Merge BorrowKind::Unique into BorrowKind::Mut

[zirconium-n]

Fixes#112072

Might have conflict with #112070

r? @lcnr

I'm not sure what's the suitable change in a couple places.

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

This PR changes MIR

cc @oli-obk, @RalfJung, @JakobDegen, @davidtwco, @celinval, @vakaras

Some changes might have occurred in exhaustiveness checking

cc @Nadrieril

[zirconium-n]

I commented on places I'm not sure.

[lcnr]

quickly skimmed it but need to take some time to think about the places where Unique currently differs from Mut, will only get to that next week probably

[zirconium-n]

quickly skimmed it but need to take some time to think about the places where Unique currently differs from Mut, will only get to that next week probably

From my intuition, Unique is just a Mut except it does not require a mutable binding. (Maybe that's oversimplification?)

So here's some personal observation:

0. Diagnostics obviously should differ.

1. These ones should unify Unique and Mut. Since they are checking exactly whether it's a mutable borrow.
   

   rust/compiler/rustc_const_eval/src/transform/check_consts/resolver.rs

   Lines 103 to 110 in e29821f

   	fnref_allows_mutation(&self,kind: mir::BorrowKind,place: mir::Place<'tcx>) -> bool{
   	match kind {
   	mir::BorrowKind::Mut{ .. } => true,
   	mir::BorrowKind::Shared | mir::BorrowKind::Shallow | mir::BorrowKind::Unique => {
   	self.shared_borrow_allows_mutation(place)
   	}
   	}
   	}

   
   

   rust/compiler/rustc_middle/src/mir/mod.rs

   Lines 2030 to 2035 in e29821f

   	pubfnmutability(&self) -> Mutability{
   	match*self{
   	BorrowKind::Shared | BorrowKind::Shallow | BorrowKind::Unique => Mutability::Not,
   	BorrowKind::Mut{ .. } => Mutability::Mut,
   	}
   	}

   
   

   rust/compiler/rustc_mir_build/src/check_unsafety.rs

   Lines 438 to 454 in e29821f

   	ExprKind::Borrow{ borrow_kind, arg } => {
   	letmut visitor = LayoutConstrainedPlaceVisitor::new(self.thir,self.tcx);
   	visit::walk_expr(&mut visitor, expr);
   	if visitor.found{
   	match borrow_kind {
   	BorrowKind::Shallow | BorrowKind::Shared | BorrowKind::Unique
   	if !self.thir[arg].ty.is_freeze(self.tcx,self.param_env) =>
   	{
   	self.requires_unsafe(expr.span,BorrowOfLayoutConstrainedField)
   	}
   	BorrowKind::Mut{ .. } => {
   	self.requires_unsafe(expr.span,MutationOfLayoutConstrainedField)
   	}
   	BorrowKind::Shallow | BorrowKind::Shared | BorrowKind::Unique => {}
   	}
   	}
   	}

2. Same to 1. But I think we can't create a ref uniq binding anyway so it probably does not matter?
   

   rust/compiler/rustc_mir_build/src/check_unsafety.rs

   Lines 247 to 268 in e29821f

   	PatKind::Binding{mode:BindingMode::ByRef(borrow_kind), ty, .. } => {
   	ifself.inside_adt{
   	let ty::Ref(_, ty, _) = ty.kind()else{
   	span_bug!(
   	pat.span,
   	"BindingMode::ByRef in pattern, but found non-reference type {}",
   	ty
   	);
   	};
   	match borrow_kind {
   	BorrowKind::Shallow | BorrowKind::Shared | BorrowKind::Unique => {
   	if !ty.is_freeze(self.tcx,self.param_env){
   	self.requires_unsafe(pat.span,BorrowOfLayoutConstrainedField);
   	}
   	}
   	BorrowKind::Mut{ .. } => {
   	self.requires_unsafe(pat.span,MutationOfLayoutConstrainedField);
   	}
   	}
   	}
   	visit::walk_pat(self, pat);
   	}

3. Same to 1. but this is just for diagnostic I guess.
   

   rust/compiler/rustc_middle/src/mir/mod.rs

   Lines 2044 to 2050 in e29821f

   	// FIXME: won't be used after diagnostic migration
   	pubfndescribe_mutability(&self) -> &str{
   	match*self{
   	BorrowKind::Shared | BorrowKind::Shallow | BorrowKind::Unique => "immutable",
   	BorrowKind::Mut{ .. } => "mutable",
   	}
   	}

