make softfloat type usage more explict

[spoonincode]

See #638 & #702 & their comments for a little initial context.

We should be more explicit across the code when a float128_t (or potentially other softfloat types) is intended to really be a softfloat float128_t. Otherwise it's at risk of getting mixed up with a boost::float128_t or a future std::float128_t.

Some possibilities include placing the softfloat types in to a namespace and/or some sort of using soft_float128_t = softfloat::float128_t alias.

[spoonincode]

moved to AntelopeIO/spring#522