More GVN for PtrMetadata

`PtrMetadata` doesn't care about `*const`/`*mut`/`&`/`&mut`, so GVN away those casts in its argument. This includes updating MIR to allow calling PtrMetadata on references too, not just raw pointers. That means that `[T]::len` can be just `_0 = PtrMetadata(_1)`, for example. # Conflicts: # tests/mir-opt/pre-codegen/slice_index.slice_get_unchecked_mut_range.PreCodegen.after.panic-abort.mir # tests/mir-opt/pre-codegen/slice_index.slice_get_unchecked_mut_range.PreCodegen.after.panic-unwind.mir
rust-lang · Jun 21, 2024 · 4a7b6c0 · 4a7b6c0
1 parent 31d8696
commit 4a7b6c0
Show file tree

Hide file tree

Showing 14 changed files with 319 additions and 78 deletions.
diff --git a/compiler/rustc_codegen_ssa/src/mir/rvalue.rs b/compiler/rustc_codegen_ssa/src/mir/rvalue.rs
@@ -639,7 +639,9 @@ impl<'a, 'tcx, Bx: BuilderMethods<'a, 'tcx>> FunctionCx<'a, 'tcx, Bx> {
                         (OperandValue::Immediate(llval), operand.layout)
                     }
                     mir::UnOp::PtrMetadata => {
-                        debug_assert!(operand.layout.ty.is_unsafe_ptr());
+                        debug_assert!(
+                            operand.layout.ty.is_unsafe_ptr() || operand.layout.ty.is_ref(),
+                        );
                         let (_, meta) = operand.val.pointer_parts();
                         assert_eq!(operand.layout.fields.count() > 1, meta.is_some());
                         if let Some(meta) = meta {

diff --git a/compiler/rustc_const_eval/src/interpret/operator.rs b/compiler/rustc_const_eval/src/interpret/operator.rs
@@ -460,7 +460,7 @@ impl<'tcx, M: Machine<'tcx>> InterpCx<'tcx, M> {
                 let res = ScalarInt::truncate_from_uint(res, layout.size).0;
                 Ok(ImmTy::from_scalar(res.into(), layout))
             }
-            ty::RawPtr(..) => {
+            ty::RawPtr(..) | ty::Ref(..) => {
                 assert_eq!(un_op, PtrMetadata);
                 let (_, meta) = val.to_scalar_and_meta();
                 Ok(match meta {

diff --git a/compiler/rustc_middle/src/mir/syntax.rs b/compiler/rustc_middle/src/mir/syntax.rs
@@ -1446,10 +1446,12 @@ pub enum UnOp {
     Not,
     /// The `-` operator for negation
     Neg,
-    /// Get the metadata `M` from a `*const/mut impl Pointee<Metadata = M>`.
+    /// Gets the metadata `M` from a `*const`/`*mut`/`&`/`&mut` to
+    /// `impl Pointee<Metadata = M>`.
     ///
     /// For example, this will give a `()` from `*const i32`, a `usize` from
-    /// `*mut [u8]`, or a pointer to a vtable from a `*const dyn Foo`.
+    /// `&mut [u8]`, or a `ptr::DynMetadata<dyn Foo>` (internally a pointer)
+    /// from a `*mut dyn Foo`.
     ///
     /// Allowed only in [`MirPhase::Runtime`]; earlier it's an intrinsic.
     PtrMetadata,

diff --git a/compiler/rustc_mir_transform/src/gvn.rs b/compiler/rustc_mir_transform/src/gvn.rs
@@ -836,12 +836,8 @@ impl<'body, 'tcx> VnState<'body, 'tcx> {
                 }
                 Value::BinaryOp(op, lhs, rhs)
             }
-            Rvalue::UnaryOp(op, ref mut arg) => {
-                let arg = self.simplify_operand(arg, location)?;
-                if let Some(value) = self.simplify_unary(op, arg) {
-                    return Some(value);
-                }
-                Value::UnaryOp(op, arg)
+            Rvalue::UnaryOp(op, ref mut arg_op) => {
+                return self.simplify_unary(op, arg_op, location);
             }
             Rvalue::Discriminant(ref mut place) => {
                 let place = self.simplify_place_value(place, location)?;
@@ -971,8 +967,71 @@ impl<'body, 'tcx> VnState<'body, 'tcx> {
     }
 
