cat: Improve performance on Linux (#1978)

* cat: Improve performance, especially on Linux

* cat: Don't use io::copy for splice fallback

On my MacBook Pro 2020, it is around 25% faster to not use io::copy.

* cat: Only fall back to generic copy if first splice fails

* cat: Don't double buffer stdout

* cat: Don't use experimental or-pattern syntax

* cat: Remove nix symbol use from non-Linux

Cargo.lock

@@ -1,5 +1,7 @@
 # This file is automatically @generated by Cargo.
 # It is not intended for manual editing.
+version = 3
+
 [[package]]
 name = "advapi32-sys"
 version = "0.2.0"
@@ -1581,7 +1583,8 @@ name = "uu_cat"
 version = "0.0.4"
 dependencies = [
  "clap",
- "quick-error",
+ "nix 0.20.0",
+ "thiserror",
  "unix_socket",
  "uucore",
  "uucore_procs",

src/uu/cat/Cargo.toml

@@ -16,13 +16,16 @@ path = "src/cat.rs"
 
 [dependencies]
 clap = "2.33"
-quick-error = "1.2.3"
+thiserror = "1.0"
 uucore = { version=">=0.0.7", package="uucore", path="../../uucore", features=["fs"] }
 uucore_procs = { version=">=0.0.5", package="uucore_procs", path="../../uucore_procs" }
 
 [target.'cfg(unix)'.dependencies]
 unix_socket = "0.5.0"
 
+[target.'cfg(any(target_os = "linux", target_os = "android"))'.dependencies]
+nix = "0.20"
+
 [[bin]]
 name = "cat"
 path = "src/main.rs"

src/uu/cat/src/cat.rs

@@ -3,14 +3,13 @@
 // (c) Jordi Boggiano <j.boggiano@seld.be>
 // (c) Evgeniy Klyuchikov <evgeniy.klyuchikov@gmail.com>
 // (c) Joshua S. Miller <jsmiller@uchicago.edu>
+// (c) Árni Dagur <arni@dagur.eu>
 //
 // For the full copyright and license information, please view the LICENSE
 // file that was distributed with this source code.
 
 // spell-checker:ignore (ToDO) nonprint nonblank nonprinting
 
-#[macro_use]
-extern crate quick_error;
 #[cfg(unix)]
 extern crate unix_socket;
 #[macro_use]
@@ -18,9 +17,9 @@ extern crate uucore;
 
 // last synced with: cat (GNU coreutils) 8.13
 use clap::{App, Arg};
-use quick_error::ResultExt;
 use std::fs::{metadata, File};
-use std::io::{self, stderr, stdin, stdout, BufWriter, Read, Write};
+use std::io::{self, Read, Write};
+use thiserror::Error;
 use uucore::fs::is_stdin_interactive;
 
 /// Unix domain socket support
@@ -31,12 +30,44 @@ use std::os::unix::fs::FileTypeExt;
 #[cfg(unix)]
 use unix_socket::UnixStream;
 
+#[cfg(any(target_os = "linux", target_os = "android"))]
+use nix::errno::Errno;
+/// Linux splice support
+#[cfg(any(target_os = "linux", target_os = "android"))]
+use nix::fcntl::{splice, SpliceFFlags};
+#[cfg(any(target_os = "linux", target_os = "android"))]
+use nix::unistd::pipe;
+#[cfg(any(target_os = "linux", target_os = "android"))]
+use std::os::unix::io::{AsRawFd, RawFd};
+
 static NAME: &str = "cat";
 static VERSION: &str = env!("CARGO_PKG_VERSION");
 static SYNTAX: &str = "[OPTION]... [FILE]...";
 static SUMMARY: &str = "Concatenate FILE(s), or standard input, to standard output
  With no FILE, or when FILE is -, read standard input.";
 
