Give an AllocId to ConstValue::Slice.

Author: cjgillot

compiler/rustc_codegen_cranelift/src/constant.rs

@@ -196,8 +196,7 @@ pub(crate) fn codegen_const_value<'tcx>(
                 .offset_i64(fx, i64::try_from(offset.bytes()).unwrap()),
             layout,
         ),
-        ConstValue::Slice { data, meta } => {
-            let alloc_id = fx.tcx.reserve_and_set_memory_alloc(data);
+        ConstValue::Slice { alloc_id, meta, phantom: _ } => {
             let ptr = pointer_for_allocation(fx, alloc_id).get_addr(fx);
             let len = fx.bcx.ins().iconst(fx.pointer_type, meta as i64);
             CValue::by_val_pair(ptr, len, layout)

compiler/rustc_codegen_ssa/src/mir/operand.rs

@@ -100,14 +100,11 @@ impl<'a, 'tcx, V: CodegenObject> OperandRef<'tcx, V> {
                 OperandValue::Immediate(llval)
             }
             ConstValue::ZeroSized => return OperandRef::zero_sized(layout),
-            ConstValue::Slice { data, meta } => {
+            ConstValue::Slice { alloc_id, meta, phantom: _ } => {
                 let Abi::ScalarPair(a_scalar, _) = layout.abi else {
                     bug!("from_const: invalid ScalarPair layout: {:#?}", layout);
                 };
-                let a = Scalar::from_pointer(
-                    Pointer::new(bx.tcx().reserve_and_set_memory_alloc(data).into(), Size::ZERO),
-                    &bx.tcx(),
-                );
+                let a = Scalar::from_pointer(Pointer::new(alloc_id.into(), Size::ZERO), &bx.tcx());
                 let a_llval = bx.scalar_to_backend(
                     a,
                     a_scalar,

compiler/rustc_const_eval/src/const_eval/eval_queries.rs

@@ -210,10 +210,9 @@ pub(super) fn op_to_const<'tcx>(
                 // We know `offset` is relative to the allocation, so we can use `into_parts`.
                 let (prov, offset) = a.to_pointer(ecx).expect(msg).into_parts();
                 let alloc_id = prov.expect(msg).alloc_id();
-                let data = ecx.tcx.global_alloc(alloc_id).unwrap_memory();
                 assert!(offset == abi::Size::ZERO, "{}", msg);
                 let meta = b.to_target_usize(ecx).expect(msg);
-                ConstValue::Slice { data, meta }
+                ConstValue::Slice { alloc_id, meta, phantom: std::marker::PhantomData }
             }
             Immediate::Uninit => bug!("`Uninit` is not a valid value for {}", op.layout.ty),
         },

compiler/rustc_const_eval/src/interpret/intrinsics.rs

@@ -10,9 +10,7 @@ use rustc_middle::ty::{Ty, TyCtxt};
 use rustc_middle::{
     mir::{
         self,
-        interpret::{
-            Allocation, ConstAllocation, GlobalId, InterpResult, PointerArithmetic, Scalar,
-        },
+        interpret::{AllocId, GlobalId, InterpResult, PointerArithmetic, Scalar},
         BinOp, ConstValue, NonDivergingIntrinsic,
     },
     ty::layout::TyAndLayout,
@@ -28,10 +26,11 @@ use super::{
 use crate::fluent_generated as fluent;
 
 /// Directly returns an `Allocation` containing an absolute path representation of the given type.
-pub(crate) fn alloc_type_name<'tcx>(tcx: TyCtxt<'tcx>, ty: Ty<'tcx>) -> ConstAllocation<'tcx> {
+pub(crate) fn alloc_type_name<'tcx>(tcx: TyCtxt<'tcx>, ty: Ty<'tcx>) -> (AllocId, u64) {
     let path = crate::util::type_name(tcx, ty);
-    let alloc = Allocation::from_bytes_byte_aligned_immutable(path.into_bytes());
-    tcx.mk_const_alloc(alloc)
+    let bytes = path.into_bytes();
+    let len = bytes.len().try_into().unwrap();
+    (tcx.allocate_bytes(bytes), len)
 }
 
