miri: do not apply aliasing restrictions to Box with custom allocator

Author: RalfJung

compiler/rustc_const_eval/src/interpret/validity.rs

@@ -17,8 +17,8 @@ use rustc_middle::mir::interpret::{
     ExpectedKind, InterpError, InvalidMetaKind, Misalignment, PointerKind, Provenance,
     ValidationErrorInfo, ValidationErrorKind, ValidationErrorKind::*,
 };
-use rustc_middle::ty;
 use rustc_middle::ty::layout::{LayoutOf, TyAndLayout};
+use rustc_middle::ty::{self, Ty};
 use rustc_span::symbol::{sym, Symbol};
 use rustc_target::abi::{
     Abi, FieldIdx, Scalar as ScalarAbi, Size, VariantIdx, Variants, WrappingRange,
@@ -783,7 +783,11 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValueVisitor<'mir, 'tcx, M>
     }
 
     #[inline]
-    fn visit_box(&mut self, op: &OpTy<'tcx, M::Provenance>) -> InterpResult<'tcx> {
+    fn visit_box(
+        &mut self,
+        _box_ty: Ty<'tcx>,
+        op: &OpTy<'tcx, M::Provenance>,
+    ) -> InterpResult<'tcx> {
         self.check_safe_pointer(op, PointerKind::Box)?;
         Ok(())
     }

compiler/rustc_const_eval/src/interpret/visitor.rs

@@ -3,7 +3,7 @@
 
 use rustc_index::IndexVec;
 use rustc_middle::mir::interpret::InterpResult;
-use rustc_middle::ty;
+use rustc_middle::ty::{self, Ty};
 use rustc_target::abi::FieldIdx;
 use rustc_target::abi::{FieldsShape, VariantIdx, Variants};
 
@@ -47,10 +47,10 @@ pub trait ValueVisitor<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>>: Sized {
         Ok(())
     }
     /// Visits the given value as the pointer of a `Box`. There is nothing to recurse into.
-    /// The type of `v` will be a raw pointer, but this is a field of `Box<T>` and the
-    /// pointee type is the actual `T`.
+    /// The type of `v` will be a raw pointer to `T`, but this is a field of `Box<T>` and the
+    /// pointee type is the actual `T`. `box_ty` provides the full type of the `Box` itself.
     #[inline(always)]
-    fn visit_box(&mut self, _v: &Self::V) -> InterpResult<'tcx> {
+    fn visit_box(&mut self, _box_ty: Ty<'tcx>, _v: &Self::V) -> InterpResult<'tcx> {
         Ok(())
     }
 
@@ -144,7 +144,7 @@ pub trait ValueVisitor<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>>: Sized {
                 assert_eq!(nonnull_ptr.layout().fields.count(), 1);
                 let raw_ptr = self.ecx().project_field(&nonnull_ptr, 0)?; // the actual raw ptr
                 // ... whose only field finally is a raw ptr we can dereference.
-                self.visit_box(&raw_ptr)?;
+                self.visit_box(ty, &raw_ptr)?;
 
                 // The second `Box` field is the allocator, which we recursively check for validity
                 // like in regular structs.

compiler/rustc_hir_analysis/src/check/intrinsic.rs

@@ -657,6 +657,10 @@ pub fn check_intrinsic_type(
             sym::simd_shuffle => (3, 0, vec![param(0), param(0), param(1)], param(2)),
             sym::simd_shuffle_generic => (2, 1, vec![param(0), param(0)], param(1)),
 
+            sym::retag_box_to_raw => {
+                (2, 0, vec![Ty::new_mut_ptr(tcx, param(0))], Ty::new_mut_ptr(tcx, param(0)))
+            }
+
             other => {
                 tcx.dcx().emit_err(UnrecognizedIntrinsicFunction { span, name: other });
                 return;

compiler/rustc_middle/src/ty/sty.rs

@@ -2008,13 +2008,10 @@ impl<'tcx> Ty<'tcx> {
                     // Single-argument Box is always global. (for "minicore" tests)
                     return true;
                 };
-                if let Some(alloc_adt) = alloc.expect_ty().ty_adt_def() {
+                alloc.expect_ty().ty_adt_def().is_some_and(|alloc_adt| {
                     let global_alloc = tcx.require_lang_item(LangItem::GlobalAlloc, None);
                     alloc_adt.did() == global_alloc
-                } else {
-                    // Allocator is not an ADT...
-                    false
-                }
+                })
             }
             _ => false,
         }

compiler/rustc_span/src/symbol.rs

@@ -1450,6 +1450,7 @@ symbols! {
         residual,
         result,
         resume,
+        retag_box_to_raw,
         return_position_impl_trait_in_trait,
         return_type_notation,
         rhs,

