relax CLUT inlining a bit

I decided to take another pass at writing the CLUT code in skcms
as "normal", mostly to get rid of the macros now that we're C++
and can use templates.

While I was at it, I moved the dimension from being a fundamental
property of each clut() function to an argument, and took away any
mention of force inlining or even inlining.  In all this cut skcms'
code size by about 30%.

Change-Id: I4db09ce0bd7886358c3f30fbf1fc89d88ec83221
Reviewed-on: https://skia-review.googlesource.com/c/161585
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@google.com>
1 file changed