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
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
// Copyright 2016 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.

use collections::hash_map::HashMap;
use env::{self, split_paths};
use ffi::OsStr;
use os::unix::ffi::OsStrExt;
use fmt;
use io::{self, Error, ErrorKind};
use path::{Path, PathBuf};
use sys::fd::FileDesc;
use sys::fs::{File, OpenOptions};
use sys::pipe::{self, AnonPipe};
use sys::{cvt, syscall};

////////////////////////////////////////////////////////////////////////////////
// Command
////////////////////////////////////////////////////////////////////////////////

pub struct Command {
    // Currently we try hard to ensure that the call to `.exec()` doesn't
    // actually allocate any memory. While many platforms try to ensure that
    // memory allocation works after a fork in a multithreaded process, it's
    // been observed to be buggy and somewhat unreliable, so we do our best to
    // just not do it at all!
    //
    // Along those lines, the `argv` and `envp` raw pointers here are exactly
    // what's gonna get passed to `execvp`. The `argv` array starts with the
    // `program` and ends with a NULL, and the `envp` pointer, if present, is
    // also null-terminated.
    //
    // Right now we don't support removing arguments, so there's no much fancy
    // support there, but we support adding and removing environment variables,
    // so a side table is used to track where in the `envp` array each key is
    // located. Whenever we add a key we update it in place if it's already
    // present, and whenever we remove a key we update the locations of all
    // other keys.
    program: String,
    args: Vec<String>,
    env: HashMap<String, String>,

    cwd: Option<String>,
    uid: Option<u32>,
    gid: Option<u32>,
    saw_nul: bool,
    closures: Vec<Box<FnMut() -> io::Result<()> + Send + Sync>>,
    stdin: Option<Stdio>,
    stdout: Option<Stdio>,
    stderr: Option<Stdio>,
}

// passed back to std::process with the pipes connected to the child, if any
// were requested
pub struct StdioPipes {
    pub stdin: Option<AnonPipe>,
    pub stdout: Option<AnonPipe>,
    pub stderr: Option<AnonPipe>,
}

// passed to do_exec() with configuration of what the child stdio should look
// like
struct ChildPipes {
    stdin: ChildStdio,
    stdout: ChildStdio,
    stderr: ChildStdio,
}

enum ChildStdio {
    Inherit,
    Explicit(usize),
    Owned(FileDesc),
}

pub enum Stdio {
    Inherit,
    Null,
    MakePipe,
    Fd(FileDesc),
}

impl Command {
    pub fn new(program: &OsStr) -> Command {
        Command {
            program: program.to_str().unwrap().to_owned(),
            args: Vec::new(),
            env: HashMap::new(),
            cwd: None,
            uid: None,
            gid: None,
            saw_nul: false,
            closures: Vec::new(),
            stdin: None,
            stdout: None,
            stderr: None,
        }
    }

    pub fn arg(&mut self, arg: &OsStr) {
        self.args.push(arg.to_str().unwrap().to_owned());
    }

    pub fn env(&mut self, key: &OsStr, val: &OsStr) {
        self.env.insert(key.to_str().unwrap().to_owned(), val.to_str().unwrap().to_owned());
    }

    pub fn env_remove(&mut self, key: &OsStr) {
        self.env.remove(key.to_str().unwrap());
    }

    pub fn env_clear(&mut self) {
        self.env.clear();
    }

    pub fn cwd(&mut self, dir: &OsStr) {
        self.cwd = Some(dir.to_str().unwrap().to_owned());
    }
    pub fn uid(&mut self, id: u32) {
        self.uid = Some(id);
    }
    pub fn gid(&mut self, id: u32) {
        self.gid = Some(id);
    }

    pub fn before_exec(&mut self,
                       f: Box<FnMut() -> io::Result<()> + Send + Sync>) {
        self.closures.push(f);
    }

    pub fn stdin(&mut self, stdin: Stdio) {
        self.stdin = Some(stdin);
    }
    pub fn stdout(&mut self, stdout: Stdio) {
        self.stdout = Some(stdout);
    }
    pub fn stderr(&mut self, stderr: Stdio) {
        self.stderr = Some(stderr);
    }

    pub fn spawn(&mut self, default: Stdio, needs_stdin: bool)
                 -> io::Result<(Process, StdioPipes)> {
         const CLOEXEC_MSG_FOOTER: &'static [u8] = b"NOEX";

