Cargo: Up version to v0.0.2

Signed-off-by: Levente Kurusa <lkurusa@acm.org>

.travis.yml

@@ -0,0 +1,11 @@
+language: rust
+rust:
+  - stable
+  - beta
+  - nightly
+matrix:
+  allow_failures:
+    - rust: nightly
+  fast_finish: true
+script:
+  - cargo build --verbose --all

Cargo.toml

@@ -4,8 +4,12 @@ description = "Native Rust crate for managing control groups on Linux"
 repository = "https://github.com/levex/cgroups-rs"
 keywords = ["linux", "cgroup", "containers", "isolation"]
 categories = ["os", "api-bindings", "os::unix-apis"]
-license = "BSD-2-Clause OR Apache-2.0"
+license = "MIT OR Apache-2.0"
-version = "0.0.1"
+version = "0.0.2"
 authors = ["Levente Kurusa <lkurusa@acm.org>"]
 
 [dependencies]
+
+[dev-dependencies]
+nix = "0.11.0"
+libc = "0.2.43"

LICENSE

@@ -1,21 +1,6 @@
-MIT License
+This crate is licensed under either of
 
-Copyright (c) 2018 Levente Kurusa
+- "Apache License, Version 2.0, (See LICENSE-Apache-2.0 file); or
+- "MIT license" (See LICENSE-MIT file), 
 
-Permission is hereby granted, free of charge, to any person obtaining a copy
+at your option.
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.

LICENSE-Apache-2.0

@@ -0,0 +1,13 @@
+Copyright 2018 Levente Kurusa <lkurusa@acm.org>
+
+Licensed under the Apache License, Version 2.0 (the "License"); 
+you may not use this file except in compliance with the License. 
+You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software 
+distributed under the License is distributed on an "AS IS" BASIS, 
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 
+See the License for the specific language governing permissions and 
+limitations under the License.

LICENSE-MIT

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2018 Levente Kurusa
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -1,2 +1,38 @@
-# cgroups-rs
+# cgroups-rs ![Build](https://travis-ci.org/levex/cgroups-rs.svg?branch=master)
 Native Rust library for managing control groups under Linux
+
+# Example
+
+## Create a control group, and limit the pid resource
+
+``` rust
+// Acquire a handle for the V1 cgroup hierarchy.
+let hier = ::hierarchies::V1::new();
+// Create a control group named "example" in the hierarchy.
+let cg = Cgroup::new(&hier, String::from("example"), 0);
+{
+    // Get a handle to the pids controller of the control group.
+    let pids: &PidController = cg.controller_of().expect("No pids controller in V1 hierarchy!");
+    // Set the maximum amount of processes in the cgroup.
+    pids.set_pid_max(PidMax::Value(10));
+    // Check that this has had the desired effect by reading the value back from the kernel.
+    assert_eq!(pids.get_pid_max(), Some(PidMax::Value(10)));
+}
+// Once done, delete the control group (and its associated controllers).
+cg.delete();
+```
+
+# Disclaimer
+
+This crate is licensed under:
+
+- MIT License (see LICENSE-MIT); or
+- Apache 2.0 LIcense (see LICENSE-Apache-2.0),
+
+at your option.
+
+Please note that this crate is under heavy development, we will use sematic
+versioning, but during the `0.0.*` phase, no guarantees are made about
+backwards compatibility.
+
+Regardless, check back often and thanks for taking a look!

src/blkio.rs

@@ -6,7 +6,8 @@ use std::path::PathBuf;
 use std::io::{Read, Write};
 use std::fs::File;
 
-use {BlkIoResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use {CgroupError, BlkIoResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use CgroupError::*;
 
 /// A controller that allows controlling the `blkio` subsystem of a Cgroup.
 ///
@@ -105,28 +106,28 @@ impl Controller for BlkIoController {
         let res: &BlkIoResources = &res.blkio;
 
         if res.update_values {
-            self.set_weight(res.weight as u64);
+            let _ = self.set_weight(res.weight as u64);
-            self.set_leaf_weight(res.leaf_weight as u64);
+            let _ = self.set_leaf_weight(res.leaf_weight as u64);
 
             for dev in &res.weight_device {
-                self.set_weight_for_device(format!("{}:{} {}",
+                let _ = self.set_weight_for_device(format!("{}:{} {}",
                                 dev.major, dev.minor, dev.weight));
             }
 
             for dev in &res.throttle_read_bps_device {
-                self.throttle_read_bps_for_device(dev.major, dev.minor, dev.rate);
+                let _ = self.throttle_read_bps_for_device(dev.major, dev.minor, dev.rate);
             }
 
             for dev in &res.throttle_write_bps_device {
-                self.throttle_write_bps_for_device(dev.major, dev.minor, dev.rate);
+                let _ = self.throttle_write_bps_for_device(dev.major, dev.minor, dev.rate);
             }
 
             for dev in &res.throttle_read_iops_device {
-                self.throttle_read_iops_for_device(dev.major, dev.minor, dev.rate);
+                let _ = self.throttle_read_iops_for_device(dev.major, dev.minor, dev.rate);
             }
 
             for dev in &res.throttle_write_iops_device {
-                self.throttle_write_iops_for_device(dev.major, dev.minor, dev.rate);
+                let _ = self.throttle_write_iops_for_device(dev.major, dev.minor, dev.rate);
             }
         }
     }
@@ -152,16 +153,20 @@ impl<'a> From<&'a Subsystem> for &'a BlkIoController {
     }
 }
 
-fn read_string_from(mut file: File) -> Option<String> {
+fn read_string_from(mut file: File) -> Result<String, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    match file.read_to_string(&mut string) {
-    Some(string.trim().to_string())
+        Ok(_) => Ok(string.trim().to_string()),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }
 
-fn read_u64_from(mut file: File) -> Option<u64> {
+fn read_u64_from(mut file: File) -> Result<u64, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    match file.read_to_string(&mut string) {
-    string.trim().parse().ok()
+        Ok(_) => string.trim().parse().map_err(|_| ParseError),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }
 
 impl BlkIoController {
@@ -270,68 +275,68 @@ impl BlkIoController {
 
     /// Set the leaf weight on the control group's tasks, i.e., how are they weighted against the
     /// descendant control groups' tasks.
-    pub fn set_leaf_weight(self: &Self, w: u64) {
+    pub fn set_leaf_weight(self: &Self, w: u64) -> Result<(), CgroupError> {
         self.open_path("blkio.leaf_weight", true).and_then(|mut file| {
-            file.write_all(w.to_string().as_ref()).ok()
+            file.write_all(w.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Same as `set_leaf_weight()`, but settable per each block device.
-    pub fn set_leaf_weight_for_device(self: &Self, d: String) {
+    pub fn set_leaf_weight_for_device(self: &Self, d: String) -> Result<(), CgroupError> {
         self.open_path("blkio.leaf_weight_device", true).and_then(|mut file| {
-            file.write_all(d.as_ref()).ok()
+            file.write_all(d.as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Reset the statistics the kernel has gathered so far and start fresh.
-    pub fn reset_stats(self: &Self) {
+    pub fn reset_stats(self: &Self) -> Result<(), CgroupError> {
         self.open_path("blkio.leaf_weight_device", true).and_then(|mut file| {
-            file.write_all("1".to_string().as_ref()).ok()
+            file.write_all("1".to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Throttle the bytes per second rate of read operation affecting the block device
     /// `major:minor` to `bps`.
-    pub fn throttle_read_bps_for_device(self: &Self, major: u64, minor: u64, bps: u64) {
+    pub fn throttle_read_bps_for_device(self: &Self, major: u64, minor: u64, bps: u64) -> Result<(), CgroupError> {
         self.open_path("blkio.throttle.read_bps_device", true).and_then(|mut file| {
-            file.write_all(format!("{}:{} {}", major, minor, bps).to_string().as_ref()).ok()
+            file.write_all(format!("{}:{} {}", major, minor, bps).to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Throttle the I/O operations per second rate of read operation affecting the block device
     /// `major:minor` to `bps`.
-    pub fn throttle_read_iops_for_device(self: &Self, major: u64, minor: u64, iops: u64) {
+    pub fn throttle_read_iops_for_device(self: &Self, major: u64, minor: u64, iops: u64) -> Result<(), CgroupError> {
         self.open_path("blkio.throttle.read_iops_device", true).and_then(|mut file| {
-            file.write_all(format!("{}:{} {}", major, minor, iops).to_string().as_ref()).ok()
+            file.write_all(format!("{}:{} {}", major, minor, iops).to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
     /// Throttle the bytes per second rate of write operation affecting the block device
     /// `major:minor` to `bps`.
-    pub fn throttle_write_bps_for_device(self: &Self, major: u64, minor: u64, bps: u64) {
+    pub fn throttle_write_bps_for_device(self: &Self, major: u64, minor: u64, bps: u64) -> Result<(), CgroupError> {
         self.open_path("blkio.throttle.write_bps_device", true).and_then(|mut file| {
-            file.write_all(format!("{}:{} {}", major, minor, bps).to_string().as_ref()).ok()
+            file.write_all(format!("{}:{} {}", major, minor, bps).to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Throttle the I/O operations per second rate of write operation affecting the block device
     /// `major:minor` to `bps`.
-    pub fn throttle_write_iops_for_device(self: &Self, major: u64, minor: u64, iops: u64) {
+    pub fn throttle_write_iops_for_device(self: &Self, major: u64, minor: u64, iops: u64) -> Result<(), CgroupError> {
         self.open_path("blkio.throttle.write_iops_device", true).and_then(|mut file| {
-            file.write_all(format!("{}:{} {}", major, minor, iops).to_string().as_ref()).ok()
+            file.write_all(format!("{}:{} {}", major, minor, iops).to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Set the weight of the control group's tasks.
-    pub fn set_weight(self: &Self, w: u64) {
+    pub fn set_weight(self: &Self, w: u64) -> Result<(), CgroupError> {
         self.open_path("blkio.leaf_weight", true).and_then(|mut file| {
-            file.write_all(w.to_string().as_ref()).ok()
+            file.write_all(w.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Same as `set_weight()`, but settable per each block device.
-    pub fn set_weight_for_device(self: &Self, d: String) {
+    pub fn set_weight_for_device(self: &Self, d: String) -> Result<(), CgroupError> {
         self.open_path("blkio.weight_device", true).and_then(|mut file| {
-            file.write_all(d.as_ref()).ok()
+            file.write_all(d.as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 }

src/cgroup.rs

@@ -1,6 +1,6 @@
 //! This module handles cgroup operations. Start here!
 
