think you would restrict us to saying it is a 5700ml Hemi V8 
would you?


Brian


More information about the Alsa-devel mailing list