         if self.saw_nul {
             return Err(io::Error::new(ErrorKind::InvalidInput,
                                       "nul byte found in provided data"));
         }

         let (ours, theirs) = self.setup_io(default, needs_stdin)?;
         let (input, output) = pipe::anon_pipe()?;

         let pid = unsafe {
             match cvt(syscall::clone(0))? {
                 0 => {
                     drop(input);
                     let err = self.do_exec(theirs);
                     let errno = err.raw_os_error().unwrap_or(syscall::EINVAL) as u32;
                     let bytes = [
                         (errno >> 24) as u8,
                         (errno >> 16) as u8,
                         (errno >>  8) as u8,
                         (errno >>  0) as u8,
                         CLOEXEC_MSG_FOOTER[0], CLOEXEC_MSG_FOOTER[1],
                         CLOEXEC_MSG_FOOTER[2], CLOEXEC_MSG_FOOTER[3]
                     ];
                     // pipe I/O up to PIPE_BUF bytes should be atomic, and then
                     // we want to be sure we *don't* run at_exit destructors as
                     // we're being torn down regardless
                     assert!(output.write(&bytes).is_ok());
                     let _ = syscall::exit(1);
                     panic!("failed to exit");
                 }
                 n => n,
             }
         };

         let mut p = Process { pid: pid, status: None };
         drop(output);
         let mut bytes = [0; 8];

         // loop to handle EINTR
         loop {
             match input.read(&mut bytes) {
                 Ok(0) => return Ok((p, ours)),
                 Ok(8) => {
                     assert!(combine(CLOEXEC_MSG_FOOTER) == combine(&bytes[4.. 8]),
                             "Validation on the CLOEXEC pipe failed: {:?}", bytes);
                     let errno = combine(&bytes[0.. 4]);
                     assert!(p.wait().is_ok(),
                             "wait() should either return Ok or panic");
                     return Err(Error::from_raw_os_error(errno))
                 }
                 Err(ref e) if e.kind() == ErrorKind::Interrupted => {}
                 Err(e) => {
                     assert!(p.wait().is_ok(),
                             "wait() should either return Ok or panic");
                     panic!("the CLOEXEC pipe failed: {:?}", e)
                 },
                 Ok(..) => { // pipe I/O up to PIPE_BUF bytes should be atomic
                     assert!(p.wait().is_ok(),
                             "wait() should either return Ok or panic");
                     panic!("short read on the CLOEXEC pipe")
                 }
             }
         }

         fn combine(arr: &[u8]) -> i32 {
             let a = arr[0] as u32;
             let b = arr[1] as u32;
             let c = arr[2] as u32;
             let d = arr[3] as u32;

             ((a << 24) | (b << 16) | (c << 8) | (d << 0)) as i32
         }
    }

    pub fn exec(&mut self, default: Stdio) -> io::Error {
        if self.saw_nul {
            return io::Error::new(ErrorKind::InvalidInput,
                                  "nul byte found in provided data")
        }

        match self.setup_io(default, true) {
            Ok((_, theirs)) => unsafe { self.do_exec(theirs) },
            Err(e) => e,
        }
    }

    // And at this point we've reached a special time in the life of the
    // child. The child must now be considered hamstrung and unable to
    // do anything other than syscalls really. Consider the following
    // scenario:
    //
    //      1. Thread A of process 1 grabs the malloc() mutex
    //      2. Thread B of process 1 forks(), creating thread C
    //      3. Thread C of process 2 then attempts to malloc()
    //      4. The memory of process 2 is the same as the memory of
    //         process 1, so the mutex is locked.
    //
    // This situation looks a lot like deadlock, right? It turns out
    // that this is what pthread_atfork() takes care of, which is
    // presumably implemented across platforms. The first thing that
    // threads to *before* forking is to do things like grab the malloc
    // mutex, and then after the fork they unlock it.
    //
    // Despite this information, libnative's spawn has been witnessed to
    // deadlock on both macOS and FreeBSD. I'm not entirely sure why, but
    // all collected backtraces point at malloc/free traffic in the
    // child spawned process.
    //
    // For this reason, the block of code below should contain 0
    // invocations of either malloc of free (or their related friends).
    //
    // As an example of not having malloc/free traffic, we don't close
    // this file descriptor by dropping the FileDesc (which contains an
    // allocation). Instead we just close it manually. This will never
    // have the drop glue anyway because this code never returns (the
    // child will either exec() or invoke syscall::exit)
    unsafe fn do_exec(&mut self, stdio: ChildPipes) -> io::Error {
        macro_rules! t {
            ($e:expr) => (match $e {
                Ok(e) => e,
                Err(e) => return e,
            })
        }