-use {CgroupPid, Resources, ControllIdentifier, Controller, Hierarchy, Subsystem};
+use {CgroupError, CgroupPid, Resources, ControllIdentifier, Controller, Hierarchy, Subsystem};
 
 use std::convert::From;
 
@@ -17,12 +17,15 @@ use std::convert::From;
 /// > specialized behaviour.
 ///
 /// This crate is an attempt at providing a Rust-native way of managing these cgroups.
-pub struct Cgroup {
+pub struct Cgroup<'b> {
     /// The list of subsystems that control this cgroup
     subsystems: Vec<Subsystem>,
+
+    /// The hierarchy.
+    hier: &'b Hierarchy,
 }
 
-impl Cgroup {
+impl<'b> Cgroup<'b> {
 
     /// Create this control group.
     fn create(self: &Self) {
@@ -38,15 +41,29 @@ impl Cgroup {
     /// Note that if the handle goes out of scope and is dropped, the control group is _not_
     /// destroyed.
     pub fn new(hier: &Hierarchy, path: String) -> Cgroup {
+        let cg = Cgroup::load(hier, path);
+        cg.create();
+        cg
+    }
+
+    /// Create a handle for a control group in the hierarchy `hier`, with name `path`.
+    ///
+    /// Returns a handle to the control group (that possibly does not exist until `create()` has
+    /// been called on the cgroup.
+    ///
+    /// Note that if the handle goes out of scope and is dropped, the control group is _not_
+    /// destroyed.
+    pub fn load(hier: &Hierarchy, path: String) -> Cgroup {
         let mut subsystems = hier.subsystems();
-        subsystems = subsystems.into_iter().map(|x| x.enter(&path)).collect::<Vec<_>>();
+        if path != ""  {
+            subsystems = subsystems.into_iter().map(|x| x.enter(&path)).collect::<Vec<_>>();
+        }
 
         let cg = Cgroup {
-            //name: path,
             subsystems: subsystems,
+            hier: hier,
         };
 
-        cg.create();
         cg
     }
 
@@ -92,7 +109,7 @@ impl Cgroup {
     ///
     /// ## Example:
     ///
-    /// ```
+    /// ```text
     /// let pids: &PidController = control_group.controller_of()
     ///                             .expect("No pids controller attached!");
     /// let cpu: &CpuController = control_group.controller_of()
@@ -114,8 +131,29 @@ impl Cgroup {
         None
     }
 
-    /// Attach a task to the control group.
+    /// Removes a task from the control group.
-    pub fn add_task(self: &Self, pid: CgroupPid) {
+    ///
-        self.subsystems().iter().for_each(|sub| sub.to_controller().add_task(&pid));
+    /// Note that this means that the task will be moved back to the root control group in the
+    /// hierarchy and any rules applied to that control group will _still_ apply to the task.
+    pub fn remove_task(self: &Self, pid: CgroupPid) {
+        let _ = self.hier.root_control_group().add_task(pid);
     }
+
+    /// Attach a task to the control group.
+    pub fn add_task(self: &Self, pid: CgroupPid) -> Result<(), CgroupError> {
+        self.subsystems().iter().try_for_each(|sub| sub.to_controller().add_task(&pid))
+    }
+
+    /// Returns an Iterator that can be used to iterate over the tasks that are currently in the
+    /// control group.
+    pub fn tasks(self: &Self) -> Vec<CgroupPid> {
+        /* Collect the tasks from all subsystems */
+        let mut v = self.subsystems().iter()
+            .map(|x| x.to_controller().tasks())
+            .fold(vec![], |mut acc, mut x| { acc.append(&mut x); acc });
+        v.sort();
+        v.dedup();
+        v
+    }
+
 }

src/cpu.rs

@@ -6,7 +6,7 @@
 use std::path::PathBuf;
 use std::io::{Read, Write};
 
-use {CpuResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use {CgroupError, CpuResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
 
 /// A controller that allows controlling the `cpu` subsystem of a Cgroup.
 /// 
@@ -40,9 +40,9 @@ impl Controller for CpuController {
 
         if res.update_values {
             /* apply pid_max */
-            self.set_shares(res.shares);
+            let _ = self.set_shares(res.shares);
-            self.set_cfs_period(res.period);
+            let _ = self.set_cfs_period(res.period);
-            self.set_cfs_quota(res.quota as u64);
+            let _ = self.set_cfs_quota(res.quota as u64);
             /* TODO: rt properties (CONFIG_RT_GROUP_SCHED) are not yet supported */
         }
     }
@@ -84,8 +84,11 @@ impl CpuController {
         Cpu {
             stat: self.open_path("cpu.stat", false).and_then(|mut file| {
                 let mut s = String::new();
-                let _ = file.read_to_string(&mut s);
+                let res = file.read_to_string(&mut s);
-                Some(s)
+                match res {
+                    Ok(_) => Ok(s),
+                    Err(e) => Err(CgroupError::ReadError(e)),
+                }
             }).unwrap_or("".to_string()),
         }
     }
@@ -96,25 +99,25 @@ impl CpuController {
     /// For example, setting control group `A`'s `shares` to `100`, and control group `B`'s
     /// `shares` to `200` ensures that control group `B` receives twice as much as CPU bandwidth.
     /// (Assuming both `A` and `B` are of the same parent)
-    pub fn set_shares(self: &Self, shares: u64) {
+    pub fn set_shares(self: &Self, shares: u64) -> Result<(), CgroupError> {
         self.open_path("cpu.shares", true).and_then(|mut file| {
-            file.write_all(shares.to_string().as_ref()).ok()
+            file.write_all(shares.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Specify a period (when using the CFS scheduler) of time in microseconds for how often this
     /// control group's access to the CPU should be reallocated.
-    pub fn set_cfs_period(self: &Self, us: u64) {
+    pub fn set_cfs_period(self: &Self, us: u64) -> Result<(), CgroupError> {
         self.open_path("cpu.cfs_period_us", true).and_then(|mut file| {
-            file.write_all(us.to_string().as_ref()).ok()
+            file.write_all(us.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Specify a quota (when using the CFS scheduler) of time in microseconds for which all tasks
     /// in this control group can run during one period (see: `set_cfs_period()`).
-    pub fn set_cfs_quota(self: &Self, us: u64) {
+    pub fn set_cfs_quota(self: &Self, us: u64) -> Result<(), CgroupError> {
         self.open_path("cpu.cfs_quota_us", true).and_then(|mut file| {
-            file.write_all(us.to_string().as_ref()).ok()
+            file.write_all(us.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 }

src/cpuacct.rs

@@ -6,7 +6,7 @@ use std::path::PathBuf;
 use std::io::{Read, Write};
 use std::fs::File;
 
-use {Controllers, Resources, Subsystem, ControllIdentifier, Controller};
+use {CgroupError, Controllers, Resources, Subsystem, ControllIdentifier, Controller};
 
 /// A controller that allows controlling the `cpuacct` subsystem of a Cgroup.
 ///
@@ -28,7 +28,7 @@ pub struct CpuAcct {
     /// time spent is `user` time or `system` time.
     ///
     /// An example is as follows:
-    /// ```
+    /// ```text
     /// cpu user system
     /// 0 8348363768 0
     /// 1 8324369100 0
@@ -79,10 +79,24 @@ impl<'a> From<&'a Subsystem> for &'a CpuAcctController {
     }
 }
 
-fn read_u64_from(mut file: File) -> Option<u64> {
+fn read_u64_from(mut file: File) -> Result<u64, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    let res = file.read_to_string(&mut string);
-    string.trim().parse().ok()
+    match res {
+        Ok(_) => match string.trim().parse() {
+            Ok(e) => Ok(e),
+            Err(_) => Err(CgroupError::ParseError),
+        },
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
+}
+
+fn read_string_from(mut file: File) -> Result<String, CgroupError> {
+    let mut string = String::new();
+    match file.read_to_string(&mut string) {
+        Ok(_) => Ok(string.trim().to_string()),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }
 
 impl CpuAcctController {
@@ -101,38 +115,18 @@ impl CpuAcctController {
     pub fn cpuacct(self: &Self) -> CpuAcct {
         CpuAcct {
             stat: self.open_path("cpuacct.stat", false)
-                    .and_then(|mut file| {
+                    .and_then(|file| read_string_from(file)).unwrap_or("".to_string()),
-                        let mut string = String::new();
-                        let _ = file.read_to_string(&mut string);
-                        Some(string.trim().to_string())
-                    }).unwrap_or("".to_string()),
             usage: self.open_path("cpuacct.usage", false)
                     .and_then(|file| read_u64_from(file))
                     .unwrap_or(0),
             usage_all: self.open_path("cpuacct.usage_all", false)
-                    .and_then(|mut file| {
+                    .and_then(|file| read_string_from(file)).unwrap_or("".to_string()),
-                        let mut string = String::new();
-                        let _ = file.read_to_string(&mut string);
-                        Some(string.trim().to_string())
-                    }).unwrap_or("".to_string()),
             usage_percpu: self.open_path("cpuacct.usage_percpu", false)
-                    .and_then(|mut file| {
+                    .and_then(|file| read_string_from(file)).unwrap_or("".to_string()),
-                        let mut string = String::new();
-                        let _ = file.read_to_string(&mut string);
-                        Some(string.trim().to_string())
-                    }).unwrap_or("".to_string()),
             usage_percpu_sys: self.open_path("cpuacct.usage_percpu_sys", false)
-                    .and_then(|mut file| {
+                    .and_then(|file| read_string_from(file)).unwrap_or("".to_string()),
-                        let mut string = String::new();
-                        let _ = file.read_to_string(&mut string);
-                        Some(string.trim().to_string())
-                    }).unwrap_or("".to_string()),
             usage_percpu_user: self.open_path("cpuacct.usage_percpu_user", false)
-                    .and_then(|mut file| {
+                    .and_then(|file| read_string_from(file)).unwrap_or("".to_string()),
-                        let mut string = String::new();
-                        let _ = file.read_to_string(&mut string);
-                        Some(string.trim().to_string())
-                    }).unwrap_or("".to_string()),
             usage_sys: self.open_path("cpuacct.usage_sys", false)
                     .and_then(|file| read_u64_from(file))
                     .unwrap_or(0),
@@ -143,9 +137,9 @@ impl CpuAcctController {
     }
 
