In general it is not correct to inline a callee with a target features
that are subset of the callee. Require target features to match exactly
during inlining.

The exact match could be potentially relaxed, but this would require
identifying specific feature that are allowed to differ, those that need
to match, and those that can be present in caller but not in callee.

This resolves MIR part of rust-lang#116573. For other concerns with respect to
the previous implementation also see areInlineCompatible in LLVM.

This flag was stabilized in rustc 1.60.0 as `-C instrument-coverage`, but the
old unstable flag was kept around as an alias to ease migration.

…ErrorGuaranteed`, even if that error is only emitted by `check_modwitem_types`

…rrors

Remove support for alias `-Z instrument-coverage`

This flag was stabilized in rustc 1.60.0 (2022-04-07) as `-C instrument-coverage`, but the old unstable flag was kept around (with a warning) as an alias to ease migration.

It should now be reasonable to remove the somewhat tricky code that implemented that alias.

Fixes rust-lang#116980.

…oli-obk

Require target features to match exactly during inlining

In general it is not correct to inline a callee with a target features
that are subset of the callee. Require target features to match exactly
during inlining.

The exact match could be potentially relaxed, but this would require
identifying specific feature that are allowed to differ, those that need
to match, and those that can be present in caller but not in callee.

This resolves MIR part of rust-lang#116573. For other concerns with respect to
the previous implementation also see areInlineCompatible in LLVM.

…=chenyukang

Fix unwrap suggestion for async fn

Use `body_fn_sig` to get the expected return type of the function instead of `ret_coercion` in `FnCtxt`. This avoids accessing the `ret_coercion` when it's already mutably borrowed (e.g. when checking `return` expressions).

Fixes rust-lang#117144

r? `@chenyukang`

…2, r=bjorn3

implement C ABI lowering for CSKY

fix rust-lang/compiler-builtins#551

 ​Reference: [CSKY ABI Manual](https://occ-oss-prod.oss-cn-hangzhou.aliyuncs.com/resource//1695027452256/T-HEAD_800_Series_ABI_Standards_Manual.pdf)
 ​
Reference: [Clang CSKY lowering code](https://github.com/llvm/llvm-project/blob/4a074f32a6914f2a8d7215d78758c24942dddc3d/clang/lib/CodeGen/Targets/CSKY.cpp#L76-L162)

r? `@bjorn3`

Work around the fact that `check_mod_type_wf` may spuriously return `ErrorGuaranteed`

Even if that error is only emitted by `check_mod_item_types`.

fixes rust-lang#117153

A cleaner refactoring would merge/chain these queries in ways that ensure we only actually get an `ErrorGuaranteed` if there was an error emitted.

…piler-errors

compiletest: Display compilation errors in mir-opt tests

Previously when compilation failed the `check_mir_dump` would panic first, so we would never display the compiler output.

…errors

Make `Iterator` a lang item

r? `@compiler-errors`

pulled out of rust-lang#116447

We're doing this change on its own, because iterator was the one diagnostic item that was load bearing on us correctly emitting errors about `diagnostic_item` mis-uses. It was used in some diagnostics as an early abort, before the actual checks of the diagnostic, so effectively the compiler was *unconditionally* checking for the iterator diagnostic item, even if it didn't emit any diagnostics. Changing those uses to use the lang item, caused us not to invoke the `all_diagnostic_items` query anymore, which then caused us to miss some issues around diagnostic items until they were actually used.

The reason we keep the diagnostic item around is that clippy uses it a lot and having `Iterator` be a lang item and a diagnostic item at the same time doesn't cost us anything, but makes clippy's internal code simpler