miri/const eval: support `MaybeDangling` by WaffleLapkin · Pull Request #150446 · rust-lang/rust

WaffleLapkin · 2025-12-27T21:10:03Z

View all comments

r? RalfJung

rustbot · 2025-12-27T21:10:07Z

Some changes occurred to the CTFE / Miri interpreter

cc @rust-lang/miri

Some changes occurred to the CTFE machinery

cc @RalfJung, @oli-obk, @lcnr

The Miri subtree was changed

cc @rust-lang/miri

compiler/rustc_const_eval/src/interpret/validity.rs

src/tools/miri/src/borrow_tracker/stacked_borrows/mod.rs

theemathas · 2025-12-28T08:47:06Z

Does this have insta-stable behavior change for ManuallyDrop in consteval?

RalfJung · 2025-12-28T16:24:14Z

It should only affect the behavior of code that still has UB until MaybeDangling is stabilized.

RalfJung · 2026-01-04T11:52:30Z

It is very import that we land #150447 before landing this, to avoid a situation where we generate LLVM IR with UB but Miri reports no UB.

src/tools/miri/src/borrow_tracker/stacked_borrows/mod.rs

src/tools/miri/tests/pass/move_manually_drop.rs

WaffleLapkin · 2026-02-26T16:29:50Z

@RalfJung I think I've addressed the review comments and this is ready for review too :)

(still blocked on the compiler change though)

RalfJung · 2026-02-26T22:26:32Z

(still blocked on the compiler change though)

By "the compiler change" you mean #150447 ?

WaffleLapkin · 2026-02-27T06:56:27Z

yes

rustbot · 2026-03-05T10:55:38Z

Some changes occurred in compiler/rustc_codegen_gcc

cc @antoyo, @GuillaumeGomez

Some changes occurred in compiler/rustc_codegen_cranelift

cc @bjorn3

Some changes occurred in tests/codegen-llvm/sanitizer

cc @rcvalle

RalfJung · 2026-03-05T16:06:02Z

Please rebase so it's easier to see the diff. :) Also, CI doesn't seem entirely happy yet.
@rustbot author

rustbot · 2026-03-05T16:06:06Z

Error: shortcut handler unexpectedly failed in this comment: failed to add labels

Please file an issue on GitHub at triagebot if there's a problem with this bot, or reach out on #triagebot on Zulip.

apiraino · 2026-03-05T16:11:27Z

(retry for debugging)

@rustbot author

Kobzol · 2026-03-05T16:52:06Z

(It was a 500 GitHub error)

rustbot · 2026-03-06T11:12:41Z

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

RalfJung

Looks good overall, thanks! All nits are minor.

View changes since this review

RalfJung · 2026-03-06T11:17:46Z

compiler/rustc_const_eval/src/interpret/validity.rs

    /// If this is `Some`, then `reset_provenance_and_padding` must be true (but not vice versa:
    /// we might not track data vs padding bytes if the operand isn't stored in memory anyway).
    data_bytes: Option<RangeSet>,
+    may_dangle: bool,


Please add a doc comment.

compiler/rustc_const_eval/src/interpret/validity.rs

RalfJung · 2026-03-06T11:19:37Z

src/tools/miri/tests/fail/validity/maybe_dangling_null.rs

This isn't even a borrow checking thing. Please move the test to tests/fail/validity and have it disable Stacked Borrows to ensure we don't rely on the aliasing model for catching this.