     /// Reset the statistics the kernel has gathered about the control group.
-    pub fn reset(self: &Self) {
+    pub fn reset(self: &Self) -> Result<(), CgroupError> {
         self.open_path("cpuacct.usage", true).and_then(|mut file| {
-            file.write_all(b"0").ok()
+            file.write_all(b"0").map_err(CgroupError::WriteError)
-        });
+        })
     }
 }

src/cpuset.rs

@@ -6,7 +6,8 @@ use std::path::PathBuf;
 use std::io::{Read, Write};
 use std::fs::File;
 
-use {CpuResources, Resources, Controller, ControllIdentifier, Subsystem, Controllers};
+use {CgroupError, CpuResources, Resources, Controller, ControllIdentifier, Subsystem, Controllers};
+use CgroupError::*;
 
 /// A controller that allows controlling the `cpuset` subsystem of a Cgroup.
 /// 
@@ -84,8 +85,8 @@ impl Controller for CpuSetController {
 
         if res.update_values {
             /* apply pid_max */
-            self.set_cpus(&res.cpus);
+            let _ = self.set_cpus(&res.cpus);
-            self.set_mems(&res.mems);
+            let _ = self.set_mems(&res.mems);
         }
     }
 }
@@ -110,10 +111,20 @@ impl<'a> From<&'a Subsystem> for &'a CpuSetController {
     }
 }
 
-fn read_u64_from(mut file: File) -> Option<u64> {
+fn read_string_from(mut file: File) -> Result<String, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    match file.read_to_string(&mut string) {
-    string.trim().parse().ok()
+        Ok(_) => Ok(string.trim().to_string()),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
+}
+
+fn read_u64_from(mut file: File) -> Result<u64, CgroupError> {
+    let mut string = String::new();
+    match file.read_to_string(&mut string) {
+        Ok(_) => string.trim().parse().map_err(|_| ParseError),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }
 
 impl CpuSetController {
@@ -137,122 +148,96 @@ impl CpuSetController {
                 }).map(|x| x == 1).unwrap_or(false)
             },
             cpus: {
-                self.open_path("cpuset.cpus", false).and_then(|mut file| {
+                self.open_path("cpuset.cpus", false).and_then(read_string_from).unwrap_or("".to_string())
-                    let mut string = String::new();
-                    let _ = file.read_to_string(&mut string);
-                    Some(string.trim().to_string())
-                }).unwrap()
             },
             effective_cpus: {
-                self.open_path("cpuset.effective_cpus", false).and_then(|mut file| {
+                self.open_path("cpuset.effective_cpus", false).and_then(read_string_from).unwrap_or("".to_string())
-                    let mut string = String::new();
-                    let _ = file.read_to_string(&mut string);
-                    Some(string.trim().to_string())
-                }).unwrap()
             },
             effective_mems: {
-                self.open_path("cpuset.effective_mems", false).and_then(|mut file| {
+                self.open_path("cpuset.effective_mems", false).and_then(read_string_from).unwrap_or("".to_string())
-                    let mut string = String::new();
-                    let _ = file.read_to_string(&mut string);
-                    Some(string.trim().to_string())
-                }).unwrap()
             },
             mem_exclusive: {
-                self.open_path("cpuset.mem_exclusive", false).and_then(|file| {
+                self.open_path("cpuset.mem_exclusive", false).and_then(read_u64_from)
-                    read_u64_from(file)
+                    .map(|x| x == 1).unwrap_or(false)
-                }).map(|x| x == 1).unwrap_or(false)
             },
             mem_hardwall: {
-                self.open_path("cpuset.mem_hardwall", false).and_then(|file| {
+                self.open_path("cpuset.mem_hardwall", false).and_then(read_u64_from)
-                    read_u64_from(file)
+                    .map(|x| x == 1).unwrap_or(false)
-                }).map(|x| x == 1).unwrap_or(false)
             },
             memory_migrate: {
-                self.open_path("cpuset.memory_migrate", false).and_then(|file| {
+                self.open_path("cpuset.memory_migrate", false).and_then(read_u64_from)
-                    read_u64_from(file)
+                    .map(|x| x == 1).unwrap_or(false)
-                }).map(|x| x == 1).unwrap_or(false)
             },
             memory_pressure: {
-                self.open_path("cpuset.memory_pressure", false).and_then(|file| {
+                self.open_path("cpuset.memory_pressure", false).and_then(read_u64_from).unwrap_or(0)
-                    read_u64_from(file)
-                }).unwrap_or(0)
             },
             memory_pressure_enabled: {
-                self.open_path("cpuset.memory_pressure_enabled", false).and_then(|file| {
+                self.open_path("cpuset.memory_pressure_enabled", false).and_then(read_u64_from)
-                    read_u64_from(file)
+                    .map(|x| x == 1).ok()
-                }).map(|x| x == 1)
             },
             memory_spread_page: {
-                self.open_path("cpuset.memory_spread_page", false).and_then(|file| {
+                self.open_path("cpuset.memory_spread_page", false).and_then(read_u64_from)
-                    read_u64_from(file)
+                    .map(|x| x == 1).unwrap_or(false)
-                }).map(|x| x == 1).unwrap_or(false)
             },
             memory_spread_slab: {
-                self.open_path("cpuset.memory_spread_slab", false).and_then(|file| {
+                self.open_path("cpuset.memory_spread_slab", false).and_then(read_u64_from)
-                    read_u64_from(file)
+                    .map(|x| x == 1).unwrap_or(false)
-                }).map(|x| x == 1).unwrap_or(false)
             },
             mems: {
-                self.open_path("cpuset.mems", false).and_then(|mut file| {
+                self.open_path("cpuset.mems", false).and_then(read_string_from).unwrap_or("".to_string())
-                    let mut string = String::new();
-                    let _ = file.read_to_string(&mut string);
-                    Some(string.trim().to_string())
-                }).unwrap()
             },
             sched_load_balance: {
-                self.open_path("cpuset.sched_load_balance", false).and_then(|file| {
+                self.open_path("cpuset.sched_load_balance", false).and_then(read_u64_from)
-                    read_u64_from(file)
+                    .map(|x| x == 1).unwrap_or(false)
-                }).map(|x| x == 1).unwrap_or(false)
             },
             sched_relax_domain_level: {
-                self.open_path("cpuset.sched_relax_domain_level", false).and_then(|file| {
+                self.open_path("cpuset.sched_relax_domain_level", false).and_then(read_u64_from)
-                    read_u64_from(file)
+                    .unwrap_or(0)
-                }).unwrap_or(0)
             },
         }
     }
 
     /// Control whether the CPUs selected via `set_cpus()` should be exclusive to this control
     /// group or not.
-    pub fn set_cpu_exclusive(self: &Self, b: bool) {
+    pub fn set_cpu_exclusive(self: &Self, b: bool) -> Result<(), CgroupError> {
         self.open_path("cpuset.cpu_exclusive", true).and_then(|mut file| {
             if b {
-                file.write_all(b"1").ok()
+                file.write_all(b"1").map_err(CgroupError::WriteError)
             } else {
-                file.write_all(b"0").ok()
+                file.write_all(b"0").map_err(CgroupError::WriteError)
             }
-        });
+        })
     }
 
     /// Control whether the memory nodes selected via `set_memss()` should be exclusive to this control
     /// group or not.
-    pub fn set_mem_exclusive(self: &Self, b: bool) {
+    pub fn set_mem_exclusive(self: &Self, b: bool) -> Result<(), CgroupError> {
         self.open_path("cpuset.mem_exclusive", true).and_then(|mut file| {
             if b {
-                file.write_all(b"1").ok()
+                file.write_all(b"1").map_err(CgroupError::WriteError)
             } else {
-                file.write_all(b"0").ok()
+                file.write_all(b"0").map_err(CgroupError::WriteError)
             }
-        });
+        })
     }
 