        if let Some(fd) = stdio.stderr.fd() {
            t!(cvt(syscall::dup2(fd, 2, &[])));
            let mut flags = t!(cvt(syscall::fcntl(2, syscall::F_GETFD, 0)));
            flags &= ! syscall::O_CLOEXEC;
            t!(cvt(syscall::fcntl(2, syscall::F_SETFD, flags)));
        }
        if let Some(fd) = stdio.stdout.fd() {
            t!(cvt(syscall::dup2(fd, 1, &[])));
            let mut flags = t!(cvt(syscall::fcntl(1, syscall::F_GETFD, 0)));
            flags &= ! syscall::O_CLOEXEC;
            t!(cvt(syscall::fcntl(1, syscall::F_SETFD, flags)));
        }
        if let Some(fd) = stdio.stdin.fd() {
            t!(cvt(syscall::dup2(fd, 0, &[])));
            let mut flags = t!(cvt(syscall::fcntl(0, syscall::F_GETFD, 0)));
            flags &= ! syscall::O_CLOEXEC;
            t!(cvt(syscall::fcntl(0, syscall::F_SETFD, flags)));
        }

        if let Some(g) = self.gid {
            t!(cvt(syscall::setregid(g as usize, g as usize)));
        }
        if let Some(u) = self.uid {
            t!(cvt(syscall::setreuid(u as usize, u as usize)));
        }
        if let Some(ref cwd) = self.cwd {
            t!(cvt(syscall::chdir(cwd)));
        }

        for callback in self.closures.iter_mut() {
            t!(callback());
        }

        let mut args: Vec<[usize; 2]> = Vec::new();
        args.push([self.program.as_ptr() as usize, self.program.len()]);
        for arg in self.args.iter() {
            args.push([arg.as_ptr() as usize, arg.len()]);
        }

        for (key, val) in self.env.iter() {
            env::set_var(key, val);
        }

        let program = if self.program.contains(':') || self.program.contains('/') {
            Some(PathBuf::from(&self.program))
        } else if let Ok(path_env) = ::env::var("PATH") {
            let mut program = None;
            for mut path in split_paths(&path_env) {
                path.push(&self.program);
                if path.exists() {
                    program = Some(path);
                    break;
                }
            }
            program
        } else {
            None
        };

        if let Some(program) = program {
            if let Err(err) = syscall::execve(program.as_os_str().as_bytes(), &args) {
                io::Error::from_raw_os_error(err.errno as i32)
            } else {
                panic!("return from exec without err");
            }
        } else {
            io::Error::from_raw_os_error(syscall::ENOENT)
        }
    }


    fn setup_io(&self, default: Stdio, needs_stdin: bool)
                -> io::Result<(StdioPipes, ChildPipes)> {
        let null = Stdio::Null;
        let default_stdin = if needs_stdin {&default} else {&null};
        let stdin = self.stdin.as_ref().unwrap_or(default_stdin);
        let stdout = self.stdout.as_ref().unwrap_or(&default);
        let stderr = self.stderr.as_ref().unwrap_or(&default);
        let (their_stdin, our_stdin) = stdin.to_child_stdio(true)?;
        let (their_stdout, our_stdout) = stdout.to_child_stdio(false)?;
        let (their_stderr, our_stderr) = stderr.to_child_stdio(false)?;
        let ours = StdioPipes {
            stdin: our_stdin,
            stdout: our_stdout,
            stderr: our_stderr,
        };
        let theirs = ChildPipes {
            stdin: their_stdin,
            stdout: their_stdout,
            stderr: their_stderr,
        };
        Ok((ours, theirs))
    }
}

impl Stdio {
    fn to_child_stdio(&self, readable: bool)
                      -> io::Result<(ChildStdio, Option<AnonPipe>)> {
        match *self {
            Stdio::Inherit => Ok((ChildStdio::Inherit, None)),

