Constify mem::swap and ptr::swap[_nonoverlapping]

This commit is contained in:
Albin Hedman 2021-03-13 20:33:27 +01:00
parent 62cf244563
commit 64e2248794
No known key found for this signature in database
GPG Key ID: 1F501ECD1BBD70A7
4 changed files with 12 additions and 27 deletions

View File

@ -1902,18 +1902,6 @@ pub(crate) fn is_aligned_and_not_null<T>(ptr: *const T) -> bool {
!ptr.is_null() && ptr as usize % mem::align_of::<T>() == 0
}
/// Checks whether the regions of memory starting at `src` and `dst` of size
/// `count * size_of::<T>()` do *not* overlap.
pub(crate) fn is_nonoverlapping<T>(src: *const T, dst: *const T, count: usize) -> bool {
let src_usize = src as usize;
let dst_usize = dst as usize;
let size = mem::size_of::<T>().checked_mul(count).unwrap();
let diff = if src_usize > dst_usize { src_usize - dst_usize } else { dst_usize - src_usize };
// If the absolute distance between the ptrs is at least as big as the size of the buffer,
// they do not overlap.
diff >= size
}
/// Sets `count * size_of::<T>()` bytes of memory starting at `dst` to
/// `val`.
///

View File

@ -98,6 +98,7 @@
#![feature(const_slice_from_raw_parts)]
#![feature(const_slice_ptr_len)]
#![feature(const_size_of_val)]
#![feature(const_swap)]
#![feature(const_align_of_val)]
#![feature(const_type_id)]
#![feature(const_type_name)]

View File

@ -682,7 +682,8 @@ pub unsafe fn uninitialized<T>() -> T {
/// ```
#[inline]
#[stable(feature = "rust1", since = "1.0.0")]
pub fn swap<T>(x: &mut T, y: &mut T) {
#[rustc_const_unstable(feature = "const_swap", issue = "83163")]
pub const fn swap<T>(x: &mut T, y: &mut T) {
// SAFETY: the raw pointers have been created from safe mutable references satisfying all the
// constraints on `ptr::swap_nonoverlapping_one`
unsafe {

View File

@ -67,7 +67,7 @@
use crate::cmp::Ordering;
use crate::fmt;
use crate::hash;
use crate::intrinsics::{self, abort, is_aligned_and_not_null, is_nonoverlapping};
use crate::intrinsics::{self, abort, is_aligned_and_not_null};
use crate::mem::{self, MaybeUninit};
#[stable(feature = "rust1", since = "1.0.0")]
@ -394,7 +394,8 @@ pub const fn slice_from_raw_parts_mut<T>(data: *mut T, len: usize) -> *mut [T] {
/// ```
#[inline]
#[stable(feature = "rust1", since = "1.0.0")]
pub unsafe fn swap<T>(x: *mut T, y: *mut T) {
#[rustc_const_unstable(feature = "const_swap", issue = "83163")]
pub const unsafe fn swap<T>(x: *mut T, y: *mut T) {
// Give ourselves some scratch space to work with.
// We do not have to worry about drops: `MaybeUninit` does nothing when dropped.
let mut tmp = MaybeUninit::<T>::uninit();
@ -451,16 +452,8 @@ pub unsafe fn swap<T>(x: *mut T, y: *mut T) {
/// ```
#[inline]
#[stable(feature = "swap_nonoverlapping", since = "1.27.0")]
pub unsafe fn swap_nonoverlapping<T>(x: *mut T, y: *mut T, count: usize) {
if cfg!(debug_assertions)
&& !(is_aligned_and_not_null(x)
&& is_aligned_and_not_null(y)
&& is_nonoverlapping(x, y, count))
{
// Not panicking to keep codegen impact smaller.
abort();
}
#[rustc_const_unstable(feature = "const_swap", issue = "83163")]
pub const unsafe fn swap_nonoverlapping<T>(x: *mut T, y: *mut T, count: usize) {
let x = x as *mut u8;
let y = y as *mut u8;
let len = mem::size_of::<T>() * count;
@ -470,7 +463,8 @@ pub unsafe fn swap_nonoverlapping<T>(x: *mut T, y: *mut T, count: usize) {
}
#[inline]
pub(crate) unsafe fn swap_nonoverlapping_one<T>(x: *mut T, y: *mut T) {
#[rustc_const_unstable(feature = "const_swap", issue = "83163")]
pub(crate) const unsafe fn swap_nonoverlapping_one<T>(x: *mut T, y: *mut T) {
// For types smaller than the block optimization below,
// just swap directly to avoid pessimizing codegen.
if mem::size_of::<T>() < 32 {
@ -488,7 +482,8 @@ pub(crate) unsafe fn swap_nonoverlapping_one<T>(x: *mut T, y: *mut T) {
}
#[inline]
unsafe fn swap_nonoverlapping_bytes(x: *mut u8, y: *mut u8, len: usize) {
#[rustc_const_unstable(feature = "const_swap", issue = "83163")]
const unsafe fn swap_nonoverlapping_bytes(x: *mut u8, y: *mut u8, len: usize) {
// The approach here is to utilize simd to swap x & y efficiently. Testing reveals
// that swapping either 32 bytes or 64 bytes at a time is most efficient for Intel
// Haswell E processors. LLVM is more able to optimize if we give a struct a