library/alloc/src/boxed.rs

@@ -155,6 +155,7 @@ use core::error::Error;
 use core::fmt;
 use core::future::Future;
 use core::hash::{Hash, Hasher};
+use core::intrinsics::retag_box_to_raw;
 use core::iter::FusedIterator;
 use core::marker::Tuple;
 use core::marker::Unsize;
@@ -1110,8 +1111,16 @@ impl<T: ?Sized, A: Allocator> Box<T, A> {
     #[unstable(feature = "allocator_api", issue = "32838")]
     #[inline]
     pub fn into_raw_with_allocator(b: Self) -> (*mut T, A) {
-        let (leaked, alloc) = Box::into_unique(b);
-        (leaked.as_ptr(), alloc)
+        // This is the transition point from `Box` to raw pointers. For Stacked Borrows, these casts
+        // are relevant -- if this is a global allocator Box and we just get the pointer from `b.0`,
+        // it will have `Unique` permission, which is not what we want from a raw pointer. We could
+        // fix that by going through `&mut`, but then if this is *not* a global allocator Box, we'd
+        // be adding uniqueness assertions that we do not want. So for Miri's sake we pass this
+        // pointer through an intrinsic for box-to-raw casts, which can do the right thing wrt the
+        // aliasing model.
+        let b = mem::ManuallyDrop::new(b);
+        let alloc = unsafe { ptr::read(&b.1) };
+        (unsafe { retag_box_to_raw::<T, A>(b.0.as_ptr()) }, alloc)
     }
 
     #[unstable(
@@ -1122,13 +1131,8 @@ impl<T: ?Sized, A: Allocator> Box<T, A> {
     #[inline]
     #[doc(hidden)]
     pub fn into_unique(b: Self) -> (Unique<T>, A) {
-        // Box is recognized as a "unique pointer" by Stacked Borrows, but internally it is a
-        // raw pointer for the type system. Turning it directly into a raw pointer would not be
-        // recognized as "releasing" the unique pointer to permit aliased raw accesses,
-        // so all raw pointer methods have to go through `Box::leak`. Turning *that* to a raw pointer
-        // behaves correctly.
-        let alloc = unsafe { ptr::read(&b.1) };
-        (Unique::from(Box::leak(b)), alloc)
+        let (ptr, alloc) = Box::into_raw_with_allocator(b);
+        unsafe { (Unique::from(&mut *ptr), alloc) }
     }
 
     /// Returns a reference to the underlying allocator.
@@ -1184,7 +1188,7 @@ impl<T: ?Sized, A: Allocator> Box<T, A> {
     where
         A: 'a,
     {
-        unsafe { &mut *mem::ManuallyDrop::new(b).0.as_ptr() }
+        unsafe { &mut *Box::into_raw(b) }
     }
 
     /// Converts a `Box<T>` into a `Pin<Box<T>>`. If `T` does not implement [`Unpin`], then

library/core/src/intrinsics.rs

@@ -2695,6 +2695,19 @@ pub unsafe fn vtable_size(_ptr: *const ()) -> usize {
     unreachable!()
 }
 
+/// Retag a box pointer as part of casting it to a raw pointer. This is the `Box` equivalent of
+/// `(x: &mut T) as *mut T`. The input pointer must be the pointer of a `Box` (passed as raw pointer
+/// to avoid all questions around move semantics and custom allocators), and `A` must be the `Box`'s
+/// allocator.
+#[unstable(feature = "core_intrinsics", issue = "none")]
+#[rustc_nounwind]
+#[cfg_attr(not(bootstrap), rustc_intrinsic)]
+#[cfg_attr(bootstrap, inline)]
+pub unsafe fn retag_box_to_raw<T: ?Sized, A>(ptr: *mut T) -> *mut T {
+    // Miri needs to adjust the provenance, but for regular codegen this is not needed
+    ptr
+}
+
 // Some functions are defined here because they accidentally got made
 // available in this module on stable. See <https://github.com/rust-lang/rust/issues/15702>.
 // (`transmute` also falls into this category, but it cannot be wrapped due to the

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

@@ -5,7 +5,7 @@ use std::num::NonZero;
 use smallvec::SmallVec;
 
 use rustc_data_structures::fx::{FxHashMap, FxHashSet};
-use rustc_middle::mir::RetagKind;
+use rustc_middle::{mir::RetagKind, ty::Ty};
 use rustc_target::abi::Size;
 
 use crate::*;
@@ -291,6 +291,19 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriInterpCxExt<'mir, 'tcx> {
         }
     }
 
