Commit 934cec29 authored by Victor Marmol's avatar Victor Marmol

Updating cAdvisor godep.

This adds support for Docker 1.6.
parent 88cf777d
......@@ -131,11 +131,6 @@
"Rev": "ed6685369740035b0af9675bf9add52d0af7657b"
},
{
"ImportPath": "github.com/docker/docker/pkg/symlink",
"Comment": "v1.4.1-1714-ged66853",
"Rev": "ed6685369740035b0af9675bf9add52d0af7657b"
},
{
"ImportPath": "github.com/docker/docker/pkg/system",
"Comment": "v1.4.1-1714-ged66853",
"Rev": "ed6685369740035b0af9675bf9add52d0af7657b"
......@@ -157,8 +152,8 @@
},
{
"ImportPath": "github.com/docker/libcontainer",
"Comment": "v1.4.0-52-gd7dea0e",
"Rev": "d7dea0e925315bab640115053204c16718839b1e"
"Comment": "v1.4.0-446-gae812bd",
"Rev": "ae812bdca78084dc322037225d170e1883521d87"
},
{
"ImportPath": "github.com/docker/spdystream",
......@@ -212,83 +207,83 @@
},
{
"ImportPath": "github.com/google/cadvisor/api",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/container",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/events",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/fs",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/healthz",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/http",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/info/v1",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/info/v2",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/manager",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/metrics",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/pages",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/storage",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/summary",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/utils",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/validate",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/cadvisor/version",
"Comment": "0.11.0-14-ga1c6887",
"Rev": "a1c688751c991f5002d4d36e8e0682186bc78566"
"Comment": "0.11.0-26-g417cd5e",
"Rev": "417cd5ead23d33eff717a41d3be80ea73bf307cd"
},
{
"ImportPath": "github.com/google/gofuzz",
......@@ -415,6 +410,10 @@
"Rev": "e4ec8152c15fc46bd5056ce65997a07c7d415325"
},
{
"ImportPath": "github.com/syndtr/gocapability/capability",
"Rev": "3c85049eaeb429febe7788d9c7aac42322a377fe"
},
{
"ImportPath": "github.com/vaughan0/go-ini",
"Rev": "a98ad7ee00ec53921f08832bc06ecf7fd600e6a1"
},
......
Copyright (c) 2014-2015 The Docker & Go Authors. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:
* Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above
copyright notice, this list of conditions and the following disclaimer
in the documentation and/or other materials provided with the
distribution.
* Neither the name of Google Inc. nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Package symlink implements EvalSymlinksInScope which is an extension of filepath.EvalSymlinks
from the [Go standard library](https://golang.org/pkg/path/filepath).
The code from filepath.EvalSymlinks has been adapted in fs.go.
Please read the LICENSE.BSD file that governs fs.go and LICENSE.APACHE for fs_test.go.
// Copyright 2012 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE.BSD file.
// This code is a modified version of path/filepath/symlink.go from the Go standard library.
package symlink
import (
"bytes"
"errors"
"os"
"path/filepath"
"strings"
)
// FollowSymlinkInScope is a wrapper around evalSymlinksInScope that returns an absolute path
func FollowSymlinkInScope(path, root string) (string, error) {
path, err := filepath.Abs(path)
if err != nil {
return "", err
}
root, err = filepath.Abs(root)
if err != nil {
return "", err
}
return evalSymlinksInScope(path, root)
}
// evalSymlinksInScope will evaluate symlinks in `path` within a scope `root` and return
// a result guaranteed to be contained within the scope `root`, at the time of the call.
// Symlinks in `root` are not evaluated and left as-is.
// Errors encountered while attempting to evaluate symlinks in path will be returned.
// Non-existing paths are valid and do not constitute an error.
// `path` has to contain `root` as a prefix, or else an error will be returned.
// Trying to break out from `root` does not constitute an error.
//
// Example:
// If /foo/bar -> /outside,
// FollowSymlinkInScope("/foo/bar", "/foo") == "/foo/outside" instead of "/oustide"
//
// IMPORTANT: it is the caller's responsibility to call evalSymlinksInScope *after* relevant symlinks
// are created and not to create subsequently, additional symlinks that could potentially make a
// previously-safe path, unsafe. Example: if /foo/bar does not exist, evalSymlinksInScope("/foo/bar", "/foo")
// would return "/foo/bar". If one makes /foo/bar a symlink to /baz subsequently, then "/foo/bar" should
// no longer be considered safely contained in "/foo".
