// Copyright 2018-2019 the Deno authors. All rights reserved. MIT license. // Think of Resources as File Descriptors. They are integers that are allocated // by the privileged side of Deno to refer to various resources. The simplest // example are standard file system files and stdio - but there will be other // resources added in the future that might not correspond to operating system // level File Descriptors. To avoid confusion we call them "resources" not "file // descriptors". This module implements a global resource table. Ops (AKA // handlers) look up resources by their integer id here. use crate::deno_error; use crate::deno_error::bad_resource; use crate::deno_error::DenoError; use crate::deno_error::DenoResult; use crate::http_body::HttpBody; use crate::repl::Repl; use crate::state::WorkerChannels; use deno::Buf; use futures; use futures::Future; use futures::Poll; use futures::Sink; use futures::Stream; use hyper; use std; use std::collections::HashMap; use std::io::{Error, Read, Seek, SeekFrom, Write}; use std::net::{Shutdown, SocketAddr}; use std::process::ExitStatus; use std::sync::atomic::AtomicUsize; use std::sync::atomic::Ordering; use std::sync::{Arc, Mutex}; use tokio; use tokio::io::{AsyncRead, AsyncWrite}; use tokio::net::TcpStream; use tokio::sync::mpsc; use tokio_process; pub type ResourceId = u32; // Sometimes referred to RID. // These store Deno's file descriptors. These are not necessarily the operating // system ones. type ResourceTable = HashMap; #[cfg(not(windows))] use std::os::unix::io::FromRawFd; #[cfg(windows)] use std::os::windows::io::FromRawHandle; #[cfg(windows)] extern crate winapi; lazy_static! { // Starts at 3 because stdio is [0-2]. static ref NEXT_RID: AtomicUsize = AtomicUsize::new(3); static ref RESOURCE_TABLE: Mutex = Mutex::new({ let mut m = HashMap::new(); // TODO Load these lazily during lookup? m.insert(0, Repr::Stdin(tokio::io::stdin())); m.insert(1, Repr::Stdout({ #[cfg(not(windows))] let stdout = unsafe { std::fs::File::from_raw_fd(1) }; #[cfg(windows)] let stdout = unsafe { std::fs::File::from_raw_handle(winapi::um::processenv::GetStdHandle( winapi::um::winbase::STD_OUTPUT_HANDLE)) }; tokio::fs::File::from_std(stdout) })); m.insert(2, Repr::Stderr(tokio::io::stderr())); m }); } // Internal representation of Resource. enum Repr { Stdin(tokio::io::Stdin), Stdout(tokio::fs::File), Stderr(tokio::io::Stderr), FsFile(tokio::fs::File), // Since TcpListener might be closed while there is a pending accept task, // we need to track the task so that when the listener is closed, // this pending task could be notified and die. // Currently TcpListener itself does not take care of this issue. // See: https://github.com/tokio-rs/tokio/issues/846 TcpListener(tokio::net::TcpListener, Option), TcpStream(tokio::net::TcpStream), HttpBody(HttpBody), Repl(Arc>), // Enum size is bounded by the largest variant. // Use `Box` around large `Child` struct. // https://rust-lang.github.io/rust-clippy/master/index.html#large_enum_variant Child(Box), ChildStdin(tokio_process::ChildStdin), ChildStdout(tokio_process::ChildStdout), ChildStderr(tokio_process::ChildStderr), Worker(WorkerChannels), } /// If the given rid is open, this returns the type of resource, E.G. "worker". /// If the rid is closed or was never open, it returns None. pub fn get_type(rid: ResourceId) -> Option { let table = RESOURCE_TABLE.lock().unwrap(); table.get(&rid).map(inspect_repr) } pub fn table_entries() -> Vec<(u32, String)> { let table = RESOURCE_TABLE.lock().unwrap(); table .iter() .map(|(key, value)| (*key, inspect_repr(&value))) .collect() } #[test] fn test_table_entries() { let mut entries = table_entries(); entries.sort(); assert_eq!(entries[0], (0, String::from("stdin"))); assert_eq!(entries[1], (1, String::from("stdout"))); assert_eq!