 /// The logic for all nullary intrinsics is implemented here. These intrinsics don't get evaluated
@@ -47,8 +46,8 @@ pub(crate) fn eval_nullary_intrinsic<'tcx>(
     Ok(match name {
         sym::type_name => {
             ensure_monomorphic_enough(tcx, tp_ty)?;
-            let alloc = alloc_type_name(tcx, tp_ty);
-            ConstValue::Slice { data: alloc, meta: alloc.inner().size().bytes() }
+            let (alloc_id, len) = alloc_type_name(tcx, tp_ty);
+            ConstValue::Slice { alloc_id, meta: len, phantom: std::marker::PhantomData }
         }
         sym::needs_drop => {
             ensure_monomorphic_enough(tcx, tp_ty)?;

compiler/rustc_const_eval/src/interpret/operand.rs

@@ -780,9 +780,8 @@ impl<'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> InterpCx<'mir, 'tcx, M> {
             }
             mir::ConstValue::Scalar(x) => adjust_scalar(x)?.into(),
             mir::ConstValue::ZeroSized => Immediate::Uninit,
-            mir::ConstValue::Slice { data, meta } => {
+            mir::ConstValue::Slice { alloc_id, meta, phantom: _ } => {
                 // This is const data, no mutation allowed.
-                let alloc_id = self.tcx.reserve_and_set_memory_alloc(data);
                 let ptr = Pointer::new(CtfeProvenance::from(alloc_id).as_immutable(), Size::ZERO);
                 Immediate::new_slice(self.global_base_pointer(ptr)?.into(), meta, self)
             }

compiler/rustc_middle/src/mir/consts.rs

@@ -5,7 +5,7 @@ use rustc_session::{config::RemapPathScopeComponents, RemapFileNameExt};
 use rustc_span::{Span, DUMMY_SP};
 use rustc_target::abi::{HasDataLayout, Size};
 
-use crate::mir::interpret::{alloc_range, AllocId, ConstAllocation, ErrorHandled, Scalar};
+use crate::mir::interpret::{alloc_range, AllocId, ErrorHandled, Scalar};
 use crate::mir::{pretty_print_const_value, Promoted};
 use crate::ty::print::with_no_trimmed_paths;
 use crate::ty::GenericArgsRef;
@@ -28,8 +28,8 @@ pub struct ConstAlloc<'tcx> {
 
 /// Represents a constant value in Rust. `Scalar` and `Slice` are optimizations for
 /// array length computations, enum discriminants and the pattern matching logic.
-#[derive(Copy, Clone, Debug, Eq, PartialEq, TyEncodable, TyDecodable, Hash)]
-#[derive(HashStable, Lift)]
+#[derive(Copy, Clone, Debug, Eq, PartialEq, TyEncodable, TyDecodable, Lift, Hash)]
+#[derive(HashStable)]
 pub enum ConstValue<'tcx> {
     /// Used for types with `layout::abi::Scalar` ABI.
     ///
@@ -48,10 +48,11 @@ pub enum ConstValue<'tcx> {
     Slice {
         /// The allocation storing the slice contents.
         /// This always points to the beginning of the allocation.
-        data: ConstAllocation<'tcx>,
+        alloc_id: AllocId,
         /// The metadata field of the reference.
         /// This is a "target usize", so we use `u64` as in the interpreter.
         meta: u64,
+        phantom: std::marker::PhantomData<&'tcx ()>,
     },
 
     /// A value not representable by the other variants; needs to be stored in-memory.
@@ -73,7 +74,7 @@ pub enum ConstValue<'tcx> {
 #[cfg(all(any(target_arch = "x86_64", target_arch = "aarch64"), target_pointer_width = "64"))]
 static_assert_size!(ConstValue<'_>, 24);
 
