Implement a global value numbering MIR optimization
The aim of this pass is to avoid repeated computations by reusing past assignments. It is based on an analysis of SSA locals, in order to perform a restricted form of common subexpression elimination.
By opportunity, this pass allows for some simplifications by combining assignments. For instance, this pass could be able to see through projections of aggregates to directly reuse the aggregate field (not in this PR).
We handle references by assigning a different "provenance" index to each `Ref`/`AddressOf` rvalue. This ensure that we do not spuriously merge borrows that should not be merged. Meanwhile, we consider all the derefs of an immutable reference to a freeze type to give the same value:
```rust
_a = *_b // _b is &Freeze
_c = *_b // replaced by _c = _a
```
Amend style guide section for formatting where clauses in type aliases
This PR has two parts:
1. Amend wording about breaking before or after the `=`, which is a style guide bugfix to align it with current rustfmt behavior.
2. Explain how to format trailing (#89122) where clauses, which are preferred in both GATs (#90076) and type aliases (#114662).
r? `@joshtriplett`
make link_llvm_intrinsics and platform_intrinsics features internal
These are both a lot like `feature(intrinsics)`, just slightly different syntax, so IMO it should be treated the same (also in terms of: if you get ICEs with this feature, that's on you -- we are not doing "nice" type-checking for intrinsics).
Rollup of 4 pull requests
Successful merges:
- #115934 (Split out the stable part of smir into its own crate to prevent accidental usage of forever unstable things)
- #116149 (Anonymize binders for `refining_impl_trait` check)
- #116178 (Add test for `const async fn`)
- #116187 (Add context to `let: Ty = loop { break };`)
r? `@ghost`
`@rustbot` modify labels: rollup
Anonymize binders for `refining_impl_trait` check
We're naively using the equality impl for `ty::Clause` in the refinement check, which is okay *except* for binders, which carry some information about where they come from in the AST. Those locations are not gonna be equal between traits and impls, so anonymize those clauses so that this doesn't matter.
Fixes#116135
Split out the stable part of smir into its own crate to prevent accidental usage of forever unstable things
Some groundwork for being able to work on https://github.com/rust-lang/project-stable-mir/issues/27 at all
r? `@spastorino`
Skip MIR pass `UnreachablePropagation` when coverage is enabled
When coverage instrumentation and MIR opts are both enabled, coverage relies on two assumptions:
- MIR opts that would delete `StatementKind::Coverage` statements instead move them into bb0 and change them to `CoverageKind::Unreachable`.
- MIR opts won't delete all `CoverageKind::Counter` statements from an instrumented function.
Most MIR opts naturally satisfy the second assumption, because they won't remove coverage statements from bb0, but `UnreachablePropagation` can do so if it finds that bb0 is unreachable. If this happens, LLVM thinks the function isn't instrumented, and it vanishes from coverage reports.
A proper solution won't be possible until after per-function coverage info lands in #116046, but for now we can avoid the problem by turning off this particular pass when coverage instrumentation is enabled.
---
cc `@cjgillot` since I found this while investigating coverage problems encountered by #113970
`@rustbot` label +A-code-coverage +A-mir-opt
Don't store lazyness in `DefKind::TyAlias`
1. Don't store lazyness of a type alias in its `DefKind`, but instead via a query.
2. This allows us to treat type aliases as lazy if `#[feature(lazy_type_alias)]` *OR* if the alias contains a TAIT, rather than having checks for both in separate parts of the codebase.
r? `@oli-obk` cc `@fmease`
Only prevent field projections into opaque types, not types containing opaque types
fixes https://github.com/rust-lang/rust/issues/115778
I did not think that original condition through properly... I'll also need to check the similar check around the other `ProjectionKind::OpaqueCast` creation site (this one is in hir, the other one is in mir), but I'll do that change in another PR that doesn't go into a beta backport.
subst -> instantiate
continues #110793, there are still quite a few uses of `subst` and `substitute`, but changing them all in the same PR was a bit too much, so I've stopped here for now.
Correct codegen of `ConstValue::Indirect` scalar and scalar pair
This concerns 3 tricky cases with `ConstValue::Indirect`:
- if we want a non-pointer scalar;
- if we have non-zero offset;
- if offset points to uninit memory => generate `poison` instead of an ICE. This case could happen in unreachable code, trying to extract a field from the wrong variant.
Those cases are not currently emitted by the compiler, but are exercised by https://github.com/rust-lang/rust/pull/116012.
Gate and validate `#[rustc_safe_intrinsic]`
Copied over from #116159:
> This was added as ungated in https://github.com/rust-lang/rust/pull/100719/files#diff-09c366d3ad3ec9a42125253b610ca83cad6b156aa2a723f6c7e83eddef7b1e8fR502, probably because the author looked at the surrounding attributes, which are ungated because they are gated specially behind the staged_api feature.