     #[instrument(level = "trace", skip(self), ret)]
-    fn simplify_unary(&mut self, op: UnOp, value: VnIndex) -> Option<VnIndex> {
-        let value = match (op, self.get(value)) {
+    fn simplify_unary(
+        &mut self,
+        op: UnOp,
+        arg_op: &mut Operand<'tcx>,
+        location: Location,
+    ) -> Option<VnIndex> {
+        let mut arg_index = self.simplify_operand(arg_op, location)?;
+
+        // PtrMetadata doesn't care about *const vs *mut vs & vs &mut,
+        // so start by removing those distinctions so we can update the `Operand`
+        if op == UnOp::PtrMetadata {
+            let mut was_updated = false;
+            loop {
+                match self.get(arg_index) {
+                    // Pointer casts that preserve metadata, such as
+                    // `*const [i32]` <-> `*mut [i32]` <-> `*mut [f32]`.
+                    // It's critical that this not eliminate cases like
+                    // `*const [T]` -> `*const T` which remove metadata.
+                    // We run on potentially-generic MIR, though, so unlike codegen
+                    // we can't always know exactly what the metadata are.
+                    // Thankfully, equality on `ptr_metadata_ty_or_tail` gives us
+                    // what we need: `Ok(meta_ty)` if the metadata is known, or
+                    // `Err(tail_ty)` if not. Matching metadata is ok, but if
+                    // that's not known, then matching tail types is also ok,
+                    // allowing things like `*mut (?A, ?T)` <-> `*mut (?B, ?T)`.
+                    // FIXME: Would it be worth trying to normalize, rather than
+                    // passing the identity closure?  Or are the types in the
+                    // Cast realistically about as normalized as we can get anyway?
+                    Value::Cast { kind: CastKind::PtrToPtr, value: inner, from, to }
+                        if from
+                            .builtin_deref(true)
+                            .unwrap()
+                            .ptr_metadata_ty_or_tail(self.tcx, |t| t)
+                            == to
+                                .builtin_deref(true)
+                                .unwrap()
+                                .ptr_metadata_ty_or_tail(self.tcx, |t| t) =>
+                    {
+                        arg_index = *inner;
+                        was_updated = true;
+                        continue;
+                    }
+
+                    // `&mut *p`, `&raw *p`, etc don't change metadata.
+                    Value::Address { place, kind: _, provenance: _ }
+                        if let PlaceRef { local, projection: [PlaceElem::Deref] } =
+                            place.as_ref()
+                            && let Some(local_index) = self.locals[local] =>
+                    {
+                        arg_index = local_index;
+                        was_updated = true;
+                        continue;
+                    }
+
+                    _ => {
+                        if was_updated && let Some(op) = self.try_as_operand(arg_index, location) {
+                            *arg_op = op;
+                        }
+                        break;
+                    }
+                }
+            }
+        }
+
+        let value = match (op, self.get(arg_index)) {
             (UnOp::Not, Value::UnaryOp(UnOp::Not, inner)) => return Some(*inner),
             (UnOp::Neg, Value::UnaryOp(UnOp::Neg, inner)) => return Some(*inner),
             (UnOp::Not, Value::BinaryOp(BinOp::Eq, lhs, rhs)) => {
@@ -984,9 +1043,26 @@ impl<'body, 'tcx> VnState<'body, 'tcx> {
             (UnOp::PtrMetadata, Value::Aggregate(AggregateTy::RawPtr { .. }, _, fields)) => {
                 return Some(fields[1]);
             }
-            _ => return None,
+            // We have an unsizing cast, which assigns the length to fat pointer metadata.
+            (
+                UnOp::PtrMetadata,
+                Value::Cast {
+                    kind: CastKind::PointerCoercion(ty::adjustment::PointerCoercion::Unsize),
+                    from,
+                    to,
+                    ..
+                },
+            ) if let ty::Slice(..) = to.builtin_deref(true).unwrap().kind()
+                && let ty::Array(_, len) = from.builtin_deref(true).unwrap().kind() =>
+            {
+                return self.insert_constant(Const::from_ty_const(
+                    *len,
+                    self.tcx.types.usize,
+                    self.tcx,
+                ));
+            }
+            _ => Value::UnaryOp(op, arg_index),
         };
-
         Some(self.insert(value))
     }
 

diff --git a/compiler/rustc_mir_transform/src/validate.rs b/compiler/rustc_mir_transform/src/validate.rs
@@ -1116,12 +1116,17 @@ impl<'a, 'tcx> Visitor<'tcx> for TypeChecker<'a, 'tcx> {
                     UnOp::PtrMetadata => {
                         if !matches!(self.mir_phase, MirPhase::Runtime(_)) {
                             // It would probably be fine to support this in earlier phases,
-                            // but at the time of writing it's only ever introduced from intrinsic lowering,
+                            // but at the time of writing it's only ever introduced from intrinsic lowering
+                            // or other runtime-phase optimization passes,
                             // so earlier things can just `bug!` on it.
                             self.fail(location, "PtrMetadata should be in runtime MIR only");
                         }
 