-impl<'tcx> ConstValue<'tcx> {
+impl ConstValue<'_> {
     #[inline]
     pub fn try_to_scalar(&self) -> Option<Scalar> {
         match *self {
@@ -94,11 +95,11 @@ impl<'tcx> ConstValue<'tcx> {
         self.try_to_scalar_int()?.try_into().ok()
     }
 
-    pub fn try_to_target_usize(&self, tcx: TyCtxt<'tcx>) -> Option<u64> {
+    pub fn try_to_target_usize(&self, tcx: TyCtxt<'_>) -> Option<u64> {
         self.try_to_scalar_int()?.try_to_target_usize(tcx).ok()
     }
 
-    pub fn try_to_bits_for_ty(
+    pub fn try_to_bits_for_ty<'tcx>(
         &self,
         tcx: TyCtxt<'tcx>,
         param_env: ty::ParamEnv<'tcx>,
@@ -125,12 +126,15 @@ impl<'tcx> ConstValue<'tcx> {
     }
 
     /// Must only be called on constants of type `&str` or `&[u8]`!
-    pub fn try_get_slice_bytes_for_diagnostics(&self, tcx: TyCtxt<'tcx>) -> Option<&'tcx [u8]> {
-        let (data, start, end) = match self {
+    pub fn try_get_slice_bytes_for_diagnostics<'tcx>(
+        &self,
+        tcx: TyCtxt<'tcx>,
+    ) -> Option<&'tcx [u8]> {
+        let (alloc_id, start, len) = match self {
             ConstValue::Scalar(_) | ConstValue::ZeroSized => {
                 bug!("`try_get_slice_bytes` on non-slice constant")
             }
-            &ConstValue::Slice { data, meta } => (data, 0, meta),
+            &ConstValue::Slice { alloc_id, meta, phantom: _ } => (alloc_id, 0, meta),
             &ConstValue::Indirect { alloc_id, offset } => {
                 // The reference itself is stored behind an indirection.
                 // Load the reference, and then load the actual slice contents.
@@ -162,26 +166,29 @@ impl<'tcx> ConstValue<'tcx> {
                 }
                 // Non-empty slice, must have memory. We know this is a relative pointer.
                 let (inner_prov, offset) = ptr.into_parts();
-                let data = tcx.global_alloc(inner_prov?.alloc_id()).unwrap_memory();
-                (data, offset.bytes(), offset.bytes() + len)
+                (inner_prov?.alloc_id(), offset.bytes(), offset.bytes() + len)
             }
         };
 
+        let data = tcx.global_alloc(alloc_id).unwrap_memory();
+
         // This is for diagnostics only, so we are okay to use `inspect_with_uninit_and_ptr_outside_interpreter`.
         let start = start.try_into().unwrap();
-        let end = end.try_into().unwrap();
+        let end = start + usize::try_from(len).unwrap();
         Some(data.inner().inspect_with_uninit_and_ptr_outside_interpreter(start..end))
     }
 
     /// Check if a constant may contain provenance information. This is used by MIR opts.
     /// Can return `true` even if there is no provenance.
-    pub fn may_have_provenance(&self, tcx: TyCtxt<'tcx>, size: Size) -> bool {
+    pub fn may_have_provenance(&self, tcx: TyCtxt<'_>, size: Size) -> bool {
         match *self {
             ConstValue::ZeroSized | ConstValue::Scalar(Scalar::Int(_)) => return false,
             ConstValue::Scalar(Scalar::Ptr(..)) => return true,
             // It's hard to find out the part of the allocation we point to;
             // just conservatively check everything.
-            ConstValue::Slice { data, meta: _ } => !data.inner().provenance().ptrs().is_empty(),
+            ConstValue::Slice { alloc_id, meta: _, phantom: _ } => {
+                !tcx.global_alloc(alloc_id).unwrap_memory().inner().provenance().ptrs().is_empty()
+            }
             ConstValue::Indirect { alloc_id, offset } => !tcx
                 .global_alloc(alloc_id)
                 .unwrap_memory()
@@ -424,9 +431,8 @@ impl<'tcx> Const<'tcx> {
     /// taking into account even pointer identity tests.
     pub fn is_deterministic(&self) -> bool {
         // Some constants may generate fresh allocations for pointers they contain,
-        // so using the same constant twice can yield two different results:
-        // - valtrees purposefully generate new allocations
-        // - ConstValue::Slice also generate new allocations
+        // so using the same constant twice can yield two different results.
+        // Notably, valtrees purposefully generate new allocations.
         match self {
             Const::Ty(c) => match c.kind() {
                 ty::ConstKind::Param(..) => true,
@@ -444,11 +450,11 @@ impl<'tcx> Const<'tcx> {
                 | ty::ConstKind::Placeholder(..) => bug!(),
             },
             Const::Unevaluated(..) => false,
-            // If the same slice appears twice in the MIR, we cannot guarantee that we will
-            // give the same `AllocId` to the data.
-            Const::Val(ConstValue::Slice { .. }, _) => false,
             Const::Val(
-                ConstValue::ZeroSized | ConstValue::Scalar(_) | ConstValue::Indirect { .. },
+                ConstValue::Slice { .. }
+                | ConstValue::ZeroSized
+                | ConstValue::Scalar(_)
+                | ConstValue::Indirect { .. },
                 _,
             ) => true,
         }

