Auto merge of #109035 - scottmcm:ptr-read-should-know-undef, r=WaffleLapkin,JakobDegen

Ensure `ptr::read` gets all the same LLVM `load` metadata that dereferencing does

I was looking into `array::IntoIter` optimization, and noticed that it wasn't annotating the loads with `noundef` for simple things like `array::IntoIter<i32, N>`.  Trying to narrow it down, it seems that was because `MaybeUninit::assume_init_read` isn't marking the load as initialized (<https://rust.godbolt.org/z/Mxd8TPTnv>), which is unfortunate since that's basically its reason to exist.

The root cause is that `ptr::read` is currently implemented via the *untyped* `copy_nonoverlapping`, and thus the `load` doesn't get any type-aware metadata: no `noundef`, no `!range`.  This PR solves that by lowering `ptr::read(p)` to `copy *p` in MIR, for which the backends already do the right thing.

Fortuitiously, this also improves the IR we give to LLVM for things like `mem::replace`, and fixes a couple of long-standing bugs where `ptr::read` on `Copy` types was worse than `*`ing them.

Zulip conversation: <https://rust-lang.zulipchat.com/#narrow/stream/219381-t-libs/topic/Move.20array.3A.3AIntoIter.20to.20ManuallyDrop/near/341189936>

cc `@erikdesjardins` `@JakobDegen` `@workingjubilee` `@the8472`

Fixes #106369
Fixes #73258

4 files changed, 36 insertions, 3 deletions

diff --git a/compiler/rustc_hir_analysis/src/check/intrinsic.rs b/compiler/rustc_hir_analysis/src/check/intrinsic.rs
index 20b6561f8b2..8c364a4f3b2 100644
--- a/compiler/rustc_hir_analysis/src/check/intrinsic.rs
+++ b/compiler/rustc_hir_analysis/src/check/intrinsic.rs
@@ -363,6 +363,8 @@ pub fn check_intrinsic_type(tcx: TyCtxt<'_>, it: &hir::ForeignItem<'_>) {
             sym::likely => (0, vec![tcx.types.bool], tcx.types.bool),
             sym::unlikely => (0, vec![tcx.types.bool], tcx.types.bool),
 
+            sym::read_via_copy => (1, vec![tcx.mk_imm_ptr(param(0))], param(0)),
+
             sym::discriminant_value => {
                 let assoc_items = tcx.associated_item_def_ids(
                     tcx.require_lang_item(hir::LangItem::DiscriminantKind, None),
diff --git a/compiler/rustc_lint_defs/src/builtin.rs b/compiler/rustc_lint_defs/src/builtin.rs
index 46ec1a2dca1..91966e75b5f 100644
--- a/compiler/rustc_lint_defs/src/builtin.rs
+++ b/compiler/rustc_lint_defs/src/builtin.rs
@@ -1026,12 +1026,13 @@ declare_lint! {
     /// ### Example
     ///
     /// ```rust,compile_fail
-    /// #![feature(const_ptr_read)]
+    /// #![feature(const_mut_refs)]
     /// const FOO: () = unsafe {
     ///     let x = &[0_u8; 4];
     ///     let y = x.as_ptr().cast::<u32>();
-    ///     y.read(); // the address of a `u8` array is unknown and thus we don't know if
-    ///     // it is aligned enough for reading a `u32`.
+    ///     let mut z = 123;
+    ///     y.copy_to_nonoverlapping(&mut z, 1); // the address of a `u8` array is unknown
+    ///     // and thus we don't know if it is aligned enough for copying a `u32`.
     /// };
     /// ```
     ///
diff --git a/compiler/rustc_mir_transform/src/lower_intrinsics.rs b/compiler/rustc_mir_transform/src/lower_intrinsics.rs
index f596cc1808f..5d7382305ae 100644
--- a/compiler/rustc_mir_transform/src/lower_intrinsics.rs
+++ b/compiler/rustc_mir_transform/src/lower_intrinsics.rs
@@ -149,6 +149,35 @@ impl<'tcx> MirPass<'tcx> for LowerIntrinsics {
                             terminator.kind = TerminatorKind::Goto { target };
                         }
                     }
+                    sym::read_via_copy => {
+                        let [arg] = args.as_slice() else {
+                            span_bug!(terminator.source_info.span, "Wrong number of arguments");
+                        };
+                        let derefed_place =
+                            if let Some(place) = arg.place() && let Some(local) = place.as_local() {
+                                tcx.mk_place_deref(local.into())
+                            } else {
+                                span_bug!(terminator.source_info.span, "Only passing a local is supported");
+                            };
+                        terminator.kind = match *target {
+                            None => {
+                                // No target means this read something uninhabited,
+                                // so it must be unreachable, and we don't need to
+                                // preserve the assignment either.
+                                TerminatorKind::Unreachable
+                            }
+                            Some(target) => {
+                                block.statements.push(Statement {
+                                    source_info: terminator.source_info,
+                                    kind: StatementKind::Assign(Box::new((
+                                        *destination,
+                                        Rvalue::Use(Operand::Copy(derefed_place)),
+                                    ))),
+                                });
+                                TerminatorKind::Goto { target }
+                            }
+                        }
+                    }
                     sym::discriminant_value => {
                         if let (Some(target), Some(arg)) = (*target, args[0].place()) {
                             let arg = tcx.mk_place_deref(arg);
diff --git a/compiler/rustc_span/src/symbol.rs b/compiler/rustc_span/src/symbol.rs
index d8b5b25aaee..0154c719ef6 100644
--- a/compiler/rustc_span/src/symbol.rs
+++ b/compiler/rustc_span/src/symbol.rs
@@ -1153,6 +1153,7 @@ symbols! {
         read_enum_variant_arg,
         read_struct,
         read_struct_field,
+        read_via_copy,
         readonly,
         realloc,
         reason,