Auto merge of #55674 - oli-obk:miri_engine_refactoring, r=RalfJung

Miri engine refactoring

r? @RalfJung

split out the "just moves stuff around" part of #55293

src/librustc/mir/interpret/allocation.rs

@@ -0,0 +1,233 @@
+// Copyright 2018 The Rust Project Developers. See the COPYRIGHT
+// file at the top-level directory of this distribution and at
+// http://rust-lang.org/COPYRIGHT.
+//
+// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+// option. This file may not be copied, modified, or distributed
+// except according to those terms.
+
+//! The virtual memory representation of the MIR interpreter
+
+use super::{Pointer, EvalResult, AllocId};
+
+use ty::layout::{Size, Align};
+use syntax::ast::Mutability;
+use std::iter;
+use mir;
+use std::ops::{Deref, DerefMut};
+use rustc_data_structures::sorted_map::SortedMap;
+
+#[derive(Clone, Debug, Eq, PartialEq, PartialOrd, Ord, Hash, RustcEncodable, RustcDecodable)]
+pub struct Allocation<Tag=(),Extra=()> {
+    /// The actual bytes of the allocation.
+    /// Note that the bytes of a pointer represent the offset of the pointer
+    pub bytes: Vec<u8>,
+    /// Maps from byte addresses to extra data for each pointer.
+    /// Only the first byte of a pointer is inserted into the map; i.e.,
+    /// every entry in this map applies to `pointer_size` consecutive bytes starting
+    /// at the given offset.
+    pub relocations: Relocations<Tag>,
+    /// Denotes undefined memory. Reading from undefined memory is forbidden in miri
+    pub undef_mask: UndefMask,
+    /// The alignment of the allocation to detect unaligned reads.
+    pub align: Align,
+    /// Whether the allocation is mutable.
+    /// Also used by codegen to determine if a static should be put into mutable memory,
+    /// which happens for `static mut` and `static` with interior mutability.
+    pub mutability: Mutability,
+    /// Extra state for the machine.
+    pub extra: Extra,
+}
+
+pub trait AllocationExtra<Tag>: ::std::fmt::Debug + Default + Clone {
+    /// Hook for performing extra checks on a memory read access.
+    ///
+    /// Takes read-only access to the allocation so we can keep all the memory read
+    /// operations take `&self`.  Use a `RefCell` in `AllocExtra` if you
+    /// need to mutate.
+    #[inline]
+    fn memory_read(
+        _alloc: &Allocation<Tag, Self>,
+        _ptr: Pointer<Tag>,
+        _size: Size,
+    ) -> EvalResult<'tcx> {
+        Ok(())
+    }
+
+    /// Hook for performing extra checks on a memory write access.
+    #[inline]
+    fn memory_written(
+        _alloc: &mut Allocation<Tag, Self>,
+        _ptr: Pointer<Tag>,
+        _size: Size,
+    ) -> EvalResult<'tcx> {
+        Ok(())
+    }
+}
+
+impl AllocationExtra<()> for () {}
+
+impl<Tag, Extra: Default> Allocation<Tag, Extra> {
+    /// Creates a read-only allocation initialized by the given bytes
+    pub fn from_bytes(slice: &[u8], align: Align) -> Self {
+        let mut undef_mask = UndefMask::new(Size::ZERO);
+        undef_mask.grow(Size::from_bytes(slice.len() as u64), true);
+        Self {
+            bytes: slice.to_owned(),
+            relocations: Relocations::new(),
+            undef_mask,
+            align,
+            mutability: Mutability::Immutable,
+            extra: Extra::default(),
+        }
+    }
+
+    pub fn from_byte_aligned_bytes(slice: &[u8]) -> Self {
+        Allocation::from_bytes(slice, Align::from_bytes(1, 1).unwrap())
+    }
+
+    pub fn undef(size: Size, align: Align) -> Self {
+        assert_eq!(size.bytes() as usize as u64, size.bytes());
+        Allocation {
+            bytes: vec![0; size.bytes() as usize],
+            relocations: Relocations::new(),
+            undef_mask: UndefMask::new(size),
+            align,
+            mutability: Mutability::Mutable,
+            extra: Extra::default(),
+        }
+    }
+}
+
+impl<'tcx> ::serialize::UseSpecializedDecodable for &'tcx Allocation {}