            // Make sure that the source descriptors are not an stdio
            // descriptor, otherwise the order which we set the child's
            // descriptors may blow away a descriptor which we are hoping to
            // save. For example, suppose we want the child's stderr to be the
            // parent's stdout, and the child's stdout to be the parent's
            // stderr. No matter which we dup first, the second will get
            // overwritten prematurely.
            Stdio::Fd(ref fd) => {
                if fd.raw() <= 2 {
                    Ok((ChildStdio::Owned(fd.duplicate()?), None))
                } else {
                    Ok((ChildStdio::Explicit(fd.raw()), None))
                }
            }

            Stdio::MakePipe => {
                let (reader, writer) = pipe::anon_pipe()?;
                let (ours, theirs) = if readable {
                    (writer, reader)
                } else {
                    (reader, writer)
                };
                Ok((ChildStdio::Owned(theirs.into_fd()), Some(ours)))
            }

            Stdio::Null => {
                let mut opts = OpenOptions::new();
                opts.read(readable);
                opts.write(!readable);
                let fd = File::open(Path::new("null:"), &opts)?;
                Ok((ChildStdio::Owned(fd.into_fd()), None))
            }
        }
    }
}

impl From<AnonPipe> for Stdio {
    fn from(pipe: AnonPipe) -> Stdio {
        Stdio::Fd(pipe.into_fd())
    }
}

impl From<File> for Stdio {
    fn from(file: File) -> Stdio {
        Stdio::Fd(file.into_fd())
    }
}

impl ChildStdio {
    fn fd(&self) -> Option<usize> {
        match *self {
            ChildStdio::Inherit => None,
            ChildStdio::Explicit(fd) => Some(fd),
            ChildStdio::Owned(ref fd) => Some(fd.raw()),
        }
    }
}

impl fmt::Debug for Command {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        write!(f, "{:?}", self.program)?;
        for arg in &self.args {
            write!(f, " {:?}", arg)?;
        }
        Ok(())
    }
}

////////////////////////////////////////////////////////////////////////////////
// Processes
////////////////////////////////////////////////////////////////////////////////

/// Unix exit statuses
#[derive(PartialEq, Eq, Clone, Copy, Debug)]
pub struct ExitStatus(i32);

impl ExitStatus {
    fn exited(&self) -> bool {
        self.0 & 0x7F == 0
    }

    pub fn success(&self) -> bool {
        self.code() == Some(0)
    }

    pub fn code(&self) -> Option<i32> {
        if self.exited() {
            Some((self.0 >> 8) & 0xFF)
        } else {
            None
        }
    }

    pub fn signal(&self) -> Option<i32> {
        if !self.exited() {
            Some(self.0 & 0x7F)
        } else {
            None
        }
    }
}

impl From<i32> for ExitStatus {
    fn from(a: i32) -> ExitStatus {
        ExitStatus(a)
    }
}

impl fmt::Display for ExitStatus {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        if let Some(code) = self.code() {
            write!(f, "exit code: {}", code)
        } else {
            let signal = self.signal().unwrap();
            write!(f, "signal: {}", signal)
        }
    }
}

/// The unique id of the process (this should never be negative).
pub struct Process {
    pid: usize,
    status: Option<ExitStatus>,
}

impl Process {
    pub fn id(&self) -> u32 {
        self.pid as u32
    }

    pub fn kill(&mut self) -> io::Result<()> {
        // If we've already waited on this process then the pid can be recycled
        // and used for another process, and we probably shouldn't be killing
        // random processes, so just return an error.
        if self.status.is_some() {
            Err(Error::new(ErrorKind::InvalidInput,
                           "invalid argument: can't kill an exited process"))
        } else {
            cvt(syscall::kill(self.pid, syscall::SIGKILL))?;
            Ok(())
        }
    }

    pub fn wait(&mut self) -> io::Result<ExitStatus> {
        if let Some(status) = self.status {
            return Ok(status)
        }
        let mut status = 0;
        cvt(syscall::waitpid(self.pid, &mut status, 0))?;
        self.status = Some(ExitStatus(status as i32));
        Ok(ExitStatus(status as i32))
    }

    pub fn try_wait(&mut self) -> io::Result<Option<ExitStatus>> {
        if let Some(status) = self.status {
            return Ok(Some(status))
        }
        let mut status = 0;
        let pid = cvt(syscall::waitpid(self.pid, &mut status, syscall::WNOHANG))?;
        if pid == 0 {
            Ok(None)
        } else {
            self.status = Some(ExitStatus(status as i32));
            Ok(Some(ExitStatus(status as i32)))
        }
    }
}