Strict single-letter extensions on Table 28.1 (Standard ISA extension names)

[a4lg]

This branch resolves two issues on Table 28.1 (Standard ISA extension names):

1. B, K and J are not going to be single-letter extensions, making those appearance in the Table 28.1 unnecessary (or even misleading and harmful).
   Those letters are referred in the section 28.6 (Additional Standard Extension Names) and Table 28.1 does not need them.
2. V is plural (actually V is an extension with some variable parameters such as vector length).

See commit c1c77c4 for background of this proposal for B and J.

For K, those might be helpful:

1. https://github.com/riscv/riscv-crypto/releases/tag/v0.9.3-scalar (with misa.K)
2. https://github.com/riscv/riscv-crypto/releases/tag/v0.9.4-scalar (WITHOUT misa.K)
3. https://github.com/riscv/riscv-crypto/releases/tag/v1.0.0-scalar (ratified; WITHOUT misa.K)

And how Ben Marshall responded in https://lists.riscv.org/g/tech-crypto-ext/topic/riscv_k_in_misa_risc_v/85354476 :

It says that misa.K bit is no more used. Does this mean that -misa=rv32gk is no more acceptable?

Indeed. There is no longer a single letter which enables the entire extension. You should use one of the Zk* architecture strings to enable support in the compiler.

[a4lg]

BTW, I noticed that V requires (F and) D but does it mean V implies D? If so, that should be reflected in the table, too.

[aswaterman]

In this context, I view “require” and “imply” as equivalent, meaning that e.g. RV32IV and RV32IDV and RV32IFDV are all equivalent.

[aswaterman]

I know @kasanovic aims to rewrite the Naming chapter, but I think it’s best to merge this PR anyway to reduce confusion.