Add an intrinsic for ptr::metadata

Author: scottmcm

compiler/rustc_borrowck/src/type_check/mod.rs

@@ -2398,11 +2398,11 @@ impl<'a, 'tcx> TypeChecker<'a, 'tcx> {
                             }
                         }
                     }
-                    CastKind::Transmute => {
+                    CastKind::PtrToMetadata | CastKind::Transmute => {
                         span_mirbug!(
                             self,
                             rvalue,
-                            "Unexpected CastKind::Transmute, which is not permitted in Analysis MIR",
+                            "Unexpected CastKind::{cast_kind:?}, which is not permitted in Analysis MIR",
                         );
                     }
                 }

compiler/rustc_codegen_cranelift/src/base.rs

@@ -698,6 +698,17 @@ fn codegen_stmt<'tcx>(
                         lval.write_cvalue(fx, CValue::by_val(res, dest_layout));
                     }
                 }
+                Rvalue::Cast(CastKind::PtrToMetadata, ref operand, _to_ty) => {
+                    let operand = codegen_operand(fx, operand);
+                    let meta = match operand.layout().abi {
+                        Abi::Scalar(_) => CValue::zst(dest_layout),
+                        Abi::ScalarPair(_, _) => {
+                            CValue::by_val(operand.load_scalar_pair(fx).1, dest_layout)
+                        }
+                        _ => bug!("Unexpected `PtrToMetadata` operand: {operand:?}"),
+                    };
+                    lval.write_cvalue(fx, meta);
+                }
                 Rvalue::Cast(
                     CastKind::PointerCoercion(PointerCoercion::ClosureFnPointer(_)),
                     ref operand,

compiler/rustc_codegen_cranelift/src/constant.rs

@@ -100,7 +100,7 @@ pub(crate) fn codegen_const_value<'tcx>(
     assert!(layout.is_sized(), "unsized const value");
 
     if layout.is_zst() {
-        return CValue::by_ref(crate::Pointer::dangling(layout.align.pref), layout);
+        return CValue::zst(layout);
     }
 
     match const_val {

compiler/rustc_codegen_cranelift/src/value_and_place.rs

@@ -111,6 +111,14 @@ impl<'tcx> CValue<'tcx> {
         CValue(inner, layout)
     }
 
+    /// Create an instance of a ZST
+    ///
+    /// The is represented by a dangling pointer of suitable alignment.
+    pub(crate) fn zst(layout: TyAndLayout<'tcx>) -> CValue<'tcx> {
+        assert!(layout.is_zst());
+        CValue::by_ref(crate::Pointer::dangling(layout.align.pref), layout)
+    }
+
     pub(crate) fn layout(&self) -> TyAndLayout<'tcx> {
         self.1
     }

compiler/rustc_codegen_ssa/src/mir/rvalue.rs

@@ -565,6 +565,13 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
                         };
                         OperandValue::Immediate(newval)
                     }
+                    mir::CastKind::PtrToMetadata => {
+                        match operand.val {
+                            OperandValue::Immediate(_) => OperandValue::ZeroSized,
+                            OperandValue::Pair(_, meta) => OperandValue::Immediate(meta),
+                            _ => bug!("Unexpected operand to `PtrToMetadata`: {operand:?}"),
+                        }
+                    }
                     mir::CastKind::Transmute => {
                         self.codegen_transmute_operand(bx, operand, cast).unwrap_or_else(|| {
                             bug!("Unsupported transmute-as-operand of {operand:?} to {cast:?}");

compiler/rustc_const_eval/src/interpret/cast.rs

@@ -64,6 +64,12 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
                 self.write_immediate(*res, dest)?;
             }
 
+            CastKind::PtrToMetadata => {
+                let src = self.read_immediate(src)?;
+                let res = self.ptr_to_metadata(&src, cast_layout)?;
+                self.write_immediate(*res, dest)?;
+            }
+
             CastKind::PointerCoercion(
                 PointerCoercion::MutToConstPointer | PointerCoercion::ArrayToPointer,
             ) => {
@@ -204,7 +210,7 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
         assert!(cast_to.ty.is_unsafe_ptr());
         // Handle casting any ptr to raw ptr (might be a fat ptr).
         if cast_to.size == src.layout.size {
-            // Thin or fat pointer that just hast the ptr kind of target type changed.
+            // Thin or fat pointer that just has the ptr kind of target type changed.
             return Ok(ImmTy::from_immediate(**src, cast_to));
         } else {
             // Casting the metadata away from a fat ptr.
@@ -225,6 +231,22 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
         }
     }
 