4. This one probably stay different just to be safe.
   

   rust/compiler/rustc_const_eval/src/transform/promote_consts.rs

   Lines 456 to 460 in e29821f

   	fnvalidate_ref(&mutself,kind:BorrowKind,place:&Place<'tcx>) -> Result<(),Unpromotable>{
   	match kind {
   	// Reject these borrow types just to be safe.
   	// FIXME(RalfJung): could we allow them? Should we? No point in it until we have a usecase.
   	BorrowKind::Shallow | BorrowKind::Unique => returnErr(Unpromotable),

5. This one should stay different since this is the motivation to introduce Unique.
   

   rust/compiler/rustc_borrowck/src/lib.rs

   Lines 1958 to 1969 in e29821f

   	match kind {
   	Reservation(WriteKind::MutableBorrow(
   	borrow_kind @ (BorrowKind::Unique | BorrowKind::Mut{ .. }),
   	))
   	| Write(WriteKind::MutableBorrow(
   	borrow_kind @ (BorrowKind::Unique | BorrowKind::Mut{ .. }),
   	)) => {
   	let is_local_mutation_allowed = match borrow_kind {
   	BorrowKind::Unique => LocalMutationIsAllowed::Yes,
   	BorrowKind::Mut{ .. } => is_local_mutation_allowed,
   	BorrowKind::Shared | BorrowKind::Shallow => unreachable!(),
   	};

[lcnr]

I am not sure whether specialcasing ClosureCapture borrows this much in compiler/rustc_borrowck/src/diagnostics/conflict_errors.rs is necessary or even particularily helpful, but 🤷

after dealing with the comments, r=me

thank you for working on this ❤️

[lcnr]

we should also run crater after the remaining changes as I expect the unsafeck and const checks to be theoretically breaking. Don't have the capacity to try and get an actual test for that though

[zirconium-n]

@rustbot label -S-waiting-on-author +S-waiting-on-review

we should also run crater after the remaining changes as I expect the unsafeck and const checks to be theoretically breaking. Don't have the capacity to try and get an actual test for that though

Const checks should be ok, since we forbid &mut in const so it's rejected anyway?
Not sure about unsafeck.

[lcnr]

@bors try

[bors]

⌛ Trying commit 9ce76438c3537c38c0614036add7cbe790cbdfd7 with merge d43518feaed04ed9d50bf2337350de8304ca58cf...

[lcnr]

@bors try

[bors]

⌛ Trying commit d93e93b3c821de824a888eae65a05f9687ad940c with merge ca6d8d30c65b4fcf60d8a40774045a936ec3c78d...

[bors]

☀️ Try build successful - checks-actions
Build commit: ca6d8d30c65b4fcf60d8a40774045a936ec3c78d (ca6d8d30c65b4fcf60d8a40774045a936ec3c78d)

[lcnr]

@craterbot check

[craterbot]

👌 Experiment pr-112119 created and queued.
🤖 Automatically detected try build ca6d8d30c65b4fcf60d8a40774045a936ec3c78d
🔍 You can check out the queue and this experiment's details.

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🚧 Experiment pr-112119 is now running

ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[craterbot]

🎉 Experiment pr-112119 is completed!
📊 16 regressed and 4 fixed (298526 total)
📰 Open the full report.

⚠️ If you notice any spurious failure please add them to the blacklist!
ℹ️ Crater is a tool to run experiments across parts of the Rust ecosystem. Learn more

[lcnr]

looked at a bunch and they all looked spurious, didn't get a nice testcase for them

r=me after rebase

[lcnr]

@bors r+ rollup=never

[bors]

📌 Commit b8a250f has been approved by lcnr

It is now in the queue for this repository.

[bors]

⌛ Testing commit b8a250f with merge c55d1ee...

[bors]

☀️ Test successful - checks-actions
Approved by: lcnr
Pushing c55d1ee to master...

[rust-timer]

Finished benchmarking commit (c55d1ee): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	2.3%	[2.3%, 2.3%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.0%	[3.0%, 3.0%]	1
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 657.959s -> 655.662s (-0.35%)