     /// Set the CPUs that the tasks in this control group can run on.
     ///
     /// Syntax is a comma separated list of CPUs, with an additional extension that ranges can
     /// be represented via dashes.
-    pub fn set_cpus(self: &Self, cpus: &String) {
+    pub fn set_cpus(self: &Self, cpus: &String) -> Result<(), CgroupError> {
         self.open_path("cpuset.cpus", true).and_then(|mut file| {
-            file.write_all(cpus.as_ref()).ok()
+            file.write_all(cpus.as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Set the memory nodes that the tasks in this control group can use.
     ///
     /// Syntax is the same as with `set_cpus()`.
-    pub fn set_mems(self: &Self, mems: &String) {
+    pub fn set_mems(self: &Self, mems: &String) -> Result<(), CgroupError> {
         self.open_path("cpuset.mems", true).and_then(|mut file| {
-            file.write_all(mems.as_ref()).ok()
+            file.write_all(mems.as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Controls whether the control group should be "hardwalled", i.e., whether kernel allocations
@@ -260,71 +245,71 @@ impl CpuSetController {
     ///
     /// Note that some kernel allocations, most notably those that are made in interrupt handlers
     /// may disregard this.
-    pub fn set_hardwall(self: &Self, b: bool) {
+    pub fn set_hardwall(self: &Self, b: bool) -> Result<(), CgroupError> {
         self.open_path("cpuset.mem_hardwall", true).and_then(|mut file| {
             if b {
-                file.write_all(b"1").ok()
+                file.write_all(b"1").map_err(CgroupError::WriteError)
             } else {
-                file.write_all(b"0").ok()
+                file.write_all(b"0").map_err(CgroupError::WriteError)
             }
-        });
+        })
     }
 
     /// Controls whether the kernel should attempt to rebalance the load between the CPUs specified in the
     /// `cpus` field of this control group.
-    pub fn set_load_balancing(self: &Self, b: bool) {
+    pub fn set_load_balancing(self: &Self, b: bool) -> Result<(), CgroupError> {
         self.open_path("cpuset.sched_load_balance", true).and_then(|mut file| {
             if b {
-                file.write_all(b"1").ok()
+                file.write_all(b"1").map_err(CgroupError::WriteError)
             } else {
-                file.write_all(b"0").ok()
+                file.write_all(b"0").map_err(CgroupError::WriteError)
             }
-        });
+        })
     }
 
     /// Contorl how much effort the kernel should invest in rebalacing the control group.
     ///
     /// See @CpuSet 's similar field for more information.
-    pub fn set_rebalance_relax_domain_level(self: &Self, i: i64) {
+    pub fn set_rebalance_relax_domain_level(self: &Self, i: i64) -> Result<(), CgroupError> {
         self.open_path("cpuset.sched_relax_domain_level", true).and_then(|mut file| {
-            file.write_all(i.to_string().as_ref()).ok()
+            file.write_all(i.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Control whether when using `set_mems()` the existing memory used by the tasks should be
     /// migrated over to the now-selected nodes.
-    pub fn set_memory_migration(self: &Self, b: bool) {
+    pub fn set_memory_migration(self: &Self, b: bool) -> Result<(), CgroupError> {
         self.open_path("cpuset.memory_migrate", true).and_then(|mut file| {
             if b {
-                file.write_all(b"1").ok()
+                file.write_all(b"1").map_err(CgroupError::WriteError)
             } else {
-                file.write_all(b"0").ok()
+                file.write_all(b"0").map_err(CgroupError::WriteError)
             }
-        });
+        })
     }
 
     /// Control whether filesystem buffers should be evenly split across the nodes selected via
     /// `set_mems()`.
-    pub fn set_memory_spread_page(self: &Self, b: bool) {
+    pub fn set_memory_spread_page(self: &Self, b: bool) -> Result<(), CgroupError> {
         self.open_path("cpuset.memory_spread_page", true).and_then(|mut file| {
             if b {
-                file.write_all(b"1").ok()
+                file.write_all(b"1").map_err(CgroupError::WriteError)
             } else {
-                file.write_all(b"0").ok()
+                file.write_all(b"0").map_err(CgroupError::WriteError)
             }
-        });
+        })
     }
 
     /// Control whether the kernel's slab cache for file I/O should be evenly split across the
     /// nodes selected via `set_mems()`.
-    pub fn set_memory_spread_slab(self: &Self, b: bool) {
+    pub fn set_memory_spread_slab(self: &Self, b: bool) -> Result<(), CgroupError> {
         self.open_path("cpuset.memory_spread_slab", true).and_then(|mut file| {
             if b {
-                file.write_all(b"1").ok()
+                file.write_all(b"1").map_err(CgroupError::WriteError)
             } else {
-                file.write_all(b"0").ok()
+                file.write_all(b"0").map_err(CgroupError::WriteError)
             }
-        });
+        })
     }
 
     /// Control whether the kernel should collect information to calculate memory pressure for
@@ -332,14 +317,14 @@ impl CpuSetController {
     ///
     /// Note: This is a no-operation if the control group referred by `self` is not the root
     /// control group.
-    pub fn set_enable_memory_pressure(self: &Self, b: bool) {
+    pub fn set_enable_memory_pressure(self: &Self, b: bool) -> Result<(), CgroupError> {
         /* XXX: this file should only be present in the root cpuset cg */
         self.open_path("cpuset.memory_pressure_enabled", true).and_then(|mut file| {
             if b {
-                file.write_all(b"1").ok()
+                file.write_all(b"1").map_err(CgroupError::WriteError)
             } else {
-                file.write_all(b"0").ok()
+                file.write_all(b"0").map_err(CgroupError::WriteError)
             }
-        });
+        })
     }
 }

src/devices.rs

@@ -5,7 +5,7 @@
 use std::path::PathBuf;
 use std::io::{Read, Write};
 
-use {DeviceResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use {CgroupError, DeviceResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
 
 /// A controller that allows controlling the `devices` subsystem of a Cgroup.
 ///
@@ -32,9 +32,9 @@ impl Controller for DevicesController {
                 let wstr = format!("{} {}:{} {}",
                                    i.devtype, i.major, i.minor, i.access);
                 if i.allow {
-                    self.allow_device(&wstr);
+                    let _ = self.allow_device(&wstr);
                 } else {
-                    self.deny_device(&wstr);
+                    let _ = self.deny_device(&wstr);
                 }
             }
         }
@@ -81,10 +81,10 @@ impl DevicesController {
     ///
     /// Note that `dev` can be "regex"-like: both `$major` and `$minor` can be `*` which implies
     /// that their value does not matter.
-    pub fn allow_device(self: &Self, dev: &String) {
+    pub fn allow_device(self: &Self, dev: &String) -> Result<(), CgroupError> {
         self.open_path("devices.allow", true).and_then(|mut file| {
-            file.write_all(dev.as_ref()).ok()
+            file.write_all(dev.as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Deny the control group's tasks access to the devices covered by `dev`.
@@ -96,18 +96,21 @@ impl DevicesController {
     ///
     /// Note that `dev` can be "regex"-like: both `$major` and `$minor` can be `*` which implies
     /// that their value does not matter.
-    pub fn deny_device(self: &Self, dev: &String) {
+    pub fn deny_device(self: &Self, dev: &String) -> Result<(), CgroupError> {
         self.open_path("devices.deny", true).and_then(|mut file| {
-            file.write_all(dev.as_ref()).ok()
+            file.write_all(dev.as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Get the current list of allowed devices.
-    pub fn allowed_devices(self: &Self) -> String {
+    pub fn allowed_devices(self: &Self) -> Result<String, CgroupError> {
         self.open_path("devices.list", false).and_then(|mut file| {
             let mut s = String::new();
-            let _ = file.read_to_string(&mut s);
+            let res = file.read_to_string(&mut s);
-            Some(s)
+            match res {
-        }).unwrap_or("".to_string())
+                Ok(_) => Ok(s),
+                Err(e) => Err(CgroupError::ReadError(e)),
+            }
+        })
     }
 }

src/freezer.rs

@@ -5,7 +5,7 @@
 use std::path::PathBuf;
 use std::io::{Read, Write};
 
-use {Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use {CgroupError, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
 
 /// A controller that allows controlling the `freezer` subsystem of a Cgroup.
 ///
@@ -73,30 +73,33 @@ impl FreezerController {
     }
 
     /// Freezes the processes in the control group.
-    pub fn freeze(self: &Self) {
+    pub fn freeze(self: &Self) -> Result<(), CgroupError> {
         self.open_path("freezer.state", true).and_then(|mut file| {
-            file.write_all("FROZEN".to_string().as_ref()).ok()
+            file.write_all("FROZEN".to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Thaws, that is, unfreezes the processes in the control group.
-    pub fn thaw(self: &Self) {
+    pub fn thaw(self: &Self) -> Result<(), CgroupError> {
         self.open_path("freezer.state", true).and_then(|mut file| {
-            file.write_all("THAWED".to_string().as_ref()).ok()
+            file.write_all("THAWED".to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Retrieve the state of processes in the control group.
-    pub fn state(self: &Self) -> FreezerState {
+    pub fn state(self: &Self) -> Result<FreezerState, CgroupError> {
         self.open_path("freezer.state", false).and_then(|mut file| {
             let mut s = String::new();
-            let _ = file.read_to_string(&mut s);
+            let res = file.read_to_string(&mut s);
-            match s.as_ref() {
+            match res {
-                "FROZEN" => Some(FreezerState::Frozen),
+                Ok(_) => match s.as_ref() {
-                "THAWED" => Some(FreezerState::Thawed),
+                    "FROZEN" => Ok(FreezerState::Frozen),
-                "FREEZING" => Some(FreezerState::Freezing),
+                    "THAWED" => Ok(FreezerState::Thawed),
-                _ => None,
+                    "FREEZING" => Ok(FreezerState::Freezing),
+                    _ => Err(CgroupError::ParseError),
+                },
+                Err(e) => Err(CgroupError::ReadError(e)),
             }
-        }).unwrap_or(FreezerState::Thawed)
+        })
     }
 }

src/hierarchies.rs

@@ -23,6 +23,8 @@ use ::net_prio::NetPrioController;
 use ::hugetlb::HugeTlbController;
 use ::rdma::RdmaController;
 