compiler/rustc_middle/src/mir/pretty.rs

@@ -1360,11 +1360,7 @@ pub fn write_allocations<'tcx>(
             }
             ConstValue::Scalar(interpret::Scalar::Int { .. }) => Either::Right(std::iter::empty()),
             ConstValue::ZeroSized => Either::Right(std::iter::empty()),
-            ConstValue::Slice { .. } => {
-                // `u8`/`str` slices, shouldn't contain pointers that we want to print.
-                Either::Right(std::iter::empty())
-            }
-            ConstValue::Indirect { alloc_id, .. } => {
+            ConstValue::Slice { alloc_id, .. } | ConstValue::Indirect { alloc_id, .. } => {
                 // FIXME: we don't actually want to print all of these, since some are printed nicely directly as values inline in MIR.
                 // Really we'd want `pretty_print_const_value` to decide which allocations to print, instead of having a separate visitor.
                 Either::Left(std::iter::once(alloc_id))

compiler/rustc_middle/src/ty/context.rs

@@ -72,7 +72,7 @@ use rustc_type_ir::TyKind::*;
 use rustc_type_ir::WithCachedTypeInfo;
 use rustc_type_ir::{CollectAndApply, Interner, TypeFlags};
 
-use std::borrow::Borrow;
+use std::borrow::{Borrow, Cow};
 use std::cmp::Ordering;
 use std::fmt;
 use std::hash::{Hash, Hasher};
@@ -866,7 +866,7 @@ impl<'tcx> TyCtxt<'tcx> {
     }
 
     /// Allocates a read-only byte or string literal for `mir::interpret`.
-    pub fn allocate_bytes(self, bytes: &[u8]) -> interpret::AllocId {
+    pub fn allocate_bytes<'a>(self, bytes: impl Into<Cow<'a, [u8]>>) -> interpret::AllocId {
         // Create an allocation that just contains these bytes.
         let alloc = interpret::Allocation::from_bytes_byte_aligned_immutable(bytes);
         let alloc = self.mk_const_alloc(alloc);

compiler/rustc_middle/src/ty/structural_impls.rs

@@ -16,6 +16,7 @@ use rustc_target::abi::TyAndLayout;
 use rustc_type_ir::{ConstKind, DebugWithInfcx, InferCtxtLike, WithInfcx};
 
 use std::fmt::{self, Debug};
+use std::marker::PhantomData;
 
 use super::print::PrettyPrinter;
 use super::{GenericArg, GenericArgKind, Region};
@@ -459,6 +460,13 @@ TrivialTypeTraversalAndLiftImpls! {
 ///////////////////////////////////////////////////////////////////////////
 // Lift implementations
 
