small regions refactoring

these commits should be fairly self-contained

r? rust-lang/types

interpret: refactor allocation info query

We now have an infallible function that also tells us which kind of allocation we are talking about.
Also we do longer have to distinguish between data and function allocations for liveness.

This will help us to avoid "catching" `InterpError`s in Miri.
r? `@oli-obk`

Fix span issues in object safety suggestions

Fixes #98500

Fix rustdoc argument error

Fixes #88756.

It's a take over of #88831. I cherry-picked the commits, fixed the merge conflict and the failing test.

cc `@inashivb` `@jyn514`

r? `@notriddle`

Rollup of 5 pull requests

Successful merges:

 - #97389 (Improve memory ordering diagnostics)
 - #97780 (Check ADT field is well-formed before checking it is sized)
 - #98530 (compiletest: add issue number param to `known-bug`)
 - #98556 (Fix builds on Windows (closes #98546))
 - #98561 (Fix spelling in SAFETY comment)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup

see comment added to the field in `VerifyBoundCx`.

Fix spelling in SAFETY comment

"can not" should be "cannot", and add punctuation.

Fix builds on Windows (closes #98546)

closes https://github.com/rust-lang/rust/issues/98546

compiletest: add issue number param to `known-bug`

I was getting some errors while testing this, but I'm pretty sure that was unrelated to my changes.

Closes #98436

> Basically, instead of `// known-bug`, do `// known-bug #00000` or maybe `// known-bug chalk#00`?
>
> From: https://rust-lang.zulipchat.com/#narrow/stream/326866-t-types.2Fnominated/topic/.2398095.3A.20NLL.3A.20unsound.20verification.20of.20higher.20ranked.20outlives.E2.80.A6/near/287258738

I also added an `unknown` escape-hatch because I didn't find corresponding issues for every `// known-bug`.

The syntax also ended up being `// known-bug: `, because of `set_name_value_directive`.

Check ADT field is well-formed before checking it is sized

Fixes #96810.

There is one diagnostics regression, in [`src/test/ui/generic-associated-types/bugs/issue-80626.stderr`](https://github.com/rust-lang/rust/pull/97780/files#diff-53795946378e78a0af23a10277c628ff79091c18090fdc385801ee70c1ba6963). I am not super concerned about it, since it's GAT related.
We _could_ fix it, possibly by using the `FieldSized` obligation cause code instead of `BuiltinDerivedObligation`. But that would require changing `Sized` trait confirmation and the `adt_sized_constraint` query.

Improve memory ordering diagnostics

Before:

![image](https://user-images.githubusercontent.com/783247/170234545-891cac30-eaa2-4186-847b-35cd51e00f2b.png)

After:

![image](https://user-images.githubusercontent.com/783247/170239684-645f186f-5a02-4eb9-8651-2e5fe9591352.png)

---

Before this change, the compiler suggests the failure ordering is too strong and suggests choosing a weaker ordering. After this change, it instead suggests the success ordering is not strong enough, and suggests chosing a stronger one. This is more likely to be correct.

Also, before this change, the compiler suggested downgrading an invalid AcqRel failure ordering to Relaxed, without mentioning Acquire as an option.

Rebase LLVM submodule

This is a rebase of our LLVM fork onto LLVM 14.0.5, which is intended to be the last release of the 14.x series. Additionally:

 * I've dropped three patches that were needed to build on FreeBSD 11, which is no longer necessary after #97944.
 * I've dropped some cherry-picks that were later reverted.
 * I've cherry-picked https://github.com/llvm/llvm-project/commit/caa2a829cdf905a5e8664d96a464d414b2adb42e for https://github.com/rust-lang/rust/issues/96797 (fyi `@Amanieu)`

Perform coherence checking per impl.

r? `@ghost`

We now have an infallible function that also tells us which kind of allocation we are talking about.
Also we do longer have to distinguish between data and function allocations for liveness.

"can not" should be "cannot", and add punctuation.

rustc_target: Add convenience functions for adding linker arguments

They ensure that lld and non-lld linker flavors get the same set of arguments.

The second commit also adds some tests checking for linker argument inconsistencies, and tweaks some arguments to fix those inconsistencies.

proc_macro/bridge: cache static spans in proc_macro's client thread-local state

This is the second part of https://github.com/rust-lang/rust/pull/86822, split off as requested in https://github.com/rust-lang/rust/pull/86822#pullrequestreview-1008655452. This patch removes the RPC calls required for the very common operations of `Span::call_site()`, `Span::def_site()` and `Span::mixed_site()`.

Some notes:

This part is one of the ones I don't love as a final solution from a design standpoint, because I don't like how the spans are serialized immediately at macro invocation. I think a more elegant solution might've been to reserve special IDs for `call_site`, `def_site`, and `mixed_site` at compile time (either starting at 1 or from `u32::MAX`) and making reading a Span handle automatically map these IDs to the relevant values, rather than doing extra serialization.