-                        check_kinds!(a, "Cannot PtrMetadata non-pointer type {:?}", ty::RawPtr(..));
+                        check_kinds!(
+                            a,
+                            "Cannot PtrMetadata non-pointer non-reference type {:?}",
+                            ty::RawPtr(..) | ty::Ref(..)
+                        );
                     }
                 }
             }

diff --git a/tests/mir-opt/gvn.array_len.GVN.panic-abort.diff b/tests/mir-opt/gvn.array_len.GVN.panic-abort.diff
@@ -0,0 +1,31 @@
+- // MIR for `array_len` before GVN
++ // MIR for `array_len` after GVN
+
+  fn array_len(_1: &mut [i32; 42]) -> usize {
+      debug x => _1;
+      let mut _0: usize;
+      let _2: &[i32];
+      let mut _3: &[i32; 42];
+      let mut _4: *const [i32];
+      scope 1 {
+          debug x => _2;
+      }
+
+      bb0: {
+-         StorageLive(_2);
++         nop;
+          StorageLive(_3);
+          _3 = &(*_1);
+          _2 = move _3 as &[i32] (PointerCoercion(Unsize));
+          StorageDead(_3);
+          StorageLive(_4);
+          _4 = &raw const (*_2);
+-         _0 = PtrMetadata(move _4);
++         _0 = const 42_usize;
+          StorageDead(_4);
+-         StorageDead(_2);
++         nop;
+          return;
+      }
+  }
+
diff --git a/tests/mir-opt/gvn.array_len.GVN.panic-unwind.diff b/tests/mir-opt/gvn.array_len.GVN.panic-unwind.diff
@@ -0,0 +1,31 @@
+- // MIR for `array_len` before GVN
++ // MIR for `array_len` after GVN
+
+  fn array_len(_1: &mut [i32; 42]) -> usize {
+      debug x => _1;
+      let mut _0: usize;
+      let _2: &[i32];
+      let mut _3: &[i32; 42];
+      let mut _4: *const [i32];
+      scope 1 {
+          debug x => _2;
+      }
+
+      bb0: {
+-         StorageLive(_2);
++         nop;
+          StorageLive(_3);
+          _3 = &(*_1);
+          _2 = move _3 as &[i32] (PointerCoercion(Unsize));
+          StorageDead(_3);
+          StorageLive(_4);
+          _4 = &raw const (*_2);
+-         _0 = PtrMetadata(move _4);
++         _0 = const 42_usize;
+          StorageDead(_4);
+-         StorageDead(_2);
++         nop;
+          return;
+      }
+  }
+
diff --git a/tests/mir-opt/gvn.fn_pointers.GVN.panic-abort.diff b/tests/mir-opt/gvn.fn_pointers.GVN.panic-abort.diff
@@ -8,18 +8,18 @@
       let mut _3: fn(u8) -> u8;
       let _5: ();
       let mut _6: fn(u8) -> u8;
-      let mut _9: {closure@$DIR/gvn.rs:612:19: 612:21};
+      let mut _9: {closure@$DIR/gvn.rs:614:19: 614:21};
       let _10: ();
       let mut _11: fn();
-      let mut _13: {closure@$DIR/gvn.rs:612:19: 612:21};
+      let mut _13: {closure@$DIR/gvn.rs:614:19: 614:21};
       let _14: ();
       let mut _15: fn();
       scope 1 {
           debug f => _1;
           let _4: fn(u8) -> u8;
           scope 2 {
               debug g => _4;
-              let _7: {closure@$DIR/gvn.rs:612:19: 612:21};
+              let _7: {closure@$DIR/gvn.rs:614:19: 614:21};
               scope 3 {
                   debug closure => _7;
                   let _8: fn();
@@ -62,16 +62,16 @@
           StorageDead(_6);
           StorageDead(_5);
 -         StorageLive(_7);
--         _7 = {closure@$DIR/gvn.rs:612:19: 612:21};
+-         _7 = {closure@$DIR/gvn.rs:614:19: 614:21};
 -         StorageLive(_8);
 +         nop;
-+         _7 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21};
++         _7 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21};
 +         nop;
           StorageLive(_9);
 -         _9 = _7;
 -         _8 = move _9 as fn() (PointerCoercion(ClosureFnPointer(Safe)));
-+         _9 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21};
-+         _8 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21} as fn() (PointerCoercion(ClosureFnPointer(Safe)));
++         _9 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21};
++         _8 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21} as fn() (PointerCoercion(ClosureFnPointer(Safe)));
           StorageDead(_9);
           StorageLive(_10);
           StorageLive(_11);
@@ -88,8 +88,8 @@
           StorageLive(_13);
 -         _13 = _7;
 -         _12 = move _13 as fn() (PointerCoercion(ClosureFnPointer(Safe)));
-+         _13 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21};
-+         _12 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21} as fn() (PointerCoercion(ClosureFnPointer(Safe)));
++         _13 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21};
++         _12 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21} as fn() (PointerCoercion(ClosureFnPointer(Safe)));
           StorageDead(_13);
           StorageLive(_14);
           StorageLive(_15);

