…Crate`.

Add all remaining `DefKind`s.

r? @eddyb or @Centril

~~I'm not sure if this is what you were thinking of. There are also a few places where I'm not sure what the correct choice is because I don't fully understand the meaning of some variants.~~

~~In general, it feels a bit odd to add some of these as `DefKind`s (e.g. `Arm`) because they don't feel like definitions. Are there things that it makes sense not to add?~~

Move `MapInPlace` to rustc_data_structures

Follow-up of rust-lang#67786, it fits the purpose of rustc_data_structures.
r? @petrochenkov

…ic-morse

miri-unleashed: test that we detect heap allocations

This removes the second-to-last use of `IS_SUPPORTED_IN_MIRI = false`.

r? @ecstatic-morse @oli-obk

Minor improvements to -Zprofile

- `-Zprofile` is broken with codegen units because GCOV assumes that each source file corresponds to one object file. This PR makes `-Zprofile` automatically set codegen units to 1 and gives an error if `-Ccodegen-units=X` is specified on the command line (with `X != 1`).
- The `profiler_builtins` crate is not suitable for `no_std` applications since it contains C code that depends on libc. In such cases a custom implementation of the LLVM gcov API (`llvm_gcov_init`, `llvm_gcda_*`) is needed. This PR adds `-Zno-profiler-runtime` flag which inhibits automatic injection of the `profiler_builtins` crate.

cc @whitequark who implemented the original `-Zprofile` support

Explain why we shouldn't add inline attr to into_vec

Follow-up of rust-lang#71204
r? @RalfJung