Remove the TypedArena::alloc_from_iter specialization.

It was added in #78569. It's complicated and doesn't actually help
performance.

Also, add a comment explaining why the two `alloc_from_iter` functions
are so different.
This commit is contained in:
Nicholas Nethercote 2023-10-03 11:54:46 +11:00
parent e0d7ed1f45
commit 816383c60d

View File

@ -15,7 +15,6 @@
#![feature(dropck_eyepatch)]
#![feature(new_uninit)]
#![feature(maybe_uninit_slice)]
#![feature(min_specialization)]
#![feature(decl_macro)]
#![feature(pointer_byte_offsets)]
#![feature(rustc_attrs)]
@ -44,23 +43,6 @@ fn outline<F: FnOnce() -> R, R>(f: F) -> R {
f()
}
/// An arena that can hold objects of only one type.
pub struct TypedArena<T> {
/// A pointer to the next object to be allocated.
ptr: Cell<*mut T>,
/// A pointer to the end of the allocated area. When this pointer is
/// reached, a new chunk is allocated.
end: Cell<*mut T>,
/// A vector of arena chunks.
chunks: RefCell<Vec<ArenaChunk<T>>>,
/// Marker indicating that dropping the arena causes its owned
/// instances of `T` to be dropped.
_own: PhantomData<T>,
}
struct ArenaChunk<T = u8> {
/// The raw storage for the arena chunk.
storage: NonNull<[MaybeUninit<T>]>,
@ -130,6 +112,23 @@ impl<T> ArenaChunk<T> {
const PAGE: usize = 4096;
const HUGE_PAGE: usize = 2 * 1024 * 1024;
/// An arena that can hold objects of only one type.
pub struct TypedArena<T> {
/// A pointer to the next object to be allocated.
ptr: Cell<*mut T>,
/// A pointer to the end of the allocated area. When this pointer is
/// reached, a new chunk is allocated.
end: Cell<*mut T>,
/// A vector of arena chunks.
chunks: RefCell<Vec<ArenaChunk<T>>>,
/// Marker indicating that dropping the arena causes its owned
/// instances of `T` to be dropped.
_own: PhantomData<T>,
}
impl<T> Default for TypedArena<T> {
/// Creates a new `TypedArena`.
fn default() -> TypedArena<T> {
@ -144,77 +143,6 @@ impl<T> Default for TypedArena<T> {
}
}
trait IterExt<T> {
fn alloc_from_iter(self, arena: &TypedArena<T>) -> &mut [T];
}
impl<I, T> IterExt<T> for I
where
I: IntoIterator<Item = T>,
{
// This default collects into a `SmallVec` and then allocates by copying
// from it. The specializations below for types like `Vec` are more
// efficient, copying directly without the intermediate collecting step.
// This default could be made more efficient, like
// `DroplessArena::alloc_from_iter`, but it's not hot enough to bother.
#[inline]
default fn alloc_from_iter(self, arena: &TypedArena<T>) -> &mut [T] {
let vec: SmallVec<[_; 8]> = self.into_iter().collect();
vec.alloc_from_iter(arena)
}
}
impl<T, const N: usize> IterExt<T> for std::array::IntoIter<T, N> {
#[inline]
fn alloc_from_iter(self, arena: &TypedArena<T>) -> &mut [T] {
let len = self.len();
if len == 0 {
return &mut [];
}
// Move the content to the arena by copying and then forgetting it.
let start_ptr = arena.alloc_raw_slice(len);
unsafe {
self.as_slice().as_ptr().copy_to_nonoverlapping(start_ptr, len);
mem::forget(self);
slice::from_raw_parts_mut(start_ptr, len)
}
}
}
impl<T> IterExt<T> for Vec<T> {
#[inline]
fn alloc_from_iter(mut self, arena: &TypedArena<T>) -> &mut [T] {
let len = self.len();
if len == 0 {
return &mut [];
}
// Move the content to the arena by copying and then forgetting it.
let start_ptr = arena.alloc_raw_slice(len);
unsafe {
self.as_ptr().copy_to_nonoverlapping(start_ptr, len);
self.set_len(0);
slice::from_raw_parts_mut(start_ptr, len)
}
}
}
impl<A: smallvec::Array> IterExt<A::Item> for SmallVec<A> {
#[inline]
fn alloc_from_iter(mut self, arena: &TypedArena<A::Item>) -> &mut [A::Item] {
let len = self.len();
if len == 0 {
return &mut [];
}
// Move the content to the arena by copying and then forgetting it.
let start_ptr = arena.alloc_raw_slice(len);
unsafe {
self.as_ptr().copy_to_nonoverlapping(start_ptr, len);
self.set_len(0);
slice::from_raw_parts_mut(start_ptr, len)
}
}
}
impl<T> TypedArena<T> {
/// Allocates an object in the `TypedArena`, returning a reference to it.
#[inline]
@ -270,8 +198,35 @@ impl<T> TypedArena<T> {
#[inline]
pub fn alloc_from_iter<I: IntoIterator<Item = T>>(&self, iter: I) -> &mut [T] {
// This implementation is entirely separate to
// `DroplessIterator::alloc_from_iter`, even though conceptually they
// are the same.
//
// `DroplessIterator` (in the fast case) writes elements from the
// iterator one at a time into the allocated memory. That's easy
// because the elements don't implement `Drop`. But for `TypedArena`
// they do implement `Drop`, which means that if the iterator panics we
// could end up with some allocated-but-uninitialized elements, which
// will then cause UB in `TypedArena::drop`.
//
// Instead we use an approach where any iterator panic will occur
// before the memory is allocated. This function is much less hot than
// `DroplessArena::alloc_from_iter`, so it doesn't need to be
// hyper-optimized.
assert!(mem::size_of::<T>() != 0);
iter.alloc_from_iter(self)
let mut vec: SmallVec<[_; 8]> = iter.into_iter().collect();
if vec.is_empty() {
return &mut [];
}
// Move the content to the arena by copying and then forgetting it.
let len = vec.len();
let start_ptr = self.alloc_raw_slice(len);
unsafe {
vec.as_ptr().copy_to_nonoverlapping(start_ptr, len);
vec.set_len(0);
slice::from_raw_parts_mut(start_ptr, len)
}
}
/// Grows the arena.