diff --git a/tests/mir-opt/gvn.fn_pointers.GVN.panic-unwind.diff b/tests/mir-opt/gvn.fn_pointers.GVN.panic-unwind.diff
@@ -8,18 +8,18 @@
       let mut _3: fn(u8) -> u8;
       let _5: ();
       let mut _6: fn(u8) -> u8;
-      let mut _9: {closure@$DIR/gvn.rs:612:19: 612:21};
+      let mut _9: {closure@$DIR/gvn.rs:614:19: 614:21};
       let _10: ();
       let mut _11: fn();
-      let mut _13: {closure@$DIR/gvn.rs:612:19: 612:21};
+      let mut _13: {closure@$DIR/gvn.rs:614:19: 614:21};
       let _14: ();
       let mut _15: fn();
       scope 1 {
           debug f => _1;
           let _4: fn(u8) -> u8;
           scope 2 {
               debug g => _4;
-              let _7: {closure@$DIR/gvn.rs:612:19: 612:21};
+              let _7: {closure@$DIR/gvn.rs:614:19: 614:21};
               scope 3 {
                   debug closure => _7;
                   let _8: fn();
@@ -62,16 +62,16 @@
           StorageDead(_6);
           StorageDead(_5);
 -         StorageLive(_7);
--         _7 = {closure@$DIR/gvn.rs:612:19: 612:21};
+-         _7 = {closure@$DIR/gvn.rs:614:19: 614:21};
 -         StorageLive(_8);
 +         nop;
-+         _7 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21};
++         _7 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21};
 +         nop;
           StorageLive(_9);
 -         _9 = _7;
 -         _8 = move _9 as fn() (PointerCoercion(ClosureFnPointer(Safe)));
-+         _9 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21};
-+         _8 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21} as fn() (PointerCoercion(ClosureFnPointer(Safe)));
++         _9 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21};
++         _8 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21} as fn() (PointerCoercion(ClosureFnPointer(Safe)));
           StorageDead(_9);
           StorageLive(_10);
           StorageLive(_11);
@@ -88,8 +88,8 @@
           StorageLive(_13);
 -         _13 = _7;
 -         _12 = move _13 as fn() (PointerCoercion(ClosureFnPointer(Safe)));
-+         _13 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21};
-+         _12 = const ZeroSized: {closure@$DIR/gvn.rs:612:19: 612:21} as fn() (PointerCoercion(ClosureFnPointer(Safe)));
++         _13 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21};
++         _12 = const ZeroSized: {closure@$DIR/gvn.rs:614:19: 614:21} as fn() (PointerCoercion(ClosureFnPointer(Safe)));
           StorageDead(_13);
           StorageLive(_14);
           StorageLive(_15);

diff --git a/tests/mir-opt/gvn.manual_slice_mut_len.GVN.panic-abort.diff b/tests/mir-opt/gvn.manual_slice_mut_len.GVN.panic-abort.diff
@@ -0,0 +1,41 @@
+- // MIR for `manual_slice_mut_len` before GVN
++ // MIR for `manual_slice_mut_len` after GVN
+
+  fn manual_slice_mut_len(_1: &mut [i32]) -> usize {
+      debug x => _1;
+      let mut _0: usize;
+      let _2: *mut [i32];
+      let mut _4: *mut [i32];
+      let mut _5: *const [i32];
+      scope 1 {
+          debug x => _2;
+          let _3: *const [i32];
+          scope 2 {
+              debug x => _3;
+          }
+      }
+
+      bb0: {
+-         StorageLive(_2);
++         nop;
+          _2 = &raw mut (*_1);
+-         StorageLive(_3);
++         nop;
+          StorageLive(_4);
+          _4 = _2;
+-         _3 = move _4 as *const [i32] (PtrToPtr);
++         _3 = _2 as *const [i32] (PtrToPtr);
+          StorageDead(_4);
+          StorageLive(_5);
+          _5 = _3;
+-         _0 = PtrMetadata(move _5);
++         _0 = PtrMetadata(_1);
+          StorageDead(_5);
+-         StorageDead(_3);
+-         StorageDead(_2);
++         nop;
++         nop;
+          return;
+      }
+  }
+