+#[derive(Error, Debug)]
+enum CatError {
+    /// Wrapper around `io::Error`
+    #[error("{0}")]
+    Io(#[from] io::Error),
+    /// Wrapper around `nix::Error`
+    #[cfg(any(target_os = "linux", target_os = "android"))]
+    #[error("{0}")]
+    Nix(#[from] nix::Error),
+    /// Unknown file type; it's not a regular file, socket, etc.
+    #[error("{}: unknown filetype: {}", path, ft_debug)]
+    UnknownFiletype {
+        path: String,
+        /// A debug print of the file type
+        ft_debug: String,
+    },
+    #[error("{0}: Expected a file, found directory")]
+    IsDirectory(String),
+}
+
+type CatResult<T> = Result<T, CatError>;
+
 #[derive(PartialEq)]
 enum NumberingMode {
     None,
@@ -44,39 +75,6 @@ enum NumberingMode {
     All,
 }
 
-quick_error! {
-    #[derive(Debug)]
-    enum CatError {
-        /// Wrapper for io::Error with path context
-        Input(err: io::Error, path: String) {
-            display("cat: {0}: {1}", path, err)
-            context(path: &'a str, err: io::Error) -> (err, path.to_owned())
-            cause(err)
-        }
-
-        /// Wrapper for io::Error with no context
-        Output(err: io::Error) {
-            display("cat: {0}", err) from()
-            cause(err)
-        }
-
-        /// Unknown Filetype  classification
-        UnknownFiletype(path: String) {
-            display("cat: {0}: unknown filetype", path)
-        }
-
-        /// At least one error was encountered in reading or writing
-        EncounteredErrors(count: usize) {
-            display("cat: encountered {0} errors", count)
-        }
-
-        /// Denotes an error caused by trying to `cat` a directory
-        IsDirectory(path: String) {
-            display("cat: {0}: Is a directory", path)
-        }
-    }
-}
-
 struct OutputOptions {
     /// Line numbering mode
     number: NumberingMode,
@@ -87,21 +85,56 @@ struct OutputOptions {
     /// display TAB characters as `tab`
     show_tabs: bool,
 
-    /// If `show_tabs == true`, this string will be printed in the
+    /// Show end of lines
-    /// place of tabs
+    show_ends: bool,
-    tab: String,
-
-    /// Can be set to show characters other than '\n' a the end of
-    /// each line, e.g. $
-    end_of_line: String,
 
     /// use ^ and M- notation, except for LF (\\n) and TAB (\\t)
     show_nonprint: bool,
 }
 
+impl OutputOptions {
+    fn tab(&self) -> &'static str {
+        if self.show_tabs {
+            "^I"
+        } else {
+            "\t"
+        }
+    }
+
+    fn end_of_line(&self) -> &'static str {
+        if self.show_ends {
+            "$\n"
+        } else {
+            "\n"
+        }
+    }
+
+    /// We can write fast if we can simply copy the contents of the file to
+    /// stdout, without augmenting the output with e.g. line numbers.
+    fn can_write_fast(&self) -> bool {
+        !(self.show_tabs
+            || self.show_nonprint
+            || self.show_ends
+            || self.squeeze_blank
+            || self.number != NumberingMode::None)
+    }
+}
+
+/// State that persists between output of each file. This struct is only used
+/// when we can't write fast.
+struct OutputState {
+    /// The current line number
+    line_number: usize,
+
+    /// Whether the output cursor is at the beginning of a new line
+    at_line_start: bool,
+}
+
 /// Represents an open file handle, stream, or other device
-struct InputHandle {
+struct InputHandle<R: Read> {
-    reader: Box<dyn Read>,
+    #[cfg(any(target_os = "linux", target_os = "android"))]
+    file_descriptor: RawFd,
+    reader: R,
     is_interactive: bool,
 }
 
@@ -124,8 +157,6 @@ enum InputType {
     Socket,
 }
 
-type CatResult<T> = Result<T, CatError>;
-
 mod options {
     pub static FILE: &str = "file";
     pub static SHOW_ALL: &str = "show-all";
@@ -243,30 +274,14 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
         None => vec!["-".to_owned()],
     };
 