RalfJung · 2026-03-06T11:19:52Z

compiler/rustc_const_eval/src/interpret/validity.rs

                }
            }
+            ty::Adt(adt, _) if adt.is_maybe_dangling() => {
+                let could_dangle = mem::replace(&mut self.may_dangle, true);


Suggested change

let could_dangle = mem::replace(&mut self.may_dangle, true);

let old_may_dangle = mem::replace(&mut self.may_dangle, true);

RalfJung · 2026-03-06T11:20:13Z

compiler/rustc_const_eval/src/interpret/validity.rs

+                Ub(DanglingIntPointer { addr: 0, .. }) => NullPtr { ptr_kind, maybe: false },
+                Ub(DanglingIntPointer { addr: i, .. }) => DanglingPtrNoProvenance {
+                    ptr_kind,
+                    // FIXME this says "null pointer" when null but we need translate


Suggested change

// FIXME this says "null pointer" when null but we need translate

While we're at it

RalfJung · 2026-03-06T11:20:56Z

compiler/rustc_const_eval/src/interpret/validity.rs

This comment doesn't make sense any more since we're skipping that part when may_dangle is true.

RalfJung · 2026-03-06T11:21:51Z

compiler/rustc_const_eval/src/interpret/validity.rs

        // Make sure this is non-null. We checked dereferenceability above, but if `size` is zero
        // that does not imply non-null.


Suggested change

// Make sure this is non-null. We checked dereferenceability above, but if `size` is zero

// that does not imply non-null.

// Make sure this is non-null. This is obviously needed when `may_dangle` is set,

// but even if we did check dereferenceability above that would still allow null

// pointers if `size` is zero.

src/tools/miri/tests/fail/unaligned_pointers/maybe_dangling_unalighed.rs

RalfJung · 2026-03-06T11:24:20Z

src/tools/miri/tests/pass/both_borrows/maybe_dangling.rs

Given that this case was contentious, please also add a test like

// Under the current models, we do not forbid writing through // `MaybeDangling<&i32>`. That's not yet finally decided, but meanwhile // ensure we document this and notice when it changes. fn write_through_shr(x: MaybeDangling<&i32>) { let y: *mut i32 = transmute(x); y.write(1); } let mutref = &mut 0i32; write_through_shr(transmute(mutref));

WaffleLapkin · 2026-03-06T16:41:40Z

@RalfJung I think I addressed your nits, unless I missed something ^^'

RalfJung · 2026-03-06T21:11:26Z

src/tools/miri/tests/pass/both_borrows/maybe_dangling_shared_write.rs

I meant for this to just be a new test in the other file -- no reason to split it up.

RalfJung · 2026-03-06T21:12:42Z

compiler/rustc_const_eval/src/interpret/validity.rs

            self.check_wide_ptr_meta(place.meta(), place.layout)?;
        }
-        // Make sure this is dereferenceable and all.
+


Suggested change

// Determine size and alignment of pointee.

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. labels Dec 27, 2025

rustbot assigned RalfJung Dec 27, 2025

This comment has been minimized.

Sign in to view

WaffleLapkin force-pushed the miri-maybe-dangling branch from 96d2a1d to 72213c3 Compare December 27, 2025 21:22

This comment has been minimized.

Sign in to view

WaffleLapkin force-pushed the miri-maybe-dangling branch from 72213c3 to 0d9e008 Compare December 27, 2025 21:28

This comment has been minimized.

Sign in to view

WaffleLapkin force-pushed the miri-maybe-dangling branch from 0d9e008 to ae69218 Compare December 27, 2025 21:36

This comment has been minimized.

Sign in to view

theemathas reviewed Dec 28, 2025

View reviewed changes

compiler/rustc_const_eval/src/interpret/validity.rs Outdated Show resolved Hide resolved

theemathas reviewed Dec 28, 2025

View reviewed changes

src/tools/miri/src/borrow_tracker/stacked_borrows/mod.rs Outdated Show resolved Hide resolved

WaffleLapkin marked this pull request as draft December 28, 2025 13:42

rustbot 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 Dec 28, 2025

RalfJung added the S-blocked Status: Blocked on something else such as an RFC or other implementation work. label Jan 4, 2026

RalfJung reviewed Jan 4, 2026

View reviewed changes

src/tools/miri/src/borrow_tracker/stacked_borrows/mod.rs Show resolved Hide resolved

RalfJung reviewed Jan 4, 2026

View reviewed changes

src/tools/miri/src/borrow_tracker/stacked_borrows/mod.rs Show resolved Hide resolved

RalfJung reviewed Jan 4, 2026

View reviewed changes

src/tools/miri/tests/pass/move_manually_drop.rs Outdated Show resolved Hide resolved

WaffleLapkin force-pushed the miri-maybe-dangling branch 2 times, most recently from e5b7d00 to 959b31b Compare February 26, 2026 14:28

WaffleLapkin marked this pull request as ready for review February 26, 2026 16:28

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Feb 26, 2026

This comment has been minimized.