func evalSymlinksInScope(path, root string) (string, error) {
root = filepath.Clean(root)
if path == root {
return path, nil
}
if !strings.HasPrefix(path, root) {
return "", errors.New("evalSymlinksInScope: " + path + " is not in " + root)
}
const maxIter = 255
originalPath := path
// given root of "/a" and path of "/a/b/../../c" we want path to be "/b/../../c"
path = path[len(root):]
if root == string(filepath.Separator) {
path = string(filepath.Separator) + path
}
if !strings.HasPrefix(path, string(filepath.Separator)) {
return "", errors.New("evalSymlinksInScope: " + path + " is not in " + root)
}
path = filepath.Clean(path)
// consume path by taking each frontmost path element,
// expanding it if it's a symlink, and appending it to b
var b bytes.Buffer
// b here will always be considered to be the "current absolute path inside
// root" when we append paths to it, we also append a slash and use
// filepath.Clean after the loop to trim the trailing slash
for n := 0; path != ""; n++ {
if n > maxIter {
return "", errors.New("evalSymlinksInScope: too many links in " + originalPath)
}
// find next path component, p
i := strings.IndexRune(path, filepath.Separator)
var p string
if i == -1 {
p, path = path, ""
} else {
p, path = path[:i], path[i+1:]
}
if p == "" {
continue
}
// this takes a b.String() like "b/../" and a p like "c" and turns it
// into "/b/../c" which then gets filepath.Cleaned into "/c" and then
// root gets prepended and we Clean again (to remove any trailing slash
// if the first Clean gave us just "/")
cleanP := filepath.Clean(string(filepath.Separator) + b.String() + p)
if cleanP == string(filepath.Separator) {
// never Lstat "/" itself
b.Reset()
continue
}
fullP := filepath.Clean(root + cleanP)
fi, err := os.Lstat(fullP)
if os.IsNotExist(err) {
// if p does not exist, accept it
b.WriteString(p)
b.WriteRune(filepath.Separator)
continue
}
if err != nil {
return "", err
}
if fi.Mode()&os.ModeSymlink == 0 {
b.WriteString(p + string(filepath.Separator))
continue
}
// it's a symlink, put it at the front of path
dest, err := os.Readlink(fullP)
if err != nil {
return "", err
}
if filepath.IsAbs(dest) {
b.Reset()
}
path = dest + string(filepath.Separator) + path
}
// see note above on "fullP := ..." for why this is double-cleaned and
// what's happening here
return filepath.Clean(root + filepath.Clean(string(filepath.Separator)+b.String())), nil
}
image: dockercore/libcontainer
script:
# Setup the DockerInDocker environment.
- /dind
- sed -i 's!docker/docker!docker/libcontainer!' /go/src/github.com/docker/docker/hack/make/.validate
- bash /go/src/github.com/docker/docker/hack/make/validate-dco
- bash /go/src/github.com/docker/docker/hack/make/validate-gofmt
- export GOPATH="$GOPATH:/go:$(pwd)/vendor" # Drone mucks with our GOPATH
- make direct-test
......@@ -7,9 +7,9 @@ RUN go get github.com/docker/docker/pkg/term
# setup a playground for us to spawn containers in
RUN mkdir /busybox && \
curl -sSL 'https://github.com/jpetazzo/docker-busybox/raw/buildroot-2014.02/rootfs.tar' | tar -xC /busybox
curl -sSL 'https://github.com/jpetazzo/docker-busybox/raw/buildroot-2014.11/rootfs.tar' | tar -xC /busybox
RUN curl -sSL https://raw.githubusercontent.com/docker/docker/master/project/dind -o /dind && \
RUN curl -sSL https://raw.githubusercontent.com/docker/docker/master/hack/dind -o /dind && \
chmod +x /dind
COPY . /go/src/github.com/docker/libcontainer
......
......@@ -3,4 +3,5 @@ Rohit Jnagal <jnagal@google.com> (@rjnagal)
Victor Marmol <vmarmol@google.com> (@vmarmol)
Mrunal Patel <mpatel@redhat.com> (@mrunalp)
Alexandr Morozov <lk4d4@docker.com> (@LK4D4)
Daniel, Dao Quang Minh <dqminh89@gmail.com> (@dqminh)
update-vendor.sh: Tianon Gravi <admwiggin@gmail.com> (@tianon)
......@@ -22,3 +22,12 @@ direct-build:
direct-install:
go install -v $(GO_PACKAGES)
local:
go test -v
validate:
hack/validate.sh
binary: all
docker run --rm --privileged -v $(CURDIR)/bundles:/go/bin dockercore/libcontainer make direct-install
libcontainer
Copyright 2012-2014 Docker, Inc.
Copyright 2012-2015 Docker, Inc.
This product includes software developed at Docker, Inc. (http://www.docker.com).
......
......@@ -8,7 +8,7 @@ In the design and development of libcontainer we try to follow these principles:
* Less code is better.
* Fewer components are better. Do you really need to add one more class?
* 50 lines of straightforward, readable code is better than 10 lines of magic that nobody can understand.
* Don't do later what you can do now. "//FIXME: refactor" is not acceptable in new code.
* Don't do later what you can do now. "//TODO: refactor" is not acceptable in new code.
* When hesitating between two options, choose the one that is easier to reverse.