+    fn ptr_to_metadata(
+        &self,
+        src: &ImmTy<'tcx, M::Provenance>,
+        cast_to: TyAndLayout<'tcx>,
+    ) -> InterpResult<'tcx, ImmTy<'tcx, M::Provenance>> {
+        assert!(src.layout.ty.is_unsafe_ptr());
+        return Ok(match **src {
+            Immediate::Scalar(_) => {
+                assert!(cast_to.is_zst());
+                ImmTy::uninit(cast_to)
+            }
+            Immediate::ScalarPair(_, meta) => ImmTy::from_scalar(meta, cast_to),
+            Immediate::Uninit => throw_ub!(InvalidUninitBytes(None)),
+        });
+    }
+
     pub fn pointer_expose_provenance_cast(
         &mut self,
         src: &ImmTy<'tcx, M::Provenance>,

compiler/rustc_const_eval/src/transform/validate.rs

@@ -1135,6 +1135,23 @@ impl<'a, 'tcx> Visitor<'tcx> for TypeChecker<'a, 'tcx> {
                             self.fail(location, "Can't cast {op_ty} into 'Ptr'");
                         }
                     }
+                    CastKind::PtrToMetadata => {
+                        if !op_ty.is_unsafe_ptr() {
+                            self.fail(location, "Can't get ptr metadata from {op_ty}");
+                        }
+                        let pointee_ty = op_ty.builtin_deref(true).unwrap().ty;
+
+                        // FIXME: Check metadata more generally
+                        if pointee_ty.is_slice() {
+                            if !self.mir_assign_valid_types(*target_type, self.tcx.types.usize) {
+                                self.fail(location, "slice metadata must be usize");
+                            }
+                        } else if pointee_ty.is_sized(self.tcx, self.param_env) {
+                            if *target_type != self.tcx.types.unit {
+                                self.fail(location, "metadata for pointer-to-thin must be unit");
+                            }
+                        }
+                    }
                     CastKind::FloatToFloat | CastKind::FloatToInt => {
                         if !op_ty.is_floating_point() || !target_type.is_numeric() {
                             self.fail(

compiler/rustc_hir_analysis/src/check/intrinsic.rs

@@ -129,6 +129,7 @@ pub fn intrinsic_operation_unsafety(tcx: TyCtxt<'_>, intrinsic_id: LocalDefId) -
         | sym::is_val_statically_known
         | sym::ptr_mask
         | sym::aggregate_raw_ptr
+        | sym::ptr_metadata
         | sym::ub_checks
         | sym::fadd_algebraic
         | sym::fsub_algebraic
@@ -578,6 +579,7 @@ pub fn check_intrinsic_type(
             // This type check is not particularly useful, but the `where` bounds
             // on the definition in `core` do the heavy lifting for checking it.
             sym::aggregate_raw_ptr => (3, 1, vec![param(1), param(2)], param(0)),
+            sym::ptr_metadata => (2, 1, vec![Ty::new_imm_ptr(tcx, param(0))], param(1)),
 
             sym::ub_checks => (0, 1, Vec::new(), tcx.types.bool),
 

compiler/rustc_middle/src/mir/statement.rs

@@ -425,6 +425,7 @@ impl<'tcx> Rvalue<'tcx> {
                 | CastKind::IntToFloat
                 | CastKind::FnPtrToPtr
                 | CastKind::PtrToPtr
+                | CastKind::PtrToMetadata
                 | CastKind::PointerCoercion(_)
                 | CastKind::PointerWithExposedProvenance
                 | CastKind::DynStar

compiler/rustc_middle/src/mir/syntax.rs

@@ -1325,6 +1325,17 @@ pub enum CastKind {
     IntToFloat,
     PtrToPtr,
     FnPtrToPtr,
+    /// Convert a raw pointer to an `impl Pointee<Metadata = M>` into `M`.
+    ///
+    /// For example, this will give a `()` from `*const i32`, a `usize` from
+    /// `*mut [u8]`, or a vtable from a `*const dyn Foo`.
+    ///
+    /// There's only one legal cast type based on the input type, but calculating
+    /// that type is expensive, and thus it's convenient for this to be a `Cast`
+    /// instead of an `UnOp`.
+    ///
+    /// Allowed only in [`MirPhase::Runtime`]; Earlier it's an intrinsic.
+    PtrToMetadata,
     /// Reinterpret the bits of the input as a different type.
     ///
     /// MIR is well-formed if the input and output types have different sizes,