Sign in to view

RalfJung mentioned this pull request Mar 2, 2026

Implement MaybeDangling compiler support #150447

Merged

WaffleLapkin force-pushed the miri-maybe-dangling branch from 959b31b to ac559e4 Compare March 5, 2026 10:55

rustbot added A-LLVM Area: Code generation parts specific to LLVM. Both correctness bugs and optimization-related issues. PG-exploit-mitigations Project group: Exploit mitigations labels Mar 5, 2026

This comment has been minimized.

Sign in to view

WaffleLapkin force-pushed the miri-maybe-dangling branch from ac559e4 to d14294e Compare March 5, 2026 11:20

This comment has been minimized.

Sign in to view

rustbot removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. S-blocked Status: Blocked on something else such as an RFC or other implementation work. labels Mar 5, 2026

RalfJung added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Mar 5, 2026

WaffleLapkin force-pushed the miri-maybe-dangling branch from d14294e to 955b127 Compare March 6, 2026 11:12

RalfJung requested changes Mar 6, 2026

View reviewed changes

This comment has been minimized.

Sign in to view

WaffleLapkin force-pushed the miri-maybe-dangling branch 2 times, most recently from c052e2f to 7b01b7c Compare March 6, 2026 15:33

miri/const eval: support MaybeDangling

4a26802

WaffleLapkin force-pushed the miri-maybe-dangling branch from 7b01b7c to 4a26802 Compare March 6, 2026 16:38

RalfJung reviewed Mar 6, 2026

View reviewed changes

	let could_dangle = mem::replace(&mut self.may_dangle, true);
	let old_may_dangle = mem::replace(&mut self.may_dangle, true);

		// Make sure this is non-null. We checked dereferenceability above, but if `size` is zero
		// that does not imply non-null.

-        // Make sure this is non-null. We checked dereferenceability above, but if `size` is zero
-        // that does not imply non-null.
+// Make sure this is non-null. This is obviously needed when `may_dangle` is set,
+// but even if we did check dereferenceability above that would still allow null
+// pointers if `size` is zero.

Uh oh!

Conversation

WaffleLapkin commented Dec 27, 2025 • edited by rustbot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Dec 27, 2025

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

Uh oh!

Uh oh!

theemathas commented Dec 28, 2025

Uh oh!

RalfJung commented Dec 28, 2025

Uh oh!

RalfJung commented Jan 4, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

WaffleLapkin commented Feb 26, 2026

Uh oh!

RalfJung commented Feb 26, 2026

Uh oh!

WaffleLapkin commented Feb 27, 2026

Uh oh!

This comment has been minimized.

rustbot commented Mar 5, 2026

Uh oh!

This comment has been minimized.

This comment has been minimized.

RalfJung commented Mar 5, 2026

Uh oh!

rustbot commented Mar 5, 2026

Uh oh!

apiraino commented Mar 5, 2026

Uh oh!

Kobzol commented Mar 5, 2026

Uh oh!

rustbot commented Mar 6, 2026

Uh oh!

RalfJung 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!

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!

Uh oh!

RalfJung Mar 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

WaffleLapkin commented Mar 6, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

WaffleLapkin commented Dec 27, 2025 •

edited by rustbot

Loading

RalfJung left a comment •

edited by rustbot

Loading

RalfJung Mar 6, 2026 •

edited

Loading