-    let can_write_fast = !(show_tabs
+    let options = OutputOptions {
-        || show_nonprint
+        show_ends,
-        || show_ends
+        number: number_mode,
-        || squeeze_blank
+        show_nonprint,
-        || number_mode != NumberingMode::None);
+        show_tabs,
-
+        squeeze_blank,
-    let success = if can_write_fast {
-        write_fast(files).is_ok()
-    } else {
-        let tab = if show_tabs { "^I" } else { "\t" }.to_owned();
-
-        let end_of_line = if show_ends { "$\n" } else { "\n" }.to_owned();
-
-        let options = OutputOptions {
-            end_of_line,
-            number: number_mode,
-            show_nonprint,
-            show_tabs,
-            squeeze_blank,
-            tab,
-        };
-
-        write_lines(files, &options).is_ok()
     };
+    let success = cat_files(files, &options).is_ok();
 
     if success {
         0
@@ -275,6 +290,76 @@ pub fn uumain(args: impl uucore::Args) -> i32 {
     }
 }
 
+fn cat_handle<R: Read>(
+    handle: &mut InputHandle<R>,
+    options: &OutputOptions,
+    state: &mut OutputState,
+) -> CatResult<()> {
+    if options.can_write_fast() {
+        write_fast(handle)
+    } else {
+        write_lines(handle, &options, state)
+    }
+}
+
+fn cat_path(path: &str, options: &OutputOptions, state: &mut OutputState) -> CatResult<()> {
+    if path == "-" {
+        let stdin = io::stdin();
+        let mut handle = InputHandle {
+            #[cfg(any(target_os = "linux", target_os = "android"))]
+            file_descriptor: stdin.as_raw_fd(),
+            reader: stdin,
+            is_interactive: is_stdin_interactive(),
+        };
+        return cat_handle(&mut handle, &options, state);
+    }
+    match get_input_type(path)? {
+        InputType::Directory => Err(CatError::IsDirectory(path.to_owned())),
+        #[cfg(unix)]
+        InputType::Socket => {
+            let socket = UnixStream::connect(path)?;
+            socket.shutdown(Shutdown::Write)?;
+            let mut handle = InputHandle {
+                #[cfg(any(target_os = "linux", target_os = "android"))]
+                file_descriptor: socket.as_raw_fd(),
+                reader: socket,
+                is_interactive: false,
+            };
+            cat_handle(&mut handle, &options, state)
+        }
+        _ => {
+            let file = File::open(path)?;
+            let mut handle = InputHandle {
+                #[cfg(any(target_os = "linux", target_os = "android"))]
+                file_descriptor: file.as_raw_fd(),
+                reader: file,
+                is_interactive: false,
+            };
+            cat_handle(&mut handle, &options, state)
+        }
+    }
+}
+
+fn cat_files(files: Vec<String>, options: &OutputOptions) -> Result<(), u32> {
+    let mut error_count = 0;
+    let mut state = OutputState {
+        line_number: 1,
+        at_line_start: true,
+    };
+
+    for path in &files {
+        if let Err(err) = cat_path(path, &options, &mut state) {
+            show_error!("{}", err);
+            error_count += 1;
+        }
+    }
+    if error_count == 0 {
+        Ok(())
+    } else {
+        Err(error_count)
+    }
+}
+
 /// Classifies the `InputType` of file at `path` if possible
 ///
 /// # Arguments
@@ -285,7 +370,8 @@ fn get_input_type(path: &str) -> CatResult<InputType> {
         return Ok(InputType::StdIn);
     }
 
-    match metadata(path).context(path)?.file_type() {
+    let ft = metadata(path)?.file_type();
+    match ft {
         #[cfg(unix)]
         ft if ft.is_block_device() => Ok(InputType::BlockDevice),
         #[cfg(unix)]
@@ -297,125 +383,113 @@ fn get_input_type(path: &str) -> CatResult<InputType> {
         ft if ft.is_dir() => Ok(InputType::Directory),
         ft if ft.is_file() => Ok(InputType::File),
         ft if ft.is_symlink() => Ok(InputType::SymLink),
-        _ => Err(CatError::UnknownFiletype(path.to_owned())),
+        _ => Err(CatError::UnknownFiletype {
+            path: path.to_owned(),
+            ft_debug: format!("{:?}", ft),
+        }),
     }
 }
 
