broker_tokio/io/util/buf_reader.rs
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194
use crate::io::util::DEFAULT_BUF_SIZE;
use crate::io::{AsyncBufRead, AsyncRead, AsyncWrite};
use pin_project_lite::pin_project;
use std::io::{self, Read};
use std::mem::MaybeUninit;
use std::pin::Pin;
use std::task::{Context, Poll};
use std::{cmp, fmt};
pin_project! {
/// The `BufReader` struct adds buffering to any reader.
///
/// It can be excessively inefficient to work directly with a [`AsyncRead`]
/// instance. A `BufReader` performs large, infrequent reads on the underlying
/// [`AsyncRead`] and maintains an in-memory buffer of the results.
///
/// `BufReader` can improve the speed of programs that make *small* and
/// *repeated* read calls to the same file or network socket. It does not
/// help when reading very large amounts at once, or reading just one or a few
/// times. It also provides no advantage when reading from a source that is
/// already in memory, like a `Vec<u8>`.
///
/// When the `BufReader` is dropped, the contents of its buffer will be
/// discarded. Creating multiple instances of a `BufReader` on the same
/// stream can cause data loss.
#[cfg_attr(docsrs, doc(cfg(feature = "io-util")))]
pub struct BufReader<R> {
#[pin]
pub(super) inner: R,
pub(super) buf: Box<[u8]>,
pub(super) pos: usize,
pub(super) cap: usize,
}
}
impl<R: AsyncRead> BufReader<R> {
/// Creates a new `BufReader` with a default buffer capacity. The default is currently 8 KB,
/// but may change in the future.
pub fn new(inner: R) -> Self {
Self::with_capacity(DEFAULT_BUF_SIZE, inner)
}
/// Creates a new `BufReader` with the specified buffer capacity.
pub fn with_capacity(capacity: usize, inner: R) -> Self {
unsafe {
let mut buffer = Vec::with_capacity(capacity);
buffer.set_len(capacity);
{
// Convert to MaybeUninit
let b = &mut *(&mut buffer[..] as *mut [u8] as *mut [MaybeUninit<u8>]);
inner.prepare_uninitialized_buffer(b);
}
Self {
inner,
buf: buffer.into_boxed_slice(),
pos: 0,
cap: 0,
}
}
}
/// Gets a reference to the underlying reader.
///
/// It is inadvisable to directly read from the underlying reader.
pub fn get_ref(&self) -> &R {
&self.inner
}
/// Gets a mutable reference to the underlying reader.
///
/// It is inadvisable to directly read from the underlying reader.
pub fn get_mut(&mut self) -> &mut R {
&mut self.inner
}
/// Gets a pinned mutable reference to the underlying reader.
///
/// It is inadvisable to directly read from the underlying reader.
pub fn get_pin_mut(self: Pin<&mut Self>) -> Pin<&mut R> {
self.project().inner
}
/// Consumes this `BufWriter`, returning the underlying reader.
///
/// Note that any leftover data in the internal buffer is lost.
pub fn into_inner(self) -> R {
self.inner
}
/// Returns a reference to the internally buffered data.
///
/// Unlike `fill_buf`, this will not attempt to fill the buffer if it is empty.
pub fn buffer(&self) -> &[u8] {
&self.buf[self.pos..self.cap]
}
/// Invalidates all data in the internal buffer.
#[inline]
fn discard_buffer(self: Pin<&mut Self>) {
let me = self.project();
*me.pos = 0;
*me.cap = 0;
}
}
impl<R: AsyncRead> AsyncRead for BufReader<R> {
fn poll_read(
mut self: Pin<&mut Self>,
cx: &mut Context<'_>,
buf: &mut [u8],
) -> Poll<io::Result<usize>> {
// If we don't have any buffered data and we're doing a massive read
// (larger than our internal buffer), bypass our internal buffer
// entirely.
if self.pos == self.cap && buf.len() >= self.buf.len() {
let res = ready!(self.as_mut().get_pin_mut().poll_read(cx, buf));
self.discard_buffer();
return Poll::Ready(res);
}
let mut rem = ready!(self.as_mut().poll_fill_buf(cx))?;
let nread = rem.read(buf)?;
self.consume(nread);
Poll::Ready(Ok(nread))
}
// we can't skip unconditionally because of the large buffer case in read.
unsafe fn prepare_uninitialized_buffer(&self, buf: &mut [MaybeUninit<u8>]) -> bool {
self.inner.prepare_uninitialized_buffer(buf)
}
}
impl<R: AsyncRead> AsyncBufRead for BufReader<R> {
fn poll_fill_buf(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<io::Result<&[u8]>> {
let me = self.project();
// If we've reached the end of our internal buffer then we need to fetch
// some more data from the underlying reader.
// Branch using `>=` instead of the more correct `==`
// to tell the compiler that the pos..cap slice is always valid.
if *me.pos >= *me.cap {
debug_assert!(*me.pos == *me.cap);
*me.cap = ready!(me.inner.poll_read(cx, me.buf))?;
*me.pos = 0;
}
Poll::Ready(Ok(&me.buf[*me.pos..*me.cap]))
}
fn consume(self: Pin<&mut Self>, amt: usize) {
let me = self.project();
*me.pos = cmp::min(*me.pos + amt, *me.cap);
}
}
impl<R: AsyncRead + AsyncWrite> AsyncWrite for BufReader<R> {
fn poll_write(
self: Pin<&mut Self>,
cx: &mut Context<'_>,
buf: &[u8],
) -> Poll<io::Result<usize>> {
self.get_pin_mut().poll_write(cx, buf)
}
fn poll_flush(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<io::Result<()>> {
self.get_pin_mut().poll_flush(cx)
}
fn poll_shutdown(self: Pin<&mut Self>, cx: &mut Context<'_>) -> Poll<io::Result<()>> {
self.get_pin_mut().poll_shutdown(cx)
}
}
impl<R: fmt::Debug> fmt::Debug for BufReader<R> {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
f.debug_struct("BufReader")
.field("reader", &self.inner)
.field(
"buffer",
&format_args!("{}/{}", self.cap - self.pos, self.buf.len()),
)
.finish()
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn assert_unpin() {
crate::is_unpin::<BufReader<()>>();
}
}