Auto merge of #74059 - RalfJung:miri-uninit-validation, r=oli-obk

Miri value validation: fix handling of uninit memory

Fixes https://github.com/rust-lang/miri/issues/1456
Fixes https://github.com/rust-lang/miri/issues/1467

r? @oli-obk

6 files changed, 69 insertions, 28 deletions

diff --git a/src/librustc_mir/interpret/validity.rs b/src/librustc_mir/interpret/validity.rs
index 3bb9ba37120..ab836595a7a 100644
--- a/src/librustc_mir/interpret/validity.rs
+++ b/src/librustc_mir/interpret/validity.rs
@@ -276,19 +276,21 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValidityVisitor<'rt, 'mir, '
         }
     }
 
-    fn visit_elem(
+    fn with_elem<R>(
         &mut self,
-        new_op: OpTy<'tcx, M::PointerTag>,
         elem: PathElem,
-    ) -> InterpResult<'tcx> {
+        f: impl FnOnce(&mut Self) -> InterpResult<'tcx, R>,
+    ) -> InterpResult<'tcx, R> {
         // Remember the old state
         let path_len = self.path.len();
-        // Perform operation
+        // Record new element
         self.path.push(elem);
-        self.visit_value(new_op)?;
+        // Perform operation
+        let r = f(self)?;
         // Undo changes
         self.path.truncate(path_len);
-        Ok(())
+        // Done
+        Ok(r)
     }
 
     fn check_wide_ptr_meta(
@@ -366,7 +368,7 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValidityVisitor<'rt, 'mir, '
         let place = try_validation!(
             self.ecx.ref_to_mplace(value),
             self.path,
-            err_ub!(InvalidUninitBytes { .. }) => { "uninitialized {}", kind },
+            err_ub!(InvalidUninitBytes(None)) => { "uninitialized {}", kind },
         );
         if place.layout.is_unsized() {
             self.check_wide_ptr_meta(place.meta, place.layout)?;
@@ -477,7 +479,8 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValidityVisitor<'rt, 'mir, '
                 try_validation!(
                     value.to_bool(),
                     self.path,
-                    err_ub!(InvalidBool(..)) => { "{}", value } expected { "a boolean" },
+                    err_ub!(InvalidBool(..)) | err_ub!(InvalidUninitBytes(None)) =>
+                        { "{}", value } expected { "a boolean" },
                 );
                 Ok(true)
             }
@@ -486,7 +489,8 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValidityVisitor<'rt, 'mir, '
                 try_validation!(
                     value.to_char(),
                     self.path,
-                    err_ub!(InvalidChar(..)) => { "{}", value } expected { "a valid unicode scalar value (in `0..=0x10FFFF` but not in `0xD800..=0xDFFF`)" },
+                    err_ub!(InvalidChar(..)) | err_ub!(InvalidUninitBytes(None)) =>
+                        { "{}", value } expected { "a valid unicode scalar value (in `0..=0x10FFFF` but not in `0xD800..=0xDFFF`)" },
                 );
                 Ok(true)
             }
@@ -515,7 +519,7 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValidityVisitor<'rt, 'mir, '
                 let place = try_validation!(
                     self.ecx.ref_to_mplace(self.ecx.read_immediate(value)?),
                     self.path,
-                    err_ub!(InvalidUninitBytes { .. } ) => { "uninitialized raw pointer" },
+                    err_ub!(InvalidUninitBytes(None)) => { "uninitialized raw pointer" },
                 );
                 if place.layout.is_unsized() {
                     self.check_wide_ptr_meta(place.meta, place.layout)?;
@@ -537,6 +541,7 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValidityVisitor<'rt, 'mir, '
                     self.path,
                     err_ub!(DanglingIntPointer(..)) |
                     err_ub!(InvalidFunctionPointer(..)) |
+                    err_ub!(InvalidUninitBytes(None)) |
                     err_unsup!(ReadBytesAsPointer) =>
                         { "{}", value } expected { "a function pointer" },
                 );
@@ -593,7 +598,7 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValidityVisitor<'rt, 'mir, '
         let value = try_validation!(
             value.not_undef(),
             self.path,
-            err_ub!(InvalidUninitBytes { .. }) => { "{}", value }
+            err_ub!(InvalidUninitBytes(None)) => { "{}", value }
                 expected { "something {}", wrapping_range_format(valid_range, max_hi) },
         );
         let bits = match value.to_bits_or_ptr(op.layout.size, self.ecx) {
@@ -646,6 +651,25 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValueVisitor<'mir, 'tcx, M>
         &self.ecx
     }
 