-/// Returns an InputHandle from which a Reader can be accessed or an
+/// Writes handle to stdout with no configuration. This allows a
-/// error
+/// simple memory copy.
-///
+fn write_fast<R: Read>(handle: &mut InputHandle<R>) -> CatResult<()> {
-/// # Arguments
+    let stdout = io::stdout();
-///
+    let mut stdout_lock = stdout.lock();
-/// * `path` - `InputHandler` will wrap a reader from this file path
+    #[cfg(any(target_os = "linux", target_os = "android"))]
-fn open(path: &str) -> CatResult<InputHandle> {
+    {
-    if path == "-" {
+        // If we're on Linux or Android, try to use the splice() system call
-        let stdin = stdin();
+        // for faster writing. If it works, we're done.
-        return Ok(InputHandle {
+        if !write_fast_using_splice(handle, stdout.as_raw_fd())? {
-            reader: Box::new(stdin) as Box<dyn Read>,
+            return Ok(());
-            is_interactive: is_stdin_interactive(),
-        });
-    }
-
-    match get_input_type(path)? {
-        InputType::Directory => Err(CatError::IsDirectory(path.to_owned())),
-        #[cfg(unix)]
-        InputType::Socket => {
-            let socket = UnixStream::connect(path).context(path)?;
-            socket.shutdown(Shutdown::Write).context(path)?;
-            Ok(InputHandle {
-                reader: Box::new(socket) as Box<dyn Read>,
-                is_interactive: false,
-            })
-        }
-        _ => {
-            let file = File::open(path).context(path)?;
-            Ok(InputHandle {
-                reader: Box::new(file) as Box<dyn Read>,
-                is_interactive: false,
-            })
         }
     }
+    // If we're not on Linux or Android, or the splice() call failed,
+    // fall back on slower writing.
+    let mut buf = [0; 1024 * 64];
+    while let Ok(n) = handle.reader.read(&mut buf) {
+        if n == 0 {
+            break;
+        }
+        stdout_lock.write_all(&buf[..n])?;
+    }
+    Ok(())
 }
 
-/// Writes files to stdout with no configuration.  This allows a
+/// This function is called from `write_fast()` on Linux and Android. The
-/// simple memory copy. Returns `Ok(())` if no errors were
+/// function `splice()` is used to move data between two file descriptors
-/// encountered, or an error with the number of errors encountered.
+/// without copying between kernel- and userspace. This results in a large
+/// speedup.
 ///
-/// # Arguments
+/// The `bool` in the result value indicates if we need to fall back to normal
-///
+/// copying or not. False means we don't have to.
-/// * `files` - There is no short circuit when encountering an error
+#[cfg(any(target_os = "linux", target_os = "android"))]
-/// reading a file in this vector
+#[inline]
-fn write_fast(files: Vec<String>) -> CatResult<()> {
+fn write_fast_using_splice<R: Read>(handle: &mut InputHandle<R>, writer: RawFd) -> CatResult<bool> {
-    let mut writer = stdout();
+    const BUF_SIZE: usize = 1024 * 16;
-    let mut in_buf = [0; 1024 * 64];
-    let mut error_count = 0;
 
-    for file in files {
+    let (pipe_rd, pipe_wr) = pipe()?;
-        match open(&file[..]) {
+
-            Ok(mut handle) => {
+    // We only fall back if splice fails on the first call.
-                while let Ok(n) = handle.reader.read(&mut in_buf) {
+    match splice(
-                    if n == 0 {
+        handle.file_descriptor,
-                        break;
+        None,
-                    }
+        pipe_wr,
-                    writer.write_all(&in_buf[..n]).context(&file[..])?;
+        None,
+        BUF_SIZE,
+        SpliceFFlags::empty(),
+    ) {
+        Ok(n) => {
+            if n == 0 {
+                return Ok(false);
+            }
+        }
+        Err(err) => {
+            match err.as_errno() {
+                Some(Errno::EPERM) | Some(Errno::ENOSYS) | Some(Errno::EINVAL) => {
+                    // EPERM indicates the call was blocked by seccomp.
+                    // ENOSYS indicates we're running on an ancient Kernel.
+                    // EINVAL indicates some other failure.
+                    return Ok(true);
+                }
+                _ => {
+                    // Other errors include running out of memory, etc. We
+                    // don't attempt to fall back from these.
+                    return Err(err)?;
                 }
             }
-            Err(error) => {
-                writeln!(&mut stderr(), "{}", error)?;
-                error_count += 1;
-            }
         }
     }
 
