Tier two has grown quite organically, and the two different implementations (the interpreter and JIT) have grown with it. In many places, we're repeating the same logic, or doing things in two or three different ways.

I'd like clean up tier two a bit, in the process making the tier two interpreter and JIT more similar. We enter the JIT by making a call, so I don't see a very compelling reason why the tier two interpreter needs to do anything different.

There are also some unused macros and labels lying around, and a few labels that are only used by one or two instructions. These can be cleaned up as well.

I'll start by generally tidying things up in one PR, then moving the tier two interpreter out into its own function in a follow-up PR.

Linked PRs

• GH-129709: Clean up tier two #129710