[MC][AsmParser] Rename mnemonics according to the specification

[atrosinenko]

At now, most instructions are supported with some mnemonic, but these mnemonics can differ in the public specification, pre-existing code (LLVM backend and standalone assembler) and possibly the updated spec.

This issue covers instruction mnemonics in a broad sense (for example, the <op> part in context.<op>) - those cases that can be resolved by simply changing string literals in C++ code or *.td files without changing any logic.

See #440 for conditional codes.

[atrosinenko]

As explained by @hedgar2017, replaced gas with ergs in asm mnemonics (not touching the intrinsic names yet): 3fc0963

[asl]

We will wait for the updated ISA spec to perform the rename

[asl]

@akiramenai @hedgar2017 @sayon So, what is the ultimate resolution on the assembler syntax from 1.6 ISA spec?

For example https://matter-labs.github.io/eravm-spec/spec.html#NearRetDefinition says:

Syntax
  ret in1
Legacy syntax
  ret.ok aliased as ret

How do you want to handle the legacy:

1. Make backend produce only new variant, but let assembler accept the old one as an alias (and silently "upgrade"). So, it will be just ret and ret.ok will be treated as a mnemonic alias. Disassembler will produce new mnemonic.
2. Make both backend and assembler support only new variant. Old one (ret.ok) will be rejected everywhere.

[hedgar2017]

@asl #2, let's reject all legacy.

[atrosinenko]

What is the new syntax for .first modifier of log and logl1 instructions?

[asl]

@sayon Please check the question above

[sayon]

Let's do .f for first. I will update it in spec tonight. CC @hedgar2017

[hedgar2017]

Hey @sayon @atrosinenko, do we have a single place listing all mnemonics to sort them all out?

[sayon]

Yes, in the isa redesign proposal, but i can make a separate summary table in spec for convenience -- been thinking about it.