-    match error_count {
+    loop {
-        0 => Ok(()),
+        let n = splice(
-        _ => Err(CatError::EncounteredErrors(error_count)),
+            handle.file_descriptor,
-    }
+            None,
-}
+            pipe_wr,
-
+            None,
-/// State that persists between output of each file
+            BUF_SIZE,
-struct OutputState {
+            SpliceFFlags::empty(),
-    /// The current line number
+        )?;
-    line_number: usize,
+        if n == 0 {
-
+            // We read 0 bytes from the input,
-    /// Whether the output cursor is at the beginning of a new line
+            // which means we're done copying.
-    at_line_start: bool,
+            break;
-}
-
-/// Writes files to stdout with `options` as configuration.  Returns
-/// `Ok(())` if no errors were encountered, or an error with the
-/// number of errors encountered.
-///
-/// # Arguments
-///
-/// * `files` - There is no short circuit when encountering an error
-/// reading a file in this vector
-fn write_lines(files: Vec<String>, options: &OutputOptions) -> CatResult<()> {
-    let mut error_count = 0;
-    let mut state = OutputState {
-        line_number: 1,
-        at_line_start: true,
-    };
-
-    for file in files {
-        if let Err(error) = write_file_lines(&file, options, &mut state) {
-            writeln!(&mut stderr(), "{}", error).context(&file[..])?;
-            error_count += 1;
         }
+        splice(pipe_rd, None, writer, None, BUF_SIZE, SpliceFFlags::empty())?;
     }
 
-    match error_count {
+    Ok(false)
-        0 => Ok(()),
-        _ => Err(CatError::EncounteredErrors(error_count)),
-    }
 }
 
 /// Outputs file contents to stdout in a line-by-line fashion,
 /// propagating any errors that might occur.
-fn write_file_lines(file: &str, options: &OutputOptions, state: &mut OutputState) -> CatResult<()> {
+fn write_lines<R: Read>(
-    let mut handle = open(file)?;
+    handle: &mut InputHandle<R>,
+    options: &OutputOptions,
+    state: &mut OutputState,
+) -> CatResult<()> {
     let mut in_buf = [0; 1024 * 31];
-    let mut writer = BufWriter::with_capacity(1024 * 64, stdout());
+    let stdout = io::stdout();
+    let mut writer = stdout.lock();
     let mut one_blank_kept = false;
 
     while let Ok(n) = handle.reader.read(&mut in_buf) {
@@ -433,9 +507,9 @@ fn write_file_lines(file: &str, options: &OutputOptions, state: &mut OutputState
                         write!(&mut writer, "{0:6}\t", state.line_number)?;
                         state.line_number += 1;
                     }
-                    writer.write_all(options.end_of_line.as_bytes())?;
+                    writer.write_all(options.end_of_line().as_bytes())?;
                     if handle.is_interactive {
-                        writer.flush().context(file)?;
+                        writer.flush()?;
                     }
                 }
                 state.at_line_start = true;
@@ -450,7 +524,7 @@ fn write_file_lines(file: &str, options: &OutputOptions, state: &mut OutputState
 
             // print to end of line or end of buffer
             let offset = if options.show_nonprint {
-                write_nonprint_to_end(&in_buf[pos..], &mut writer, options.tab.as_bytes())
+                write_nonprint_to_end(&in_buf[pos..], &mut writer, options.tab().as_bytes())
             } else if options.show_tabs {
                 write_tab_to_end(&in_buf[pos..], &mut writer)
             } else {
@@ -462,7 +536,7 @@ fn write_file_lines(file: &str, options: &OutputOptions, state: &mut OutputState
                 break;
             }
             // print suitable end of line
-            writer.write_all(options.end_of_line.as_bytes())?;
+            writer.write_all(options.end_of_line().as_bytes())?;
             if handle.is_interactive {
                 writer.flush()?;
             }