(entries[2], (2, String::from("stderr"))); } fn inspect_repr(repr: &Repr) -> String { let h_repr = match repr { Repr::Stdin(_) => "stdin", Repr::Stdout(_) => "stdout", Repr::Stderr(_) => "stderr", Repr::FsFile(_) => "fsFile", Repr::TcpListener(_, _) => "tcpListener", Repr::TcpStream(_) => "tcpStream", Repr::HttpBody(_) => "httpBody", Repr::Repl(_) => "repl", Repr::Child(_) => "child", Repr::ChildStdin(_) => "childStdin", Repr::ChildStdout(_) => "childStdout", Repr::ChildStderr(_) => "childStderr", Repr::Worker(_) => "worker", }; String::from(h_repr) } // Abstract async file interface. // Ideally in unix, if Resource represents an OS rid, it will be the same. #[derive(Clone, Debug)] pub struct Resource { pub rid: ResourceId, } impl Resource { // TODO Should it return a Resource instead of net::TcpStream? pub fn poll_accept(&mut self) -> Poll<(TcpStream, SocketAddr), Error> { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&self.rid); match maybe_repr { None => Err(std::io::Error::new( std::io::ErrorKind::Other, "Listener has been closed", )), Some(repr) => match repr { Repr::TcpListener(ref mut s, _) => s.poll_accept(), _ => panic!("Cannot accept"), }, } } /// Track the current task (for TcpListener resource). /// Throws an error if another task is already tracked. pub fn track_task(&mut self) -> Result<(), std::io::Error> { let mut table = RESOURCE_TABLE.lock().unwrap(); // Only track if is TcpListener. if let Some(Repr::TcpListener(_, t)) = table.get_mut(&self.rid) { // Currently, we only allow tracking a single accept task for a listener. // This might be changed in the future with multiple workers. // Caveat: TcpListener by itself also only tracks an accept task at a time. // See https://github.com/tokio-rs/tokio/issues/846#issuecomment-454208883 if t.is_some() { return Err(std::io::Error::new( std::io::ErrorKind::Other, "Another accept task is ongoing", )); } t.replace(futures::task::current()); } Ok(()) } /// Stop tracking a task (for TcpListener resource). /// Happens when the task is done and thus no further tracking is needed. pub fn untrack_task(&mut self) { let mut table = RESOURCE_TABLE.lock().unwrap(); // Only untrack if is TcpListener. if let Some(Repr::TcpListener(_, t)) = table.get_mut(&self.rid) { if t.is_some() { t.take(); } } } // close(2) is done by dropping the value. Therefore we just need to remove // the resource from the RESOURCE_TABLE. pub fn close(&self) { let mut table = RESOURCE_TABLE.lock().unwrap(); let r = table.remove(&self.rid).unwrap(); // If TcpListener, we must kill all pending accepts! if let Repr::TcpListener(_, Some(t)) = r { // Call notify on the tracked task, so that they would error out. t.notify(); } } pub fn shutdown(&mut self, how: Shutdown) -> Result<(), DenoError> { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&self.rid); match maybe_repr { None => panic!("bad rid"), Some(repr) => match repr { Repr::TcpStream(ref mut f) => { TcpStream::shutdown(f, how).map_err(DenoError::from) } _ => panic!("Cannot shutdown"), }, } } } impl Read for Resource { fn read(&mut self, _buf: &mut [u8]) -> std::io::Result { unimplemented!(); } } impl AsyncRead for Resource { fn poll_read(&mut self, buf: &mut [u8]) -> Poll { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&self.rid); match maybe_repr { None => panic!("bad rid"), Some(repr) => match repr { Repr::FsFile(ref mut f) => f.poll_read(buf), Repr::Stdin(ref mut f) => f.poll_read(buf), Repr::TcpStream(ref mut f) => f.poll_read(buf), Repr::HttpBody(ref mut f) => f.poll_read(buf), Repr::ChildStdout(ref mut f) => f.poll_read(buf), Repr::ChildStderr(ref mut f) => f.poll_read(buf), _ => panic!("Cannot read"), }, } } } impl Write for Resource { fn write(&mut self, _buf: &[u8]) -> std::io::Result { unimplemented!