+impl<'tcx> Lift<'tcx> for PhantomData<&()> {
+    type Lifted = PhantomData<&'tcx ()>;
+    fn lift_to_tcx(self, _: TyCtxt<'tcx>) -> Option<Self::Lifted> {
+        Some(PhantomData)
+    }
+}
+
 impl<'tcx, T: Lift<'tcx>> Lift<'tcx> for Option<T> {
     type Lifted = Option<T::Lifted>;
     fn lift_to_tcx(self, tcx: TyCtxt<'tcx>) -> Option<Self::Lifted> {

compiler/rustc_mir_build/src/build/expr/as_constant.rs

@@ -3,13 +3,14 @@
 use crate::build::{parse_float_into_constval, Builder};
 use rustc_ast as ast;
 use rustc_middle::mir;
-use rustc_middle::mir::interpret::{Allocation, LitToConstError, LitToConstInput, Scalar};
+use rustc_middle::mir::interpret::{LitToConstError, LitToConstInput, Scalar};
 use rustc_middle::mir::*;
 use rustc_middle::thir::*;
 use rustc_middle::ty::{
     self, CanonicalUserType, CanonicalUserTypeAnnotation, TyCtxt, UserTypeAnnotationIndex,
 };
 use rustc_target::abi::Size;
+use std::marker::PhantomData;
 
 impl<'a, 'tcx> Builder<'a, 'tcx> {
     /// Compile `expr`, yielding a compile-time constant. Assumes that
@@ -124,27 +125,39 @@ fn lit_to_mir_constant<'tcx>(
 
     let value = match (lit, &ty.kind()) {
         (ast::LitKind::Str(s, _), ty::Ref(_, inner_ty, _)) if inner_ty.is_str() => {
-            let s = s.as_str();
-            let allocation = Allocation::from_bytes_byte_aligned_immutable(s.as_bytes());
-            let allocation = tcx.mk_const_alloc(allocation);
-            ConstValue::Slice { data: allocation, meta: allocation.inner().size().bytes() }
+            let s = s.as_str().as_bytes();
+            let len = s.len();
+            let allocation = tcx.allocate_bytes(s);
+            ConstValue::Slice {
+                alloc_id: allocation,
+                meta: len.try_into().unwrap(),
+                phantom: PhantomData,
+            }
         }
         (ast::LitKind::ByteStr(data, _), ty::Ref(_, inner_ty, _))
             if matches!(inner_ty.kind(), ty::Slice(_)) =>
         {
-            let allocation = Allocation::from_bytes_byte_aligned_immutable(data as &[u8]);
-            let allocation = tcx.mk_const_alloc(allocation);
-            ConstValue::Slice { data: allocation, meta: allocation.inner().size().bytes() }
+            let len = data.len();
+            let allocation = tcx.allocate_bytes(&**data);
+            ConstValue::Slice {
+                alloc_id: allocation,
+                meta: len.try_into().unwrap(),
+                phantom: PhantomData,
+            }
         }
         (ast::LitKind::ByteStr(data, _), ty::Ref(_, inner_ty, _)) if inner_ty.is_array() => {
-            let id = tcx.allocate_bytes(data);
+            let id = tcx.allocate_bytes(&**data);
             ConstValue::Scalar(Scalar::from_pointer(id.into(), &tcx))
         }
         (ast::LitKind::CStr(data, _), ty::Ref(_, inner_ty, _)) if matches!(inner_ty.kind(), ty::Adt(def, _) if Some(def.did()) == tcx.lang_items().c_str()) =>
         {
-            let allocation = Allocation::from_bytes_byte_aligned_immutable(data as &[u8]);
-            let allocation = tcx.mk_const_alloc(allocation);
-            ConstValue::Slice { data: allocation, meta: allocation.inner().size().bytes() }
+            let len = data.len();
+            let allocation = tcx.allocate_bytes(&**data);
+            ConstValue::Slice {
+                alloc_id: allocation,
+                meta: len.try_into().unwrap(),
+                phantom: PhantomData,
+            }
         }
         (ast::LitKind::Byte(n), ty::Uint(ty::UintTy::U8)) => {
             ConstValue::Scalar(Scalar::from_uint(*n, Size::from_bytes(1)))