EAMxx: in infra, ensure comments/code reflect generality of masking treatment

[mahf708]

comments like

E3SM/components/eamxx/src/share/grid/remap/coarsening_remapper.cpp

Lines 273 to 274 in 1ad1b5f

	// If the mask comes from FieldAtLevel, it's only defined on columns (rank=1)
	// If the mask comes from vert interpolation remapper, it is defined on ncols x nlevs (rank=2)

may make it sound like this is the only case supported. I suggest a quick walk through the infra code (remap, io, etc.) to reflect the new reality of more extensive masking we are now using.

Make the assumption clearer and more general, e.g.,

// NOTE that in eamxx, masking is at most rank-2, either on COL only or on COLxLEV only

Even better, add documentation to clarify how masking works and use cases under components/eamxx/docs.

--

@bartgol I didn't assign anyone here, because this is the type of thing a curious reader may be able to do and document ... or you know, maybe that one helpful 🤖🤖🤖🤖🤖🤖🤖🤖🤖 ;) (copilot, if you're reading...)

ran into this in #7652 with annoying layout-like errors that took a bit of time to trace...

[bartgol]

If this EKAT pr gets integrated, we could maybe rework the masking framework, and take care of this comment here in the process. I am personally in favor of integrating the ekat pr, unless we find out that there is already a similar functionality in kokkos (@tcclevenger is scoping it out).

[bartgol]

The aforementioned EKAT PR has been merged. We can discuss the possibility to revisit the whole masks infrastructure.