+use ::cgroup::Cgroup;
+
 
 /// The standard, original cgroup implementation. Often referred to as "cgroupv1".
 pub struct V1 {
@@ -75,6 +77,10 @@ impl Hierarchy for V1 {
         subs
     }
 
+    fn root_control_group(self: &Self) -> Cgroup {
+        Cgroup::load(self, "".to_string())
+    }
+
     fn check_support(self: &Self, sub: Controllers) -> bool {
         let root = self.root().read_dir().unwrap();
         for entry in root {

src/hugetlb.rs

@@ -6,7 +6,8 @@ use std::path::PathBuf;
 use std::fs::File;
 use std::io::{Write, Read};
 
-use {HugePageResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use {CgroupError, HugePageResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use CgroupError::*;
 
 
 /// A controller that allows controlling the `hugetlb` subsystem of a Cgroup.
@@ -31,7 +32,7 @@ impl Controller for HugeTlbController {
 
         if res.update_values {
             for i in &res.limits {
-                self.set_limit_in_bytes(&i.size, i.limit);
+                let _ = self.set_limit_in_bytes(&i.size, i.limit);
             }
         }
     }
@@ -57,10 +58,12 @@ impl<'a> From<&'a Subsystem> for &'a HugeTlbController {
     }
 }
 
-fn read_u64_from(mut file: File) -> Option<u64> {
+fn read_u64_from(mut file: File) -> Result<u64, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    match file.read_to_string(&mut string) {
-    string.trim().parse().ok()
+        Ok(_) => string.trim().parse().map_err(|_| ParseError),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }
 
 impl HugeTlbController {
@@ -81,38 +84,38 @@ impl HugeTlbController {
     }
 
     /// Check how many times has the limit of `hugetlb_size` hugepages been hit.
-    pub fn failcnt(self: &Self, hugetlb_size: &String) -> Option<u64> {
+    pub fn failcnt(self: &Self, hugetlb_size: &String) -> Result<u64, CgroupError> {
         self.open_path(&format!("hugetlb.{}.failcnt", hugetlb_size), false)
             .and_then(read_u64_from)
     }
 
     /// Get the limit (in bytes) of how much memory can be backed by hugepages of a certain size
     /// (`hugetlb_size`).
-    pub fn limit_in_bytes(self: &Self, hugetlb_size: &String) -> Option<u64> {
+    pub fn limit_in_bytes(self: &Self, hugetlb_size: &String) -> Result<u64, CgroupError> {
         self.open_path(&format!("hugetlb.{}.limit_in_bytes", hugetlb_size), false)
             .and_then(read_u64_from)
     }
 
     /// Get the current usage of memory that is backed by hugepages of a certain size
     /// (`hugetlb_size`).
-    pub fn usage_in_bytes(self: &Self, hugetlb_size: &String) -> Option<u64> {
+    pub fn usage_in_bytes(self: &Self, hugetlb_size: &String) -> Result<u64, CgroupError> {
         self.open_path(&format!("hugetlb.{}.usage_in_bytes", hugetlb_size), false)
             .and_then(read_u64_from)
     }
 
     /// Get the maximum observed usage of memory that is backed by hugepages of a certain size
     /// (`hugetlb_size`).
-    pub fn max_usage_in_bytes(self: &Self, hugetlb_size: &String) -> Option<u64> {
+    pub fn max_usage_in_bytes(self: &Self, hugetlb_size: &String) -> Result<u64, CgroupError> {
         self.open_path(&format!("hugetlb.{}.max_usage_in_bytes", hugetlb_size), false)
             .and_then(read_u64_from)
     }
 
     /// Set the limit (in bytes) of how much memory can be backed by hugepages of a certain size
     /// (`hugetlb_size`).
-    pub fn set_limit_in_bytes(self: &Self, hugetlb_size: &String, limit: u64) {
+    pub fn set_limit_in_bytes(self: &Self, hugetlb_size: &String, limit: u64) -> Result<(), CgroupError> {
         self.open_path(&format!("hugetlb.{}.limit_in_bytes", hugetlb_size), false)
             .and_then(|mut file| {
-                file.write_all(limit.to_string().as_ref()).ok()
+                file.write_all(limit.to_string().as_ref()).map_err(CgroupError::WriteError)
-            });
+            })
     }
 }

src/lib.rs

@@ -1,6 +1,6 @@
 use std::path::PathBuf;
 use std::fs::File;
-use std::io::Write;
+use std::io::{BufRead, BufReader, Write};
 
 pub mod hierarchies;
 pub mod pid;
@@ -32,6 +32,8 @@ use net_prio::NetPrioController;
 use hugetlb::HugeTlbController;
 use rdma::RdmaController;
 
+pub use cgroup::Cgroup;
+
 /// Contains all the subsystems that are available in this crate.
 #[derive(Debug)]
 pub enum Subsystem {
@@ -63,6 +65,29 @@ pub enum Subsystem {
     Rdma(RdmaController),
 }
 
+/// The different types of errors that can occur while manipulating control groups.
+#[derive(Debug)]
+pub enum CgroupError {
+    /// An error occured while writing to a control group file.
+    WriteError(std::io::Error),
+    /// An error occured while trying to read from a control group file.
+    ReadError(std::io::Error),
+    /// An error occured while trying to parse a value from a control group file.
+    ///
+    /// In the future, there will be some information attached to this field.
+    ParseError,
+    /// You tried to do something invalid.
+    ///
+    /// This could be because you tried to set a value in a control group that is not a root
+    /// control group. Or, when using unified hierarchy, you tried to add a task in a leaf node.
+    InvalidOperation,
+    /// The path of the control group was invalid.
+    ///
+    /// This could be caused by trying to escape the control group filesystem via a string of "..".
+    /// This crate checks against this and operations will fail with this error.
+    InvalidPath,
+}
+
 #[doc(hidden)]
 #[derive(Eq, PartialEq, Debug)]
 pub enum Controllers {
@@ -147,32 +172,47 @@ pub trait Controller {
     }
 
     #[doc(hidden)]
-    fn open_path(self: &Self, p: &str, w: bool) -> Option<File> {
+    fn open_path(self: &Self, p: &str, w: bool) -> Result<File, CgroupError> {
         let mut path = self.get_path().clone();
         path.push(p);
 
         if !self.verify_path() {
-            return None;
+            return Err(CgroupError::InvalidPath);
         }
 
         if w {
             match File::create(&path) {
-                Err(_) => return None,
+                Err(e) => return Err(CgroupError::WriteError(e)),
-                Ok(file) => return Some(file),
+                Ok(file) => return Ok(file),
             }
         } else {
             match File::open(&path) {
-                Err(_) => return None,
+                Err(e) => return Err(CgroupError::ReadError(e)),
-                Ok(file) => return Some(file),
+                Ok(file) => return Ok(file),
             }
         }
     }
 
     /// Attach a task to this controller.
-    fn add_task(self: &Self, pid: &CgroupPid) {
+    fn add_task(self: &Self, pid: &CgroupPid) -> Result<(), CgroupError> {
         self.open_path("tasks", true).and_then(|mut file| {
-            file.write_all(pid.pid.to_string().as_ref()).ok()
+            file.write_all(pid.pid.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
+    }
+
+    /// Get the list of tasks that this controller has.
+    fn tasks(self: &Self) -> Vec<CgroupPid> {
+        self.open_path("tasks", false).and_then(|file| {
+            let bf = BufReader::new(file);
+            let mut v = Vec::new();
+            for line in bf.lines() {
+                if let Ok(line) = line {
+                    let n = line.trim().parse().unwrap_or(0u64);
+                    v.push(n);
+                }
+            }
+            Ok(v.into_iter().map(CgroupPid::from).collect())
+        }).unwrap_or(vec![])
     }
 }
 
@@ -186,11 +226,17 @@ pub trait ControllIdentifier {
 pub trait Hierarchy {
     /// Returns what subsystems are supported by the hierarchy.
     fn subsystems(self: &Self) -> Vec<Subsystem>;
+
     /// Returns the root directory of the hierarchy.
     fn root(self: &Self) -> PathBuf;
+
+    /// Return a handle to the root control group in the hierarchy.
+    fn root_control_group(self: &Self) -> Cgroup;
+
     /// Checks whether a certain subsystem is supported in the hierarchy.
     ///
     /// This is an internal function and should not be used.
+    #[doc(hidden)]
     fn check_support(self: &Self, sub: Controllers) -> bool;
 }
 
