linalg: enable 80-bit extended precision for whole library, xdp

[perazz]

Enable xdp precision for BLAS, LAPACK and whole linear algebra library.
This PR now seamlessly enables both 80-bit extended-precision (xdp) and 128-bit quadruple precision (qp) versions of the liear algebra library to be built.

• extended precision functions templated for whole blas, lapack
• new initials defined:

real entities -> s, d, x, q
complex entities -> c, z, y, w

• whole functionality including BLAS/LAPACK interfaces, as well as the whole testsuite, have now removed the constraints to not use xdp.
• both "high" precision LAPACK modules share use of 64-bit double precision for the algorithms that require a "lower precision" kind

cc: @jvdp1 @jalvesz @fortran-lang/stdlib

[jvdp1]

LGTM. Pending a minor comment, it could be merged IMO. It is probably better to merge this one and then review #828 and #832 in light of this PR.

[perazz]

Thanks a lot for the review @jvdp1. I agree: if we get to merge this PR first, then I can add xdp support back into the outstanding PRs, before merging them.

[jalvesz]

LGTM @perazz! It would indeed be good to merge this PR before moving to the pending linalg ones.

[perazz]

Sounds like a good plan @jvdp1 @jalvesz. Let’s wait until tomorrow, and then merge this o e if there are no further comments.

[perazz]

Thank you @jvdp1 @jalvesz, merging.