>
> I don't think we need to crater this, the attribute is entirely useless without the intrinsics feature, which is already unstable..
r? ``@Nilstrieb``
Update books
## rust-embedded/book
1 commits in 99ad2847b865e96d8ae7b333d3ee96963557e621..eac173690b8cc99094e1d88bd49dd61127fbd285
2023-09-12 07:34:44 UTC to 2023-09-12 07:34:44 UTC
- USB connector-type correction (rust-embedded/book#360)
## rust-lang/nomicon
1 commits in e3f3af69dce71cd37a785bccb7e58449197d940c..ddfa4214487686e91b21aa29afb972c08a8f0d5b
2023-09-22 17:04:10 UTC to 2023-09-22 17:04:10 UTC
- Fill "Beneath `std`" (rust-lang/nomicon#413)
## rust-lang/reference
1 commits in ee7c676fd6e287459cb407337652412c990686c0..5262e1c3b43a2c489df8f6717683a44c7a2260fd
2023-09-18 18:28:31 UTC to 2023-09-18 18:28:31 UTC
- we reserve the right to reduce our amount of UB (rust-lang/reference#1397)
## rust-lang/rustc-dev-guide
8 commits in 08bb147d51e815b96e8db7ba4cf870f201c11ff8..a13b7c28ed705891c681ce5417b3d1cdb12cecd1
2023-09-25 05:14:41 UTC to 2023-09-11 21:29:18 UTC
- Clarify all the `{AP,RP}IT{,IT}` impl trait types (rust-lang/rustc-dev-guide#1798)
- Modify build instructions for optimized build (rust-lang/rustc-dev-guide#1795)
- Remove outdated references to coverage debug code (rust-lang/rustc-dev-guide#1797)
- Add deep dive document about early/late bound parameters interacting with turbofish (rust-lang/rustc-dev-guide#1794)
- explain the MIR const vs TY const situation (rust-lang/rustc-dev-guide#1793)
- fix type name (rust-lang/rustc-dev-guide#1792)
- Clarify that `run-coverage` only runs in some of the CI jobs (rust-lang/rustc-dev-guide#1791)
- Document the `coverage-map` and `run-coverage` test suites (rust-lang/rustc-dev-guide#1790)
lint towards rejecting consts in patterns that do not implement PartialEq
I think we definitely don't want to allow such consts, so even while the general plan around structural matching is up in the air, we can start the process of getting non-PartialEq matches out of the ecosystem.
Don't use a thread to load the dep graph
This removes the use of a thread to load the dep graph. It's not currently useful as we immediately block on it.
r? `@oli-obk`
When coverage instrumentation and MIR opts are both enabled, coverage relies on
two assumptions:
- MIR opts that would delete `StatementKind::Coverage` statements instead move
them into bb0 and change them to `CoverageKind::Unreachable`.
- MIR opts won't delete all `CoverageKind::Counter` statements from an
instrumented function.
Most MIR opts naturally satisfy the second assumption, because they won't
remove coverage statements from bb0, but `UnreachablePropagation` can do so if
it finds that bb0 is unreachable. If this happens, LLVM thinks the function
isn't instrumented, and it vanishes from coverage reports.
A proper solution won't be possible until after per-function coverage info
lands in #116046, but for now we can avoid the problem by turning off this
particular pass when coverage instrumentation is enabled.
ConstParamTy: require Eq as supertrait
As discussed with `@BoxyUwu` [on Zulip](https://rust-lang.zulipchat.com/#narrow/stream/260443-project-const-generics/topic/.60ConstParamTy.60.20and.20.60Eq.60).
We want to say that valtree equality on const generic params agrees with `==`, but that only makes sense if `==` actually exists, hence we should have an appropriate bound. Valtree equality is an equivalence relation, so such a type can always be `Eq` and not just `PartialEq`.
Properly print cstr literals in `proc_macro::Literal::to_string`
Previously we printed the contents of the string, rather than the actual string literal (e.g. `the c string` instead of `c"the c string"`).
Fixes#112820
cc #105723
Use placeholders to prevent using inferred RPITIT types to imply their own well-formedness
The issue here is that we use the same signature to do RPITIT inference as we do to compute implied bounds. To fix this, when gathering the assumed wf types for the method, we replace all of the infer vars (that will be eventually used to infer RPITIT types) with type placeholders, which imply nothing about lifetime bounds.
This solution kind of sucks, but I'm not certain there's another feasible way to fix this. If anyone has a better solution, I'd be glad to hear it.
My naive first solution was, instead of using placeholders, to replace the signature with the RPITIT projections that it originally started out with. But turns out that we can't just use the unnormalized signature of the trait method in `implied_outlives_bounds` since we normalize during WF computation -- that would cause a query cycle in `collect_return_position_impl_trait_in_trait_tys`.
idk who to request review...
r? `@lcnr` or `@aliemjay` i guess.
Fixes#116060
Point at more causes of expectation of break value when possible
Follow up to #116071.
r? `@compiler-errors`
Disregard the first commit, which is in the other PR.
Pass name of object file to LLVM so it can correctly emit S_OBJNAME in pdb files on Windows
This should be the remaining fix to close https://github.com/rust-lang/rust/issues/96475
Setting ObjectFilenameForDebug in llvm::TargetOptions, so llvm it can emit S_OBJNAME in pdb files on Windows.
Without a proper pdb parsing I am not able to add a unit test for this. The string is already appearing in the pdb file so I cannot just use grep.
`@rustbot` label: +A-debuginfo