@@ -389,6 +435,7 @@ pub struct Resources {
 
 /// A structure representing a `pid`. Currently implementations exist for `u64` and
 /// `std::process::Child`.
+#[derive(Debug, PartialEq, Eq, PartialOrd, Ord)]
 pub struct CgroupPid {
     /// The process identifier
     pub pid: u64,
@@ -500,67 +547,3 @@ impl Subsystem {
         }
     }
 }
-
-
-#[cfg(test)]
-mod tests {
-    use {Resources, PidResources, Hierarchy, Controller, Controllers, Subsystem};
-    use pid::{PidMax, PidController};
-    use cgroup::Cgroup;
-
-    #[test]
-    fn create_and_delete_cgroup() {
-        let hier = ::hierarchies::V1::new();
-        let cg = Cgroup::new(&hier, String::from("ltest2"), 0);
-        {
-            let pidcontroller: &PidController = cg.controller_of().unwrap();
-            pidcontroller.set_pid_max(PidMax::Value(1337));
-            assert_eq!(pidcontroller.get_pid_max(), Some(PidMax::Value(1337)));
-        }
-        cg.delete();
-    }
-
-    #[test]
-    fn test_pid_pids_current_is_zero() {
-        let hier = ::hierarchies::V1::new();
-        let cg = Cgroup::new(&hier, String::from("ltest3"), 0);
-        {
-            let pidcontroller: &PidController = cg.controller_of().unwrap();
-            assert_eq!(pidcontroller.get_pid_current(), 0);
-        }
-        cg.delete();
-    }
-
-    #[test]
-    fn test_pid_pids_events_is_zero() {
-        let hier = ::hierarchies::V1::new();
-        let cg = Cgroup::new(&hier, String::from("ltest4"), 0);
-        {
-            let pidcontroller: &PidController = cg.controller_of().unwrap();
-            assert_eq!(pidcontroller.get_pid_events(), 0);
-        }
-        cg.delete();
-    }
-
-    #[test]
-    fn test_setting_resources() {
-        let hier = ::hierarchies::V1::new();
-        let cg = Cgroup::new(&hier, String::from("ltest5"), 0);
-        {
-            let res = Resources {
-                pid: PidResources {
-                    update_values: true,
-                    maximum_number_of_processes: PidMax::Value(512),
-                },
-                ..Default::default()
-            };
-            cg.apply(&res);
-
-            /* verify */
-            let pidcontroller: &PidController = cg.controller_of().unwrap();
-            assert_eq!(pidcontroller.get_pid_max(), Some(PidMax::Value(512)));
-        }
-        cg.delete();
-    }
-
-}

src/memory.rs

@@ -6,7 +6,8 @@ use std::path::PathBuf;
 use std::io::{Write, Read};
 use std::fs::File;
 
-use {Resources, MemoryResources, Controller, Controllers, Subsystem, ControllIdentifier};
+use {CgroupError, Resources, MemoryResources, Controller, Controllers, Subsystem, ControllIdentifier};
+use CgroupError::*;
 
 /// A controller that allows controlling the `memory` subsystem of a Cgroup.
 ///
@@ -51,7 +52,7 @@ pub struct Memory {
     /// Contains various statistics about the NUMA locality of the control group's tasks.
     ///
     /// The format of this field (as lifted from the kernel sources):
-    /// ```
+    /// ```text
     /// total=<total pages> N0=<node 0 pages> N1=<node 1 pages> ...
     /// file=<total file pages> N0=<node 0 pages> N1=<node 1 pages> ...
     /// anon=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
@@ -125,12 +126,12 @@ impl Controller for MemController {
         let memres: &MemoryResources = &res.memory;
 
         if memres.update_values {
-            self.set_limit(memres.memory_hard_limit);
+            let _ = self.set_limit(memres.memory_hard_limit);
-            self.set_soft_limit(memres.memory_soft_limit);
+            let _ = self.set_soft_limit(memres.memory_soft_limit);
-            self.set_kmem_limit(memres.kernel_memory_limit);
+            let _ = self.set_kmem_limit(memres.kernel_memory_limit);
-            self.set_memswap_limit(memres.memory_swap_limit);
+            let _ = self.set_memswap_limit(memres.memory_swap_limit);
-            self.set_tcp_limit(memres.kernel_tcp_memory_limit);
+            let _ = self.set_tcp_limit(memres.kernel_tcp_memory_limit);
-            self.set_swappiness(memres.swappiness);
+            let _ = self.set_swappiness(memres.swappiness);
         }
     }
 }
@@ -154,46 +155,29 @@ impl MemController {
     pub fn memory_stat(self: &Self) -> Memory {
         Memory {
             fail_cnt: self.open_path("memory.failcnt", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             limit_in_bytes: self.open_path("memory.limit_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             usage_in_bytes: self.open_path("memory.usage_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             max_usage_in_bytes: self.open_path("memory.max_usage_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             move_charge_at_immigrate: self.open_path("memory.move_charge_at_immigrate", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             numa_stat: self.open_path("memory.numa_stat", false)
-                            .and_then(|mut file| {
+                            .and_then(read_string_from).unwrap_or("".to_string()),
-                                let mut string = String::new();
-                                let _ = file.read_to_string(&mut string);
-                                Some(string.trim().to_string())
-                            }).unwrap_or("".to_string()),
             oom_control: self.open_path("memory.oom_control", false)
-                            .and_then(|mut file| {
+                            .and_then(read_string_from).unwrap_or("".to_string()),
-                                let mut string = String::new();
-                                let _ = file.read_to_string(&mut string);
-                                Some(string.trim().to_string())
-                            }).unwrap_or("".to_string()),
             soft_limit_in_bytes: self.open_path("memory.soft_limit_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from)
                             .unwrap_or(0),
             stat: self.open_path("memory.stat", false)
-                            .and_then(|mut file| {
+                            .and_then(read_string_from).unwrap_or("".to_string()),
-                                let mut string = String::new();
-                                let _ = file.read_to_string(&mut string);
-                                Some(string.trim().to_string())
-                            }).unwrap_or("".to_string()),
             swappiness: self.open_path("memory.swappiness", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from)
                             .unwrap_or(0),
             use_hierarchy: self.open_path("memory.use_hierarchy", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from)
                             .unwrap_or(0)
         }
     }
@@ -202,23 +186,15 @@ impl MemController {
     pub fn kmem_stat(self: &Self) -> Kmem {
         Kmem {
             fail_cnt: self.open_path("memory.kmem.failcnt", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             limit_in_bytes: self.open_path("memory.kmem.limit_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             usage_in_bytes: self.open_path("memory.kmem.usage_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             max_usage_in_bytes: self.open_path("memory.kmem.max_usage_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             slabinfo: self.open_path("memory.kmem.slabinfo", false)
-                            .and_then(|mut file| {
+                            .and_then(read_string_from).unwrap_or("".to_string()),
-                                let mut string = String::new();
-                                let _ = file.read_to_string(&mut string);
-                                Some(string.trim().to_string())
-                            }).unwrap_or("".to_string()),
         }
     }
 
@@ -227,17 +203,13 @@ impl MemController {
     pub fn kmem_tcp_stat(self: &Self) -> Tcp {
         Tcp {
             fail_cnt: self.open_path("memory.kmem.tcp.failcnt", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             limit_in_bytes: self.open_path("memory.kmem.tcp.limit_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             usage_in_bytes: self.open_path("memory.kmem.tcp.usage_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             max_usage_in_bytes: self.open_path("memory.kmem.tcp.max_usage_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
         }
     }
 
@@ -246,46 +218,42 @@ impl MemController {
     pub fn memswap(self: &Self) -> MemSwap {
         MemSwap {
             fail_cnt: self.open_path("memory.memsw.failcnt", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             limit_in_bytes: self.open_path("memory.memsw.limit_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             usage_in_bytes: self.open_path("memory.memsw.usage_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
             max_usage_in_bytes: self.open_path("memory.memsw.max_usage_in_bytes", false)
-                            .and_then(|file| read_u64_from(file))
+                            .and_then(read_u64_from).unwrap_or(0),
-                            .unwrap_or(0),
         }
     }
 
     /// Set the memory usage limit of the control group, in bytes.
-    pub fn set_limit(self: &Self, limit: u64) {
+    pub fn set_limit(self: &Self, limit: u64) -> Result<(), CgroupError> {
         self.open_path("memory.limit_in_bytes", true).and_then(|mut file| {
-            file.write_all(limit.to_string().as_ref()).ok()
+            file.write_all(limit.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Set the kernel memory limit of the control group, in bytes.
-    pub fn set_kmem_limit(self: &Self, limit: u64) {
+    pub fn set_kmem_limit(self: &Self, limit: u64) -> Result<(), CgroupError> {
         self.open_path("memory.kmem.limit_in_bytes", true).and_then(|mut file| {
-            file.write_all(limit.to_string().as_ref()).ok()
+            file.write_all(limit.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Set the memory+swap limit of the control group, in bytes.
-    pub fn set_memswap_limit(self: &Self, limit: u64) {
+    pub fn set_memswap_limit(self: &Self, limit: u64) -> Result<(), CgroupError> {
         self.open_path("memory.memsw.limit_in_bytes", true).and_then(|mut file| {
-            file.write_all(limit.to_string().as_ref()).ok()
+            file.write_all(limit.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Set how much kernel memory can be used for TCP-related buffers by the control group.
-    pub fn set_tcp_limit(self: &Self, limit: u64) {
+    pub fn set_tcp_limit(self: &Self, limit: u64) -> Result<(), CgroupError> {
         self.open_path("memory.kmem.tcp.limit_in_bytes", true).and_then(|mut file| {
-            file.write_all(limit.to_string().as_ref()).ok()
+            file.write_all(limit.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
 
@@ -293,10 +261,10 @@ impl MemController {
     ///
     /// This limit is enforced when the system is nearing OOM conditions. Contrast this with the
     /// hard limit, which is _always_ enforced.
-    pub fn set_soft_limit(self: &Self, limit: u64) {
+    pub fn set_soft_limit(self: &Self, limit: u64) -> Result<(), CgroupError> {
         self.open_path("memory.soft_limit_in_bytes", true).and_then(|mut file| {
-            file.write_all(limit.to_string().as_ref()).ok()
+            file.write_all(limit.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
 
@@ -304,10 +272,10 @@ impl MemController {
     /// group.
     ///
     /// Note that a value of zero does not imply that the process will not be swapped out.
-    pub fn set_swappiness(self: &Self, swp: u64) {
+    pub fn set_swappiness(self: &Self, swp: u64) -> Result<(), CgroupError> {
         self.open_path("memory.swappiness", true).and_then(|mut file| {
-            file.write_all(swp.to_string().as_ref()).ok()
+            file.write_all(swp.to_string().as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 }
 
@@ -331,8 +299,18 @@ impl<'a> From<&'a Subsystem> for &'a MemController {
     }
 }
 