This would also have an advantage for potential future work to allow `proc_macro` to operate more independently from the compiler (e.g. to reduce the necessity of `proc-macro2`), as methods like `Span::call_site()` could be made to function without access to the compiler backend.

That was unfortunately tricky to do at the time, as this was the first part I wrote of the patches. After the later part (#98188, #98189), the other uses of `InternedStore` are removed meaning that a custom serialization strategy for `Span` is easier to implement.

If we want to go that path, we'll still need the majority of the work to split the bridge object and introduce the `Context` trait for free methods, and it will be easier to do after `Span` is the only user of `InternedStore` (after #98189).

Rollup of 11 pull requests

Successful merges:

 - #97140 (std: use an event-flag-based thread parker on SOLID)
 - #97295 ([rustc_parse] Forbid `let`s in certain places)
 - #97743 (make const_err show up in future breakage reports)
 - #97908 (Stabilize NonZero* checked operations constness.)
 - #98297 (Transform help popup into a pocket menu)
 - #98428 (macros: use typed identifiers in diag and subdiag derive)
 - #98528 (Respect --color when building rustbuild itself)
 - #98535 (Add regression test for generic const in rustdoc)
 - #98538 (Add a ui test for issue #91883)
 - #98540 (Add regression test for #87558)
 - #98541 (Update `std::alloc::System` doc example code style)

Failed merges:

r? `@ghost`
`@rustbot` modify labels: rollup

Update `std::alloc::System` doc example code style

`return` on the last line of a block is unidiomatic so I don't think the example should be using that here

Add regression test for #87558

Fixes #87558

Add a ui test for issue #91883

closes #91883

Add regression test for generic const in rustdoc

Fixes #92859.

r? ```@lcnr```

Respect --color when building rustbuild itself

Separated out from https://github.com/rust-lang/rust/pull/95503.

macros: use typed identifiers in diag and subdiag derive

Using typed identifiers instead of strings with the Fluent identifiers in the diagnostic and subdiagnostic derives - this enables the diagnostic derive to benefit from the compile-time validation that comes with typed identifiers, namely that use of a non-existent Fluent identifier will not compile.

r? `````@oli-obk`````

Transform help popup into a pocket menu

Just like we moved the settings menu into a "pocket menu", it's doing the same to the help popup.

You can test it [here](https://rustdoc.crud.net/imperio/help-pocket-menu/doc/foo/index.html) and here is a screenshot:

![Screenshot from 2022-06-20 20-58-29](https://user-images.githubusercontent.com/3050060/174663718-538e9d11-3bf9-48b2-8909-f9bfe75af135.png)

r? ``````````@jsha``````````

Stabilize NonZero* checked operations constness.

Partial stabilization for #97547 (continued).

make const_err show up in future breakage reports

As tracked in https://github.com/rust-lang/rust/issues/71800, const_err should become a hard error Any Day Now (TM). I'd love to move forward with that sooner rather than later; it has been deny-by-default for many years and a future incompat lint since https://github.com/rust-lang/rust/pull/80394 (landed more than a year ago). Some CTFE errors are already hard errors since https://github.com/rust-lang/rust/pull/86194. But before we truly make it a hard error in all cases, we now have one more intermediate step we can take -- to make it show up in future breakage reports.

Cc `````@rust-lang/wg-const-eval`````

[rustc_parse] Forbid `let`s in certain places

Currently only forbids in locals to resolve https://github.com/rust-lang/rust/pull/94927#issuecomment-1099605024 but feel free to point any other places.

std: use an event-flag-based thread parker on SOLID

`Mutex` and `Condvar` are being replaced by more efficient implementations, which need thread parking themselves (see #93740). Therefore, the generic `Parker` needs to be replaced on all platforms where the new lock implementation will be used, which, after #96393, are SOLID, SGX and Hermit (more PRs coming soon).

SOLID, conforming to the [μITRON specification](http://www.ertl.jp/ITRON/SPEC/FILE/mitron-400e.pdf), has event flags, which are a thread parking primitive very similar to `Parker`. However, they do not make any atomic ordering guarantees (even though those can probably be assumed) and necessitate a system call even when the thread token is already available. Hence, this `Parker`, like the Windows parker, uses an extra atomic state variable.

I future-proofed the code by wrapping the event flag in a `WaitFlag` structure, as both SGX and Hermit can share the Parker implementation, they just have slightly different primitives (SGX uses signals and Hermit has a thread blocking API).

`````@kawadakk````` I assume you are the target maintainer? Could you test this for me?

Improve `derive(Debug)`

r? `@ghost`