compiler/rustc_mir_transform/src/lower_intrinsics.rs

@@ -315,6 +315,24 @@ impl<'tcx> MirPass<'tcx> for LowerIntrinsics {
 
                         terminator.kind = TerminatorKind::Goto { target };
                     }
+                    sym::ptr_metadata => {
+                        let Ok([ptr]) = <[_; 1]>::try_from(std::mem::take(args)) else {
+                            span_bug!(
+                                terminator.source_info.span,
+                                "Wrong number of arguments for ptr_metadata intrinsic",
+                            );
+                        };
+                        let target = target.unwrap();
+                        let metadata_ty = generic_args.type_at(1);
+                        block.statements.push(Statement {
+                            source_info: terminator.source_info,
+                            kind: StatementKind::Assign(Box::new((
+                                *destination,
+                                Rvalue::Cast(CastKind::PtrToMetadata, ptr.node, metadata_ty),
+                            ))),
+                        });
+                        terminator.kind = TerminatorKind::Goto { target };
+                    }
                     _ => {}
                 }
             }

compiler/rustc_smir/src/rustc_smir/convert/mir.rs

@@ -276,6 +276,7 @@ impl<'tcx> Stable<'tcx> for mir::CastKind {
             FloatToFloat => stable_mir::mir::CastKind::FloatToFloat,
             IntToFloat => stable_mir::mir::CastKind::IntToFloat,
             PtrToPtr => stable_mir::mir::CastKind::PtrToPtr,
+            PtrToMetadata => stable_mir::mir::CastKind::PtrToMetadata,
             FnPtrToPtr => stable_mir::mir::CastKind::FnPtrToPtr,
             Transmute => stable_mir::mir::CastKind::Transmute,
         }

compiler/rustc_span/src/symbol.rs

@@ -1417,6 +1417,7 @@ symbols! {
         ptr_guaranteed_cmp,
         ptr_is_null,
         ptr_mask,
+        ptr_metadata,
         ptr_null,
         ptr_null_mut,
         ptr_offset_from,

compiler/stable_mir/src/mir/body.rs

@@ -949,6 +949,7 @@ pub enum CastKind {
     FloatToFloat,
     IntToFloat,
     PtrToPtr,
+    PtrToMetadata,
     FnPtrToPtr,
     Transmute,
 }

library/core/src/intrinsics.rs

@@ -2807,6 +2807,21 @@ impl<P: ?Sized, T: ptr::Thin> AggregateRawPtr<*mut T> for *mut P {
     type Metadata = <P as ptr::Pointee>::Metadata;
 }
 
+/// Lowers in MIR to `Rvalue::Cast` with `CastKind::PtrToMetadata`.
+///
+/// This is used to implement functions like `ptr::metadata`.
+#[rustc_nounwind]
+#[unstable(feature = "core_intrinsics", issue = "none")]
+#[rustc_const_unstable(feature = "ptr_metadata", issue = "81513")]
+#[rustc_intrinsic]
+#[rustc_intrinsic_must_be_overridden]
+#[cfg(not(bootstrap))]
+pub const fn ptr_metadata<P: ptr::Pointee<Metadata = M> + ?Sized, M>(_ptr: *const P) -> M {
+    // To implement a fallback we'd have to assume the layout of the pointer,
+    // but the whole point of this intrinsic is that we shouldn't do that.
+    unreachable!()
+}
+
 // 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

library/core/src/ptr/metadata.rs

@@ -3,7 +3,7 @@
 use crate::fmt;
 use crate::hash::{Hash, Hasher};
 #[cfg(not(bootstrap))]
-use crate::intrinsics::aggregate_raw_ptr;
+use crate::intrinsics::{aggregate_raw_ptr, ptr_metadata};
 use crate::marker::Freeze;
 