() } fn flush(&mut self) -> std::io::Result<()> { unimplemented!() } } impl AsyncWrite for Resource { fn poll_write(&mut self, buf: &[u8]) -> Poll { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&self.rid); match maybe_repr { None => panic!("bad rid"), Some(repr) => match repr { Repr::FsFile(ref mut f) => f.poll_write(buf), Repr::Stdout(ref mut f) => f.poll_write(buf), Repr::Stderr(ref mut f) => f.poll_write(buf), Repr::TcpStream(ref mut f) => f.poll_write(buf), Repr::ChildStdin(ref mut f) => f.poll_write(buf), _ => panic!("Cannot write"), }, } } fn shutdown(&mut self) -> futures::Poll<(), std::io::Error> { unimplemented!() } } fn new_rid() -> ResourceId { let next_rid = NEXT_RID.fetch_add(1, Ordering::SeqCst); next_rid as ResourceId } pub fn add_fs_file(fs_file: tokio::fs::File) -> Resource { let rid = new_rid(); let mut tg = RESOURCE_TABLE.lock().unwrap(); match tg.insert(rid, Repr::FsFile(fs_file)) { Some(_) => panic!("There is already a file with that rid"), None => Resource { rid }, } } pub fn add_tcp_listener(listener: tokio::net::TcpListener) -> Resource { let rid = new_rid(); let mut tg = RESOURCE_TABLE.lock().unwrap(); let r = tg.insert(rid, Repr::TcpListener(listener, None)); assert!(r.is_none()); Resource { rid } } pub fn add_tcp_stream(stream: tokio::net::TcpStream) -> Resource { let rid = new_rid(); let mut tg = RESOURCE_TABLE.lock().unwrap(); let r = tg.insert(rid, Repr::TcpStream(stream)); assert!(r.is_none()); Resource { rid } } pub fn add_hyper_body(body: hyper::Body) -> Resource { let rid = new_rid(); let mut tg = RESOURCE_TABLE.lock().unwrap(); let body = HttpBody::from(body); let r = tg.insert(rid, Repr::HttpBody(body)); assert!(r.is_none()); Resource { rid } } pub fn add_repl(repl: Repl) -> Resource { let rid = new_rid(); let mut tg = RESOURCE_TABLE.lock().unwrap(); let r = tg.insert(rid, Repr::Repl(Arc::new(Mutex::new(repl)))); assert!(r.is_none()); Resource { rid } } pub fn add_worker(wc: WorkerChannels) -> Resource { let rid = new_rid(); let mut tg = RESOURCE_TABLE.lock().unwrap(); let r = tg.insert(rid, Repr::Worker(wc)); assert!(r.is_none()); Resource { rid } } /// Post message to worker as a host or privilged overlord pub fn post_message_to_worker( rid: ResourceId, buf: Buf, ) -> futures::sink::Send> { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&rid); match maybe_repr { Some(Repr::Worker(ref mut wc)) => { // unwrap here is incorrect, but doing it anyway wc.0.clone().send(buf) } _ => panic!("bad resource"), // futures::future::err(bad_resource()).into(), } } pub struct WorkerReceiver { rid: ResourceId, } // Invert the dumbness that tokio_process causes by making Child itself a future. impl Future for WorkerReceiver { type Item = Option; type Error = DenoError; fn poll(&mut self) -> Poll, DenoError> { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&self.rid); match maybe_repr { Some(Repr::Worker(ref mut wc)) => wc.1.poll().map_err(|err| { deno_error::new(deno_error::ErrorKind::Other, err.to_string()) }), _ => Err(bad_resource()), } } } pub fn get_message_from_worker(rid: ResourceId) -> WorkerReceiver { WorkerReceiver { rid } } pub struct WorkerReceiverStream { rid: ResourceId, } // Invert the dumbness that tokio_process causes by making Child itself a future. impl Stream for WorkerReceiverStream { type Item = Buf; type Error = DenoError; fn poll(&mut self) -> Poll, DenoError> { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&self.rid); match maybe_repr { Some(Repr::Worker(ref mut wc)) => wc.1.poll().map_err(|err| { deno_error::new(deno_error::ErrorKind::Other, err.to_string()) }), _ => Err(bad_resource()), } } } pub fn get_message_stream_from_worker(rid: ResourceId) -> WorkerReceiverStream { WorkerReceiverStream { rid } } pub struct ChildResources { pub child_rid: ResourceId, pub stdin_rid: Option, pub stdout_rid: Option, pub stderr_rid: Option, } pub fn add_child(mut c: tokio_process::Child) -> ChildResources { let child_rid = new_rid(); let mut tg = RESOURCE_TABLE.lock().unwrap(); let mut resources = ChildResources { child_rid, stdin_rid: None, stdout_rid: None, stderr_rid: None, }; if c.stdin().is_some() { let stdin = c.stdin().take().unwrap(); let rid = new_rid(); let r = tg.insert(rid, Repr::ChildStdin(stdin)); assert!