+
+#[derive(Clone, PartialEq, Eq, PartialOrd, Ord, Hash, Debug, RustcEncodable, RustcDecodable)]
+pub struct Relocations<Tag=(), Id=AllocId>(SortedMap<Size, (Tag, Id)>);
+
+impl<Tag, Id> Relocations<Tag, Id> {
+    pub fn new() -> Self {
+        Relocations(SortedMap::new())
+    }
+
+    // The caller must guarantee that the given relocations are already sorted
+    // by address and contain no duplicates.
+    pub fn from_presorted(r: Vec<(Size, (Tag, Id))>) -> Self {
+        Relocations(SortedMap::from_presorted_elements(r))
+    }
+}
+
+impl<Tag> Deref for Relocations<Tag> {
+    type Target = SortedMap<Size, (Tag, AllocId)>;
+
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
+impl<Tag> DerefMut for Relocations<Tag> {
+    fn deref_mut(&mut self) -> &mut Self::Target {
+        &mut self.0
+    }
+}
+
+////////////////////////////////////////////////////////////////////////////////
+// Undefined byte tracking
+////////////////////////////////////////////////////////////////////////////////
+
+type Block = u64;
+const BLOCK_SIZE: u64 = 64;
+
+#[derive(Clone, Debug, Eq, PartialEq, PartialOrd, Ord, Hash, RustcEncodable, RustcDecodable)]
+pub struct UndefMask {
+    blocks: Vec<Block>,
+    len: Size,
+}
+
+impl_stable_hash_for!(struct mir::interpret::UndefMask{blocks, len});
+
+impl UndefMask {
+    pub fn new(size: Size) -> Self {
+        let mut m = UndefMask {
+            blocks: vec![],
+            len: Size::ZERO,
+        };
+        m.grow(size, false);
+        m
+    }
+
+    /// Check whether the range `start..end` (end-exclusive) is entirely defined.
+    ///
+    /// Returns `Ok(())` if it's defined. Otherwise returns the index of the byte
+    /// at which the first undefined access begins.
+    #[inline]
+    pub fn is_range_defined(&self, start: Size, end: Size) -> Result<(), Size> {
+        if end > self.len {
+            return Err(self.len);
+        }
+
+        let idx = (start.bytes()..end.bytes())
+            .map(|i| Size::from_bytes(i))
+            .find(|&i| !self.get(i));
+
+        match idx {
+            Some(idx) => Err(idx),
+            None => Ok(())
+        }
+    }
+
+    pub fn set_range(&mut self, start: Size, end: Size, new_state: bool) {
+        let len = self.len;
+        if end > len {
+            self.grow(end - len, new_state);
+        }
+        self.set_range_inbounds(start, end, new_state);
+    }
+
+    pub fn set_range_inbounds(&mut self, start: Size, end: Size, new_state: bool) {
+        for i in start.bytes()..end.bytes() {
+            self.set(Size::from_bytes(i), new_state);
+        }
+    }
+
+    #[inline]
+    pub fn get(&self, i: Size) -> bool {
+        let (block, bit) = bit_index(i);
+        (self.blocks[block] & 1 << bit) != 0
+    }
+
+    #[inline]
+    pub fn set(&mut self, i: Size, new_state: bool) {
+        let (block, bit) = bit_index(i);
+        if new_state {
+            self.blocks[block] |= 1 << bit;
+        } else {
+            self.blocks[block] &= !(1 << bit);
+        }
+    }
+
+    pub fn grow(&mut self, amount: Size, new_state: bool) {
+        let unused_trailing_bits = self.blocks.len() as u64 * BLOCK_SIZE - self.len.bytes();
+        if amount.bytes() > unused_trailing_bits {
+            let additional_blocks = amount.bytes() / BLOCK_SIZE + 1;
+            assert_eq!(additional_blocks as usize as u64, additional_blocks);
+            self.blocks.extend(
+                iter::repeat(0).take(additional_blocks as usize),
+            );
+        }
+        let start = self.len;
+        self.len += amount;
+        self.set_range_inbounds(start, start + amount, new_state);
+    }
+}
+
+#[inline]
+fn bit_index(bits: Size) -> (usize, usize) {
+    let bits = bits.bytes();
+    let a = bits / BLOCK_SIZE;
+    let b = bits % BLOCK_SIZE;
+    assert_eq!(a as usize as u64, a);
+    assert_eq!(b as usize as u64, b);
+    (a as usize, b as usize)
+}