 /// Provides the pointer metadata type of any pointed-to type.
@@ -95,10 +95,17 @@ pub trait Thin = Pointee<Metadata = ()>;
 #[rustc_const_unstable(feature = "ptr_metadata", issue = "81513")]
 #[inline]
 pub const fn metadata<T: ?Sized>(ptr: *const T) -> <T as Pointee>::Metadata {
-    // SAFETY: Accessing the value from the `PtrRepr` union is safe since *const T
-    // and PtrComponents<T> have the same memory layouts. Only std can make this
-    // guarantee.
-    unsafe { PtrRepr { const_ptr: ptr }.components.metadata }
+    #[cfg(bootstrap)]
+    {
+        // SAFETY: Accessing the value from the `PtrRepr` union is safe since *const T
+        // and PtrComponents<T> have the same memory layouts. Only std can make this
+        // guarantee.
+        unsafe { PtrRepr { const_ptr: ptr }.components.metadata }
+    }
+    #[cfg(not(bootstrap))]
+    {
+        ptr_metadata(ptr)
+    }
 }
 
 /// Forms a (possibly-wide) raw pointer from a data pointer and metadata.
@@ -153,22 +160,26 @@ pub const fn from_raw_parts_mut<T: ?Sized>(
 }
 
 #[repr(C)]
+#[cfg(bootstrap)]
 union PtrRepr<T: ?Sized> {
     const_ptr: *const T,
     mut_ptr: *mut T,
     components: PtrComponents<T>,
 }
 
 #[repr(C)]
+#[cfg(bootstrap)]
 struct PtrComponents<T: ?Sized> {
     data_pointer: *const (),
     metadata: <T as Pointee>::Metadata,
 }
 
 // Manual impl needed to avoid `T: Copy` bound.
+#[cfg(bootstrap)]
 impl<T: ?Sized> Copy for PtrComponents<T> {}
 
 // Manual impl needed to avoid `T: Clone` bound.
+#[cfg(bootstrap)]
 impl<T: ?Sized> Clone for PtrComponents<T> {
     fn clone(&self) -> Self {
         *self

library/core/tests/ptr.rs

@@ -1171,3 +1171,11 @@ fn test_ptr_from_raw_parts_in_const() {
     assert_eq!(EMPTY_SLICE_PTR.addr(), 123);
     assert_eq!(EMPTY_SLICE_PTR.len(), 456);
 }
+
+#[test]
+fn test_ptr_metadata_in_const() {
+    const ARRAY_META: () = std::ptr::metadata::<[u16; 3]>(&[1, 2, 3]);
+    const SLICE_META: usize = std::ptr::metadata::<[u16]>(&[1, 2, 3]);
+    assert_eq!(ARRAY_META, ());
+    assert_eq!(SLICE_META, 3);
+}

tests/codegen/intrinsics/ptr_metadata.rs

@@ -0,0 +1,36 @@
+//@ compile-flags: -O -C no-prepopulate-passes -Z inline-mir
+//@ only-64bit (so I don't need to worry about usize)
+
+#![crate_type = "lib"]
+#![feature(core_intrinsics)]
+
+use std::intrinsics::ptr_metadata;
+
+// CHECK-LABEL: @thin_metadata(
+#[no_mangle]
+pub fn thin_metadata(p: *const ()) {
+    // CHECK: start
+    // CHECK-NEXT: ret void
+    ptr_metadata(p)
+}
+
+// CHECK-LABEL: @slice_metadata(
+#[no_mangle]
+pub fn slice_metadata(p: *const [u8]) -> usize {
+    // CHECK: start
+    // CHECK-NEXT: ret i64 %p.1
+    ptr_metadata(p)
+}
+
+// CHECK-LABEL: @dyn_byte_offset(
+#[no_mangle]
+pub unsafe fn dyn_byte_offset(
+    p: *const dyn std::fmt::Debug,
+    n: usize,
+) -> *const dyn std::fmt::Debug {
+    // CHECK: %[[Q:.+]] = getelementptr inbounds i8, ptr %p.0, i64 %n
+    // CHECK: %[[TEMP1:.+]] = insertvalue { ptr, ptr } poison, ptr %[[Q]], 0
+    // CHECK: %[[TEMP2:.+]] = insertvalue { ptr, ptr } %[[TEMP1]], ptr %p.1, 1
+    // CHECK: ret { ptr, ptr } %[[TEMP2]]
+    p.byte_add(n)
+}