+    fn read_discriminant(
+        &mut self,
+        op: OpTy<'tcx, M::PointerTag>,
+    ) -> InterpResult<'tcx, VariantIdx> {
+        self.with_elem(PathElem::EnumTag, move |this| {
+            Ok(try_validation!(
+                this.ecx.read_discriminant(op),
+                this.path,
+                err_ub!(InvalidTag(val)) =>
+                    { "{}", val } expected { "a valid enum tag" },
+                err_ub!(InvalidUninitBytes(None)) =>
+                    { "uninitialized bytes" } expected { "a valid enum tag" },
+                err_unsup!(ReadPointerAsBytes) =>
+                    { "a pointer" } expected { "a valid enum tag" },
+            )
+            .1)
+        })
+    }
+
     #[inline]
     fn visit_field(
         &mut self,
@@ -654,7 +678,7 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValueVisitor<'mir, 'tcx, M>
         new_op: OpTy<'tcx, M::PointerTag>,
     ) -> InterpResult<'tcx> {
         let elem = self.aggregate_field_path_elem(old_op.layout, field);
-        self.visit_elem(new_op, elem)
+        self.with_elem(elem, move |this| this.visit_value(new_op))
     }
 
     #[inline]
@@ -670,7 +694,7 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValueVisitor<'mir, 'tcx, M>
             ty::Generator(..) => PathElem::GeneratorState(variant_id),
             _ => bug!("Unexpected type with variant: {:?}", old_op.layout.ty),
         };
-        self.visit_elem(new_op, name)
+        self.with_elem(name, move |this| this.visit_value(new_op))
     }
 
     #[inline(always)]
@@ -693,15 +717,8 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValueVisitor<'mir, 'tcx, M>
         // Sanity check: `builtin_deref` does not know any pointers that are not primitive.
         assert!(op.layout.ty.builtin_deref(true).is_none());
 
-        // Recursively walk the type. Translate some possible errors to something nicer.
-        try_validation!(
-            self.walk_value(op),
-            self.path,
-            err_ub!(InvalidTag(val)) =>
-                { "{}", val } expected { "a valid enum tag" },
-            err_unsup!(ReadPointerAsBytes) =>
-                { "a pointer" } expected { "plain (non-pointer) bytes" },
-        );
+        // Recursively walk the value at its type.
+        self.walk_value(op)?;
 
         // *After* all of this, check the ABI.  We need to check the ABI to handle
         // types like `NonNull` where the `Scalar` info is more restrictive than what
@@ -816,6 +833,10 @@ impl<'rt, 'mir, 'tcx: 'mir, M: Machine<'mir, 'tcx>> ValueVisitor<'mir, 'tcx, M>
 
                                 throw_validation_failure!(self.path, { "uninitialized bytes" })
                             }
+                            err_unsup!(ReadPointerAsBytes) => {
+                                throw_validation_failure!(self.path, { "a pointer" } expected { "plain (non-pointer) bytes" })
+                            }
+
                             // Propagate upwards (that will also check for unexpected errors).
                             _ => return Err(err),
                         }
diff --git a/src/librustc_mir/interpret/visitor.rs b/src/librustc_mir/interpret/visitor.rs
index 903aa377a3d..6c53df40a7c 100644
--- a/src/librustc_mir/interpret/visitor.rs
+++ b/src/librustc_mir/interpret/visitor.rs
@@ -125,6 +125,15 @@ macro_rules! make_value_visitor {
             fn ecx(&$($mutability)? self)
                 -> &$($mutability)? InterpCx<'mir, 'tcx, M>;
 
