check MIR assignment type validity check during codegen

[RalfJung]

I looked at this while debugging #155477, but this makes no progress on that issue.

The first commit improves the check (subtyping should no longer be occurring here) makes the error message a bit more clear. The 2nd commit moves things around so that we can run the same check during codegen as well.

[rustbot]

Some changes occurred to the CTFE machinery

cc @oli-obk, @lcnr

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri

[rustbot]

r? @davidtwco

rustbot has assigned @davidtwco.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler, mir
• compiler, mir expanded to 72 candidates
• Random selection from 17 candidates

[rustbot]

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[RalfJung]

I didn't know what would be a good place for this check as it needs both rustc_middle and rustc_trait_selection (so apparently we cannot put it in rustc_middle). rustc_ty_utils already has those dependencies and otherwise seems fairly lightweight so adding it as dependency in rustc_const_eval and rustc_codegen_ssa is hopefully fine?

View changes since the review

[RalfJung]

Strangely, codegen does not seem to see the same types as the interpreter / Miri. I can't reproduce the ICE there.

[RalfJung]

Ah, transmute is implemented differently in codegen so we're not asking the mir_assign_valid_types question as part of the transmute. So there's not really a place outside the interpreter that could hit #155477.

It could still be worth checking mir_assign_valid_types during codegen, as a sanity check? I can also remove the 2nd commit again if you prefer. (But note that this does resolve an existing FIXME saying we need to find a better place for these utility functions.)

[rust-bors]

☔ The latest upstream changes (presumably #155083) made this pull request unmergeable. Please resolve the merge conflicts.