-fn read_u64_from(mut file: File) -> Option<u64> {
+fn read_u64_from(mut file: File) -> Result<u64, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    match file.read_to_string(&mut string) {
-    string.trim().parse().ok()
+        Ok(_) => string.trim().parse().map_err(|_| ParseError),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
+}
+
+fn read_string_from(mut file: File) -> Result<String, CgroupError> {
+    let mut string = String::new();
+    match file.read_to_string(&mut string) {
+        Ok(_) => Ok(string.trim().to_string()),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }

src/net_cls.rs

@@ -6,7 +6,8 @@ use std::path::PathBuf;
 use std::io::{Read, Write};
 use std::fs::File;
 
-use {NetworkResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use {CgroupError, NetworkResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use CgroupError::*;
 
 /// A controller that allows controlling the `net_cls` subsystem of a Cgroup.
 ///
@@ -30,7 +31,7 @@ impl Controller for NetClsController {
         let res: &NetworkResources = &res.network;
 
         if res.update_values {
-            self.set_class(res.class_id);
+            let _ = self.set_class(res.class_id);
         }
     }
 }
@@ -55,10 +56,12 @@ impl<'a> From<&'a Subsystem> for &'a NetClsController {
     }
 }
 
-fn read_u64_from(mut file: File) -> Option<u64> {
+fn read_u64_from(mut file: File) -> Result<u64, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    match file.read_to_string(&mut string) {
-    string.trim().parse().ok()
+        Ok(_) => string.trim().parse().map_err(|_| ParseError),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }
 
 impl NetClsController {
@@ -73,17 +76,17 @@ impl NetClsController {
     }
     
     /// Set the network class id of the outgoing packets of the control group's tasks.
-    pub fn set_class(self: &Self, class: u64) {
+    pub fn set_class(self: &Self, class: u64) -> Result<(), CgroupError> {
         self.open_path("net_cls.classid", true).and_then(|mut file| {
             let s = format!("{:#08X}", class);
-            file.write_all(s.as_ref()).ok()
+            file.write_all(s.as_ref()).map_err(CgroupError::WriteError)
-        });
+        })
     }
 
     /// Get the network class id of the outgoing packets of the control group's tasks.
-    pub fn get_class(self: &Self) -> u64 {
+    pub fn get_class(self: &Self) -> Result<u64, CgroupError> {
         self.open_path("net_cls.classid", false).and_then(|file| {
             read_u64_from(file)
-        }).unwrap_or(0u64)
+        })
     }
 }

src/net_prio.rs

@@ -7,7 +7,8 @@ use std::io::{BufReader, BufRead, Write, Read};
 use std::fs::File;
 use std::collections::HashMap;
 
-use {NetworkResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use {CgroupError, NetworkResources, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use CgroupError::*;
 
 /// A controller that allows controlling the `net_prio` subsystem of a Cgroup.
 ///
@@ -32,7 +33,7 @@ impl Controller for NetPrioController {
 
         if res.update_values {
             for i in &res.priorities {
-                self.set_if_prio(&i.name, i.priority);
+                let _ = self.set_if_prio(&i.name, i.priority);
             }
         }
     }
@@ -58,10 +59,12 @@ impl<'a> From<&'a Subsystem> for &'a NetPrioController {
     }
 }
 
-fn read_u64_from(mut file: File) -> Option<u64> {
+fn read_u64_from(mut file: File) -> Result<u64, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    match file.read_to_string(&mut string) {
-    string.trim().parse().ok()
+        Ok(_) => string.trim().parse().map_err(|_| ParseError),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }
 
 impl NetPrioController {
@@ -83,24 +86,39 @@ impl NetPrioController {
     }
 
     /// A map of priorities for each network interface.
-    pub fn ifpriomap(self: &Self) -> HashMap<String, u64> {
+    pub fn ifpriomap(self: &Self) -> Result<HashMap<String, u64>, CgroupError> {
-        self.open_path("net_prio.ifpriomap", false)
+        self.open_path("net_prio.ifpriomap", false) .and_then(|file| {
-            .and_then(|file| {
+            let bf = BufReader::new(file);
-                let bf = BufReader::new(file);
+            bf.lines().fold(Ok(HashMap::new()), |acc, line| {
-                Some(bf.lines().map(|line| {
+                if acc.is_err() {
+                    acc
+                } else {
+                    let mut acc = acc.unwrap();
                     let l = line.unwrap();
                     let mut sp = l.split_whitespace();
-                    (sp.nth(0).unwrap().to_string(),
+                    let ifname = sp.nth(0);
-                     sp.nth(1).unwrap().trim().parse().unwrap())
+                    let ifprio = sp.nth(1);
-                }).collect())
+                    if ifname.is_none() || ifprio.is_none() {
-            }).unwrap_or(HashMap::new())
+                        Err(CgroupError::ParseError)
+                    } else {
+                        let ifname = ifname.unwrap();
+                        let ifprio = ifprio.unwrap().trim().parse();
+                        if ifprio.is_err() {
+                            Err(CgroupError::ParseError)
+                        } else {
+                            acc.insert(ifname.to_string(), ifprio.unwrap());
+                            Ok(acc)
+                        }
+                    }
+                }
+            })
+        })
     }
 
     /// Set the priority of the network traffic on `eif` to be `prio`.
-    pub fn set_if_prio(self: &Self, eif: &String, prio: u64) {
+    pub fn set_if_prio(self: &Self, eif: &String, prio: u64) -> Result<(), CgroupError> {
-        self.open_path("net_prio.ifpriomap", true)
+        self.open_path("net_prio.ifpriomap", true).and_then(|mut file| {
-            .and_then(|mut file| {
+            file.write_all(format!("{} {}", eif, prio).as_ref()).map_err(CgroupError::WriteError)
-                Some(file.write_all(format!("{} {}", eif, prio).as_ref()))
+        })
-            });
     }
 }

src/pid.rs

@@ -4,8 +4,10 @@
 //!  [Documentation/cgroups-v1/pids.txt](https://www.kernel.org/doc/Documentation/cgroup-v1/pids.txt)
 use std::path::PathBuf;
 use std::io::{Write, Read};
+use std::fs::File;
 
-use {Resources, PidResources, Controller, ControllIdentifier, Subsystem, Controllers};
+use {CgroupError, Resources, PidResources, Controller, ControllIdentifier, Subsystem, Controllers};
+use CgroupError::*;
 
 /// A controller that allows controlling the `pids` subsystem of a Cgroup.
 #[derive(Debug, Clone)]
@@ -42,7 +44,7 @@ impl Controller for PidController {
 
         if pidres.update_values {
             /* apply pid_max */
-            self.set_pid_max(pidres.maximum_number_of_processes);
+            let _ = self.set_pid_max(pidres.maximum_number_of_processes);
         }
     }
 }
@@ -73,6 +75,14 @@ impl<'a> From<&'a Subsystem> for &'a PidController {
     }
 }
 
+fn read_u64_from(mut file: File) -> Result<u64, CgroupError> {
+    let mut string = String::new();
+    match file.read_to_string(&mut string) {
+        Ok(_) => string.trim().parse().map_err(|_| ParseError),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
+}
+
 impl PidController {
     /// Constructors a new `PidController` instance, with `oroot` serving as the controller's root
     /// directory.
@@ -86,32 +96,44 @@ impl PidController {
     }
 
     /// The number of times `fork` failed because the limit was hit.
-    pub fn get_pid_events(self: &Self) -> i64 {
+    pub fn get_pid_events(self: &Self) -> Result<u64, CgroupError> {
         self.open_path("pids.events", false).and_then(|mut file| {
             let mut string = String::new();
-            let _ = file.read_to_string(&mut string);
+            match file.read_to_string(&mut string) {
-            Some(string.split_whitespace().nth(1).unwrap().parse().unwrap_or(0))
+                Ok(_) => {
-        }).unwrap()
+                    match string.split_whitespace().nth(1) {
+                        Some(elem) => match elem.parse() {
+                            Ok(val) => Ok(val),
+                            Err(_) => Err(CgroupError::ParseError),
+                        },
+                        None => Err(CgroupError::ParseError),
+                    }
+                },
+                Err(e) => Err(CgroupError::ReadError(e)),
+            }
+        })
     }
 
     /// The number of processes currently.
-    pub fn get_pid_current(self: &Self) -> i64 {
+    pub fn get_pid_current(self: &Self) -> Result<u64, CgroupError> {
-        self.open_path("pids.current", false).and_then(|mut file| {
+        self.open_path("pids.current", false).and_then(read_u64_from)
-            let mut string = String::new();
-            let _ = file.read_to_string(&mut string);
-            Some(string.trim().parse().unwrap_or(0))
-        }).unwrap()
     }
 
     /// The maximum number of processes that can exist at one time in the control group.
-    pub fn get_pid_max(self: &Self) -> Option<PidMax> {
+    pub fn get_pid_max(self: &Self) -> Result<PidMax, CgroupError> {
         self.open_path("pids.max", false).and_then(|mut file| {
             let mut string = String::new();
-            let _ = file.read_to_string(&mut string);
+            let res = file.read_to_string(&mut string);
-            if string.trim() == "max" {
+            match res {
-                Some(PidMax::Max)
+                Ok(_) => if string.trim() == "max" {
-            } else {
+                    Ok(PidMax::Max)
-                Some(PidMax::Value(string.trim().parse().unwrap_or(0)))
+                } else {
+                    match string.trim().parse() {
+                        Ok(val) => Ok(PidMax::Value(val)),
+                        Err(_) => Err(CgroupError::ParseError),
+                    }
+                },
+                Err(e) => Err(CgroupError::ReadError(e)),
             }
         })
     }
@@ -121,17 +143,16 @@ impl PidController {
     /// Note that if `get_pid_current()` returns a higher number than what you
     /// are about to set (`max_pid`), then no processess will be killed. Additonally, attaching
     /// extra processes to a control group disregards the limit.
-    pub fn set_pid_max(self: &Self, max_pid: PidMax) {
+    pub fn set_pid_max(self: &Self, max_pid: PidMax) -> Result<(), CgroupError> {
         self.open_path("pids.max", true).and_then(|mut file| {
             let string_to_write = match max_pid {
                 PidMax::Max => "max".to_string(),
                 PidMax::Value(num) => num.to_string(),
             };
             match file.write_all(string_to_write.as_ref()) {
-                Ok(_) => (),
+                Ok(_) => Ok(()),
-                Err(e) => println!("error {:?}", e),
+                Err(e) => Err(CgroupError::WriteError(e)),
             }
-            Some(0i64)
+        })
-        });
     }
 }

src/rdma.rs

@@ -6,7 +6,7 @@ use std::path::PathBuf;
 use std::io::{Write, Read};
 use std::fs::File;
 