+            /// `read_discriminant` can be hooked for better error messages.
+            #[inline(always)]
+            fn read_discriminant(
+                &mut self,
+                op: OpTy<'tcx, M::PointerTag>,
+            ) -> InterpResult<'tcx, VariantIdx> {
+                Ok(self.ecx().read_discriminant(op)?.1)
+            }
+
             // Recursive actions, ready to be overloaded.
             /// Visits the given value, dispatching as appropriate to more specialized visitors.
             #[inline(always)]
@@ -245,7 +254,7 @@ macro_rules! make_value_visitor {
                     // with *its* fields.
                     Variants::Multiple { .. } => {
                         let op = v.to_op(self.ecx())?;
-                        let idx = self.ecx().read_discriminant(op)?.1;
+                        let idx = self.read_discriminant(op)?;
                         let inner = v.project_downcast(self.ecx(), idx)?;
                         trace!("walk_value: variant layout: {:#?}", inner.layout());
                         // recurse with the inner type
diff --git a/src/test/ui/consts/const-eval/double_check2.stderr b/src/test/ui/consts/const-eval/double_check2.stderr
index 93dd9a53ec9..513b71f0c6f 100644
--- a/src/test/ui/consts/const-eval/double_check2.stderr
+++ b/src/test/ui/consts/const-eval/double_check2.stderr
@@ -5,7 +5,7 @@ LL | / static FOO: (&Foo, &Bar) = unsafe {(
 LL | |     Union { u8: &BAR }.foo,
 LL | |     Union { u8: &BAR }.bar,
 LL | | )};
-   | |___^ type validation failed: encountered 0x05 at .1.<deref>, but expected a valid enum tag
+   | |___^ type validation failed: encountered 0x05 at .1.<deref>.<enum-tag>, but expected a valid enum tag
    |
    = note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.
 
diff --git a/src/test/ui/consts/const-eval/ub-enum.stderr b/src/test/ui/consts/const-eval/ub-enum.stderr
index 1f7593c6db9..217bfb628a0 100644
--- a/src/test/ui/consts/const-eval/ub-enum.stderr
+++ b/src/test/ui/consts/const-eval/ub-enum.stderr
@@ -2,7 +2,7 @@ error[E0080]: it is undefined behavior to use this value
   --> $DIR/ub-enum.rs:24:1
    |
 LL | const BAD_ENUM: Enum = unsafe { mem::transmute(1usize) };
-   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered 0x00000001, but expected a valid enum tag
+   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered 0x00000001 at .<enum-tag>, but expected a valid enum tag
    |
    = note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.
 
@@ -26,7 +26,7 @@ error[E0080]: it is undefined behavior to use this value
   --> $DIR/ub-enum.rs:42:1
    |
 LL | const BAD_ENUM2: Enum2 = unsafe { mem::transmute(0usize) };
-   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered 0x00000000, but expected a valid enum tag
+   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered 0x00000000 at .<enum-tag>, but expected a valid enum tag
    |
    = note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.
 
diff --git a/src/test/ui/consts/const-eval/union-ub.rs b/src/test/ui/consts/const-eval/union-ub.rs
index 848826e6ef7..512359f5b1c 100644
--- a/src/test/ui/consts/const-eval/union-ub.rs
+++ b/src/test/ui/consts/const-eval/union-ub.rs
@@ -2,6 +2,7 @@
 
 #[repr(C)]
 union DummyUnion {
+    unit: (),
     u8: u8,
     bool: bool,
 }
@@ -30,6 +31,8 @@ union Bar {
 // the value is not valid for bools
 const BAD_BOOL: bool = unsafe { DummyUnion { u8: 42 }.bool};
 //~^ ERROR it is undefined behavior to use this value
+const UNINIT_BOOL: bool = unsafe { DummyUnion { unit: () }.bool};
+//~^ ERROR it is undefined behavior to use this value
 
 // The value is not valid for any union variant, but that's fine
 // unions are just a convenient way to transmute bits around
diff --git a/src/test/ui/consts/const-eval/union-ub.stderr b/src/test/ui/consts/const-eval/union-ub.stderr
index fd3e66765c6..e8869d0d76c 100644
--- a/src/test/ui/consts/const-eval/union-ub.stderr
+++ b/src/test/ui/consts/const-eval/union-ub.stderr
@@ -1,11 +1,19 @@
 error[E0080]: it is undefined behavior to use this value
-  --> $DIR/union-ub.rs:31:1
+  --> $DIR/union-ub.rs:32:1
    |
 LL | const BAD_BOOL: bool = unsafe { DummyUnion { u8: 42 }.bool};
    | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered 0x2a, but expected a boolean
    |
    = note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.
 
-error: aborting due to previous error
+error[E0080]: it is undefined behavior to use this value
+  --> $DIR/union-ub.rs:34:1
+   |
+LL | const UNINIT_BOOL: bool = unsafe { DummyUnion { unit: () }.bool};
+   | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ type validation failed: encountered uninitialized bytes, but expected a boolean
+   |
+   = note: The rules on what exactly is undefined behavior aren't clear, so this check might be overzealous. Please open an issue on the rustc repository if you believe it should not be considered undefined behavior.
+
+error: aborting due to 2 previous errors
 
 For more information about this error, try `rustc --explain E0080`.