Semantic check of `mut` restrictions by CoCo-Japan-pan · Pull Request #159906 · rust-lang/rust

CoCo-Japan-pan · 2026-07-25T14:41:23Z

This PR implements semantic checks for mut restrictions proposed in RFC 3323, as part of the related GSoC project.
It lowers field restriction information from HIR into rustc_middle::ty::FieldDef and adds a MIR-level lint that rejects restricted field mutations and struct expressions. As parsing (#156824) and path resolution (#159125) have already been implemented, this PR provides a working implementation of mut restrictions.

Tracking issue: #105077
r? @Urgau
cc @jhpratt

rustbot · 2026-07-25T14:41:27Z

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

Urgau

Good start, mostly some tests related comments.

View changes since this review

Urgau · 2026-07-25T21:38:19Z

            if let hir::RestrictionKind::Restricted(path) = impl_restriction.kind {
-                ty::trait_def::ImplRestrictionKind::Restricted(path.res, impl_restriction.span)
+                ty::RestrictionKind::Restricted(path.res, impl_restriction.span)
            } else {
-                ty::trait_def::ImplRestrictionKind::Unrestricted
+                ty::RestrictionKind::Unrestricted
            },


Could you also make this else-if an match instead, as to be exhaustive.

Urgau · 2026-07-25T21:51:50Z

            // MIR-level lints.
            &Lint(check_inline::CheckForceInline),
            &Lint(check_call_recursion::CheckCallRecursion),
            &Lint(check_packed_ref::CheckPackedRef),
            &Lint(check_const_item_mutation::CheckConstItemMutation),
+            &Lint(check_mut_restriction::CheckMutRestriction),
            &Lint(function_item_references::FunctionItemReferences),


(Not for this PR) Most of those so called "MIR lints" actually emit hard errors of an lint. We should find another name as it's quite confusing to see a "lint" emit a hard errors.

Urgau · 2026-07-25T21:54:18Z

+struct MutRestrictionChecker<'a, 'tcx> {
+    body: &'a Body<'tcx>,
+    tcx: TyCtxt<'tcx>,
+    span: Span,


Suggested change

span: Span,

mutating_span: Span,

Urgau · 2026-07-25T21:59:12Z

+            let body_did = self.body.source.instance.def_id();
+
+            for (place_base, elem) in place.iter_projections() {
+                let ProjectionElem::Field(field_idx, _field_ty) = elem else {


Could you add some comment about why we are only checking for ProjectionElem::Field, I assume that for an index operation foo.bar[index] we always have a field operation.

Urgau · 2026-07-25T22:00:30Z

+                };
+
+                let base_ty = place_base.ty(self.body, self.tcx);
+                let ty::Adt(adt_def, _args) = *base_ty.ty.kind() else {


Same, please add a comment here for why only ty::Adt and why we ignore args.

Urgau · 2026-07-25T22:11:40Z

+//@ revisions: e2015 e2018
+//@ [e2015] edition: 2015
+//@ [e2018] edition: 2018..


Same here, no need for both revisions, let's stick 2018 an onward.

Urgau · 2026-07-25T22:15:03Z

+    //[e2018]~^ ERROR field cannot be mutated outside `crate::foo`
+    bar.2 = 1; // ok
+}
+


I think we should also check for inline assembly mutating use.

// for x86-64 asm!( "add {0}, 5", "mov {1}, 42", inout(reg) state.inout_field, // mutable inout_field out(reg) state.out_field, // mutable out_field );

Urgau · 2026-07-25T22:16:38Z

+        foo::bar::FooE::Var {
+            alpha, //[e2015]~ ERROR field cannot be mutated outside `foo::bar`
+            //[e2018]~^ ERROR field cannot be mutated outside `crate::foo::bar`
+            beta, //[e2015]~ ERROR field cannot be mutated outside `foo`
+            //[e2018]~^ ERROR field cannot be mutated outside `crate::foo`
+            gamma, // ok
+        } => {


What happens if I someone does alpha: _? Do we report an error?

Should probably also be a test.

Urgau · 2026-07-25T22:21:09Z

+        foo::bar::FooE::Tup(
+            alpha, //[e2015]~ ERROR field cannot be mutated outside `foo::bar`
+            //[e2018]~^ ERROR field cannot be mutated outside `crate::foo::bar`
+            beta, //[e2015]~ ERROR field cannot be mutated outside `foo`
+            //[e2018]~^ ERROR field cannot be mutated outside `crate::foo`
+            gamma, // ok
+        ) => {


We should also test what happens with an immutable borrowing:

&foo::bar::FooE::Tup(

Urgau · 2026-07-25T22:22:31Z

+fn change_fooe(foo: &mut foo::bar::FooE) {
+    match foo {
+        foo::bar::FooE::Default => {}
+        foo::bar::FooE::Var {
+            alpha, //[e2015]~ ERROR field cannot be mutated outside `foo::bar`
+            //[e2018]~^ ERROR field cannot be mutated outside `crate::foo::bar`
+            beta, //[e2015]~ ERROR field cannot be mutated outside `foo`
+            //[e2018]~^ ERROR field cannot be mutated outside `crate::foo`
+            gamma, // ok
+        } => {
+            *alpha = 1;
+            *beta = 1;
+            *gamma = 1;
+        }
+        foo::bar::FooE::Tup(
+            alpha, //[e2015]~ ERROR field cannot be mutated outside `foo::bar`
+            //[e2018]~^ ERROR field cannot be mutated outside `crate::foo::bar`
+            beta, //[e2015]~ ERROR field cannot be mutated outside `foo`
+            //[e2018]~^ ERROR field cannot be mutated outside `crate::foo`
+            gamma, // ok
+        ) => {
+            *alpha = 1;
+            *beta = 1;
+            *gamma = 1;
+        }
+    }
+}


@Nadrieril, can you think of another match related tests we should be testing regarding mut-restrictions?

Urgau · 2026-07-25T22:24:51Z

cc @cjgillot for the MIR related part since you know more about it than me

CoCo-Japan-pan added 6 commits July 25, 2026 17:23

Lower mut restriction to ty::FieldDef

af52084

bless thir-print

feac214

Emit error in mutation of restricted field

0cfc42a

Add UI tests for mutation of restricted fields

0e42a37

Emit error in struct expressions with restricted fields

f416665

Add UI tests for struct expressions with restricted fields

c2b967d

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. T-rustdoc Relevant to the rustdoc team, which will review and decide on the PR/issue. labels Jul 25, 2026

rustbot assigned Urgau Jul 25, 2026

rustbot added the T-rustdoc-frontend Relevant to the rustdoc-frontend team, which will review and decide on the web UI/UX output. label Jul 25, 2026

CoCo-Japan-pan changed the title ~~Mut restriction check~~ Semantic check of mut restrictions Jul 25, 2026

Urgau reviewed Jul 25, 2026

View reviewed changes

Urgau added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 25, 2026

Uh oh!

Uh oh!

Conversation

CoCo-Japan-pan commented Jul 25, 2026 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jul 25, 2026

Uh oh!

Urgau left a comment • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Urgau commented Jul 25, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

CoCo-Japan-pan commented Jul 25, 2026 •

edited by rustbot

Loading

Urgau left a comment •

edited by rustbot

Loading