(r.is_none()); resources.stdin_rid = Some(rid); } if c.stdout().is_some() { let stdout = c.stdout().take().unwrap(); let rid = new_rid(); let r = tg.insert(rid, Repr::ChildStdout(stdout)); assert!(r.is_none()); resources.stdout_rid = Some(rid); } if c.stderr().is_some() { let stderr = c.stderr().take().unwrap(); let rid = new_rid(); let r = tg.insert(rid, Repr::ChildStderr(stderr)); assert!(r.is_none()); resources.stderr_rid = Some(rid); } let r = tg.insert(child_rid, Repr::Child(Box::new(c))); assert!(r.is_none()); resources } pub struct ChildStatus { rid: ResourceId, } // Invert the dumbness that tokio_process causes by making Child itself a future. impl Future for ChildStatus { type Item = ExitStatus; type Error = DenoError; fn poll(&mut self) -> Poll { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&self.rid); match maybe_repr { Some(Repr::Child(ref mut child)) => child.poll().map_err(DenoError::from), _ => Err(bad_resource()), } } } pub fn child_status(rid: ResourceId) -> DenoResult { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&rid); match maybe_repr { Some(Repr::Child(ref mut _child)) => Ok(ChildStatus { rid }), _ => Err(bad_resource()), } } pub fn get_repl(rid: ResourceId) -> DenoResult>> { let mut table = RESOURCE_TABLE.lock().unwrap(); let maybe_repr = table.get_mut(&rid); match maybe_repr { Some(Repr::Repl(ref mut r)) => Ok(r.clone()), _ => Err(bad_resource()), } } // TODO: revamp this after the following lands: // https://github.com/tokio-rs/tokio/pull/785 pub fn get_file(rid: ResourceId) -> DenoResult { let mut table = RESOURCE_TABLE.lock().unwrap(); // We take ownership of File here. // It is put back below while still holding the lock. let maybe_repr = table.remove(&rid); match maybe_repr { Some(Repr::FsFile(r)) => { // Trait Clone not implemented on tokio::fs::File, // so convert to std File first. let std_file = r.into_std(); // Create a copy and immediately put back. // We don't want to block other resource ops. // try_clone() would yield a copy containing the same // underlying fd, so operations on the copy would also // affect the one in resource table, and we don't need // to write back. let maybe_std_file_copy = std_file.try_clone(); // Insert the entry back with the same rid. table.insert(rid, Repr::FsFile(tokio_fs::File::from_std(std_file))); if maybe_std_file_copy.is_err() { return Err(DenoError::from(maybe_std_file_copy.unwrap_err())); } let std_file_copy = maybe_std_file_copy.unwrap(); Ok(std_file_copy) } _ => Err(bad_resource()), } } pub fn lookup(rid: ResourceId) -> Option { debug!("resource lookup {}", rid); let table = RESOURCE_TABLE.lock().unwrap(); table.get(&rid).map(|_| Resource { rid }) } pub fn seek( resource: Resource, offset: i32, whence: u32, ) -> Box + Send> { // Translate seek mode to Rust repr. let seek_from = match whence { 0 => SeekFrom::Start(offset as u64), 1 => SeekFrom::Current(i64::from(offset)), 2 => SeekFrom::End(i64::from(offset)), _ => { return Box::new(futures::future::err(deno_error::new( deno_error::ErrorKind::InvalidSeekMode, format!("Invalid seek mode: {}", whence), ))); } }; match get_file(resource.rid) { Ok(mut file) => Box::new(futures::future::lazy(move || { let result = file.seek(seek_from).map(|_| {}).map_err(DenoError::from); futures::future::result(result) })), Err(err) => Box::new(futures::future::err(err)), } }