-use {Controllers, Controller, Resources, ControllIdentifier, Subsystem};
+use {CgroupError, Controllers, Controller, Resources, ControllIdentifier, Subsystem};
 
 /// A controller that allows controlling the `rdma` subsystem of a Cgroup.
 ///
@@ -48,10 +48,12 @@ impl<'a> From<&'a Subsystem> for &'a RdmaController {
     }
 }
 
-fn read_string_from(mut file: File) -> Option<String> {
+fn read_string_from(mut file: File) -> Result<String, CgroupError> {
     let mut string = String::new();
-    let _ = file.read_to_string(&mut string);
+    match file.read_to_string(&mut string) {
-    Some(string.trim().to_string())
+        Ok(_) => Ok(string.trim().to_string()),
+        Err(e) => Err(CgroupError::ReadError(e)),
+    }
 }
 
 impl RdmaController {
@@ -66,17 +68,15 @@ impl RdmaController {
     }
 
     /// Returns the current usage of RDMA/IB specific resources.
-    pub fn current(self: &Self) -> String {
+    pub fn current(self: &Self) -> Result<String, CgroupError> {
         self.open_path("rdma.current", false)
             .and_then(read_string_from)
-            .unwrap_or("".to_string())
     }
 
     /// Set a maximum usage for each RDMA/IB resource.
-    pub fn set_max(self: &Self, max: &String) {
+    pub fn set_max(self: &Self, max: &String) -> Result<(), CgroupError> {
-        self.open_path("rdma.max", true)
+        self.open_path("rdma.max", true).and_then(|mut file| {
-            .and_then(|mut file| {
+            file.write_all(max.as_ref()).map_err(CgroupError::WriteError)
-                file.write_all(max.as_ref()).ok()
+        })
-            });
     }
 }

tests/cgroup.rs

@@ -0,0 +1,29 @@
+//! Simple unit tests about the control groups system.
+extern crate cgroups;
+use cgroups::{Cgroup, CgroupPid};
+
+extern crate nix;
+extern crate libc;
+
+#[test]
+fn test_tasks_iterator() {
+    let hier = cgroups::hierarchies::V1::new();
+    let pid = libc::pid_t::from(nix::unistd::getpid()) as u64;
+    let cg = Cgroup::new(&hier, String::from("test_tasks_iterator"));
+    {
+        // Add a task to the control group.
+        cg.add_task(CgroupPid::from(pid));
+        let mut tasks = cg.tasks().into_iter();
+        // Verify that the task is indeed in the control group
+        assert_eq!(tasks.next(), Some(CgroupPid::from(pid)));
+        assert_eq!(tasks.next(), None);
+        
+        // Now, try removing it.
+        cg.remove_task(CgroupPid::from(pid));
+        tasks = cg.tasks().into_iter();
+
+        // Verify that it was indeed removed.
+        assert_eq!(tasks.next(), None);
+    }
+    cg.delete();
+}

tests/pids.rs

@@ -0,0 +1,95 @@
+//! Integration tests about the pids subsystem
+extern crate cgroups;
+use cgroups::{CgroupError, CgroupPid, Cgroup, Resources, PidResources};
+use cgroups::pid::{PidController, PidMax};
+use cgroups::Controller;
+
+extern crate nix;
+use nix::unistd::{Pid, fork, ForkResult};
+use nix::sys::wait::{waitpid, WaitStatus};
+
+extern crate libc;
+use libc::pid_t;
+
+use std::thread;
+
+#[test]
+fn create_and_delete_cgroup() {
+    let hier = cgroups::hierarchies::V1::new();
+    let cg = Cgroup::new(&hier, String::from("create_and_delete_cgroup"));
+    {
+        let pidcontroller: &PidController = cg.controller_of().unwrap();
+        pidcontroller.set_pid_max(PidMax::Value(1337));
+        assert_eq!(pidcontroller.get_pid_max(), Some(PidMax::Value(1337)));
+    }
+    cg.delete();
+}
+
+#[test]
+fn test_pids_current_is_zero() {
+    let hier = cgroups::hierarchies::V1::new();
+    let cg = Cgroup::new(&hier, String::from("test_pids_current_is_zero"));
+    {
+        let pidcontroller: &PidController = cg.controller_of().unwrap();
+        assert_eq!(pidcontroller.get_pid_current(), 0);
+    }
+    cg.delete();
+}
+
+#[test]
+fn test_pids_events_is_zero() {
+    let hier = cgroups::hierarchies::V1::new();
+    let cg = Cgroup::new(&hier, String::from("test_pids_events_is_zero"));
+    {
+        let pidcontroller: &PidController = cg.controller_of().unwrap();
+        assert_eq!(pidcontroller.get_pid_events(), 0);
+    }
+    cg.delete();
+}
+
+#[test]
+fn test_pid_events_is_not_zero() {
+    let hier = cgroups::hierarchies::V1::new();
+    let cg = Cgroup::new(&hier, String::from("test_pid_events_is_not_zero"));
+    {
+        let pids: &PidController = cg.controller_of().unwrap();
+        let before = pids.get_pid_events();
+
+        match fork() {
+            Ok(ForkResult::Parent { child, .. }) => {
+                // move the process into the control group
+                pids.add_task(&(pid_t::from(child) as u64).into());
+
+                println!("added task to cg: {:?}", child);
+
+                // Set limit to one
+                pids.set_pid_max(PidMax::Value(1));
+                println!("err = {:?}", pids.get_pid_max());
+
+                // wait on the child
+                let res = waitpid(child, None);
+                if let Ok(WaitStatus::Exited(_, e)) = res {
+                    assert_eq!(e, 0i32);
+                } else {
+                    panic!("found result: {:?}", res);
+                }
+
+                // Check pids.events
+                assert_eq!(pids.get_pid_events(), before + 1);
+            },
+            Ok(ForkResult::Child) => {
+                loop {
+                    if pids.get_pid_max() == Some(PidMax::Value(1)) {
+                        if let Err(_) = fork() {
+                            unsafe { libc::exit(0) };
+                        } else {
+                            unsafe { libc::exit(1) };
+                        }
+                    }
+                }
+            },
+            Err(_) => panic!("failed to fork"),
+        }
+    }
+    cg.delete();
+}

tests/resources.rs

@@ -0,0 +1,26 @@
+//! Integration test about setting resources using `apply()`
+extern crate cgroups;
+
+use cgroups::{Cgroup, Resources, PidResources};
+use cgroups::pid::{PidController, PidMax};
+
+#[test]
+fn pid_resources() {
+    let hier = cgroups::hierarchies::V1::new();
+    let cg = Cgroup::new(&hier, String::from("pid_resources"));
+    {
+        let res = Resources {
+            pid: PidResources {
+                update_values: true,
+                maximum_number_of_processes: PidMax::Value(512),
+            },
+            ..Default::default()
+        };
+        cg.apply(&res);
+
+        /* verify */
+        let pidcontroller: &PidController = cg.controller_of().unwrap();
+        assert_eq!(pidcontroller.get_pid_max(), Some(PidMax::Value(512)));
+    }
+    cg.delete();
+}

tools/create_cgroup.sh

@@ -0,0 +1,19 @@
+#!/bin/sh
+
+CONTROL_GROUPS=`cargo test -- --list 2>/dev/null | egrep 'test$' | egrep -v '^src' | cut -d':' -f1`
+
+echo This script will create a control group in every subsystem of the V1 hierarchy.
+echo For this, we will need your sudo privileges. Please do not trust this shell script and have a look to check that it does something that you are okay with.
+sudo -v
+
+for i in ${CONTROL_GROUPS}
+	do sudo mkdir -p /sys/fs/cgroup/{blkio,cpu,cpuacct,cpuset,devices,freezer,hugetlb,memory,net_cls,net_prio,perf_event,pids}/$i/
+
+done
+echo
+echo We will now set up permissions...
+echo
+
+for i in ${CONTROL_GROUPS}
+	do sudo chown -R ${USER} /sys/fs/cgroup/{blkio,cpu,cpuacct,cpuset,devices,freezer,hugetlb,memory,net_cls,net_prio,perf_event,pids}/$i/
+done

tools/delete_cgroup.sh

@@ -0,0 +1,14 @@
+#!/bin/sh
+
+CONTROL_GROUPS=`cargo test -- --list 2>/dev/null | egrep 'test$' | egrep -v '^src' | cut -d':' -f1`
+
+echo This script will delete the control groups created by the create_cgroup.sh shell script.
+echo
+echo It may spit out some errors, but that is fine.
+echo
+echo For this, we will need your sudo privileges. Please do not trust this shell script and have a look to check that it does something that you are okay with.
+sudo -v
+
+for i in ${CONTROL_GROUPS}
+	do sudo rmdir /sys/fs/cgroup/{blkio,cpu,cpuacct,cpuset,devices,freezer,hugetlb,memory,net_cls,net_prio,perf_event,pids}/$i/
+done