+    fn retag_box_to_raw(
+        &mut self,
+        val: &ImmTy<'tcx, Provenance>,
+        alloc_ty: Ty<'tcx>,
+    ) -> InterpResult<'tcx, ImmTy<'tcx, Provenance>> {
+        let this = self.eval_context_mut();
+        let method = this.machine.borrow_tracker.as_ref().unwrap().borrow().borrow_tracker_method;
+        match method {
+            BorrowTrackerMethod::StackedBorrows => this.sb_retag_box_to_raw(val, alloc_ty),
+            BorrowTrackerMethod::TreeBorrows => this.tb_retag_box_to_raw(val, alloc_ty),
+        }
+    }
+
     fn retag_place_contents(
         &mut self,
         kind: RetagKind,

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

@@ -865,6 +865,24 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriInterpCxExt<'mir, 'tcx> {
         this.sb_retag_reference(val, new_perm, RetagInfo { cause, in_field: false })
     }
 
+    fn sb_retag_box_to_raw(
+        &mut self,
+        val: &ImmTy<'tcx, Provenance>,
+        alloc_ty: Ty<'tcx>,
+    ) -> InterpResult<'tcx, ImmTy<'tcx, Provenance>> {
+        let this = self.eval_context_mut();
+        let is_global_alloc = alloc_ty.ty_adt_def().is_some_and(|adt| {
+            let global_alloc = this.tcx.require_lang_item(rustc_hir::LangItem::GlobalAlloc, None);
+            adt.did() == global_alloc
+        });
+        if is_global_alloc {
+            // Retag this as-if it was a mutable reference.
+            this.sb_retag_ptr_value(RetagKind::Raw, val)
+        } else {
+            Ok(val.clone())
+        }
+    }
+
     fn sb_retag_place_contents(
         &mut self,
         kind: RetagKind,
@@ -916,10 +934,18 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriInterpCxExt<'mir, 'tcx> {
                 self.ecx
             }
 
-            fn visit_box(&mut self, place: &PlaceTy<'tcx, Provenance>) -> InterpResult<'tcx> {
-                // Boxes get a weak protectors, since they may be deallocated.
-                let new_perm = NewPermission::from_box_ty(place.layout.ty, self.kind, self.ecx);
-                self.retag_ptr_inplace(place, new_perm)
+            fn visit_box(
+                &mut self,
+                box_ty: Ty<'tcx>,
+                place: &PlaceTy<'tcx, Provenance>,
+            ) -> InterpResult<'tcx> {
+                // Only boxes for the global allocator get any special treatment.
+                if box_ty.is_box_global(*self.ecx.tcx) {
+                    // Boxes get a weak protectors, since they may be deallocated.
+                    let new_perm = NewPermission::from_box_ty(place.layout.ty, self.kind, self.ecx);
+                    self.retag_ptr_inplace(place, new_perm)?;
+                }
+                Ok(())
             }
 
             fn visit_value(&mut self, place: &PlaceTy<'tcx, Provenance>) -> InterpResult<'tcx> {

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

@@ -392,6 +392,15 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriInterpCxExt<'mir, 'tcx> {
         }
     }
 
+    fn tb_retag_box_to_raw(
+        &mut self,
+        val: &ImmTy<'tcx, Provenance>,
+        _alloc_ty: Ty<'tcx>,
+    ) -> InterpResult<'tcx, ImmTy<'tcx, Provenance>> {
+        // Casts to raw pointers are NOPs in Tree Borrows.
+        Ok(val.clone())
+    }
+
     /// Retag all pointers that are stored in this place.
     fn tb_retag_place_contents(
         &mut self,
@@ -441,14 +450,22 @@ pub trait EvalContextExt<'mir, 'tcx: 'mir>: crate::MiriInterpCxExt<'mir, 'tcx> {
             /// Regardless of how `Unique` is handled, Boxes are always reborrowed.
             /// When `Unique` is also reborrowed, then it behaves exactly like `Box`
             /// except for the fact that `Box` has a non-zero-sized reborrow.
-            fn visit_box(&mut self, place: &PlaceTy<'tcx, Provenance>) -> InterpResult<'tcx> {
-                let new_perm = NewPermission::from_unique_ty(
-                    place.layout.ty,
-                    self.kind,
-                    self.ecx,
-                    /* zero_size */ false,
-                );
-                self.retag_ptr_inplace(place, new_perm)
+            fn visit_box(
+                &mut self,
+                box_ty: Ty<'tcx>,
+                place: &PlaceTy<'tcx, Provenance>,
+            ) -> InterpResult<'tcx> {
+                // Only boxes for the global allocator get any special treatment.
+                if box_ty.is_box_global(*self.ecx.tcx) {
+                    let new_perm = NewPermission::from_unique_ty(
+                        place.layout.ty,
+                        self.kind,
+                        self.ecx,
+                        /* zero_size */ false,
+                    );
+                    self.retag_ptr_inplace(place, new_perm)?;
+                }
+                Ok(())
             }
 
             fn visit_value(&mut self, place: &PlaceTy<'tcx, Provenance>) -> InterpResult<'tcx> {