* "No" is temporary; "Yes" is forever. If you're not sure about a new feature, say no. You can change your mind later.
* Containers must be portable to the greatest possible number of machines. Be suspicious of any change which makes machines less interchangeable.
......
## libcontainer - reference implementation for containers [![Build Status](https://ci.dockerproject.com/github.com/docker/libcontainer/status.svg?branch=master)](https://ci.dockerproject.com/github.com/docker/libcontainer)
## libcontainer - reference implementation for containers [![Build Status](https://jenkins.dockerproject.com/buildStatus/icon?job=Libcontainer Master)](https://jenkins.dockerproject.com/job/Libcontainer%20Master/)
### Note on API changes:
Libcontainer provides a native Go implementation for creating containers
with namespaces, cgroups, capabilities, and filesystem access controls.
It allows you to manage the lifecycle of the container performing additional operations
after the container is created.
Please bear with us while we work on making the libcontainer API stable and something that we can support long term. We are currently discussing the API with the community, therefore, if you currently depend on libcontainer please pin your dependency at a specific tag or commit id. Please join the discussion and help shape the API.
#### Background
#### Container
A container is a self contained execution environment that shares the kernel of the
host system and which is (optionally) isolated from other containers in the system.
libcontainer specifies configuration options for what a container is. It provides a native Go implementation for using Linux namespaces with no external dependencies. libcontainer provides many convenience functions for working with namespaces, networking, and management.
#### Using libcontainer
To create a container you first have to initialize an instance of a factory
that will handle the creation and initialization for a container.
#### Container
A container is a self contained execution environment that shares the kernel of the host system and which is (optionally) isolated from other containers in the system.
Because containers are spawned in a two step process you will need to provide
arguments to a binary that will be executed as the init process for the container.
To use the current binary that is spawning the containers and acting as the parent
you can use `os.Args[0]` and we have a command called `init` setup.
```go
root, err := libcontainer.New("/var/lib/container", libcontainer.InitArgs(os.Args[0], "init"))
if err != nil {
log.Fatal(err)
}
```
Once you have an instance of the factory created we can create a configuration
struct describing how the container is to be created. A sample would look similar to this:
```go
config := &configs.Config{
Rootfs: rootfs,
Capabilities: []string{
"CHOWN",
"DAC_OVERRIDE",
"FSETID",
"FOWNER",
"MKNOD",
"NET_RAW",
"SETGID",
"SETUID",
"SETFCAP",
"SETPCAP",
"NET_BIND_SERVICE",
"SYS_CHROOT",
"KILL",
"AUDIT_WRITE",
},
Namespaces: configs.Namespaces([]configs.Namespace{
{Type: configs.NEWNS},
{Type: configs.NEWUTS},
{Type: configs.NEWIPC},
{Type: configs.NEWPID},
{Type: configs.NEWNET},
}),
Cgroups: &configs.Cgroup{
Name: "test-container",
Parent: "system",
AllowAllDevices: false,
AllowedDevices: configs.DefaultAllowedDevices,
},
Devices: configs.DefaultAutoCreatedDevices,
Hostname: "testing",
Networks: []*configs.Network{
{
Type: "loopback",
Address: "127.0.0.1/0",
Gateway: "localhost",
},
},
Rlimits: []configs.Rlimit{
{
Type: syscall.RLIMIT_NOFILE,
Hard: uint64(1024),
Soft: uint64(1024),
},
},
}
```
libcontainer may be used to execute a process in a container. If a user tries to run a new process inside an existing container, the new process is added to the processes executing in the container.
Once you have the configuration populated you can create a container:
```go
container, err := root.Create("container-id", config)
```
To spawn bash as the initial process inside the container and have the
processes pid returned in order to wait, signal, or kill the process:
```go
process := &libcontainer.Process{
Args: []string{"/bin/bash"},
Env: []string{"PATH=/bin"},
User: "daemon",
Stdin: os.Stdin,
Stdout: os.Stdout,
Stderr: os.Stderr,
}
err := container.Start(process)
if err != nil {
log.Fatal(err)
}
// wait for the process to finish.
status, err := process.Wait()
if err != nil {
log.Fatal(err)
}
// destroy the container.
container.Destroy()
```
#### Root file system
Additional ways to interact with a running container are:
A container runs with a directory known as its *root file system*, or *rootfs*, mounted as the file system root. The rootfs is usually a full system tree.
```go
// return all the pids for all processes running inside the container.
processes, err := container.Processes()
// get detailed cpu, memory, io, and network statistics for the container and
// it's processes.
stats, err := container.Stats()
#### Configuration
A container is initially configured by supplying configuration data when the container is created.
// pause all processes inside the container.
container.Pause()
// resume all paused processes.
container.Resume()
```
#### nsinit
`nsinit` is a cli application which demonstrates the use of libcontainer. It is able to spawn new containers or join existing containers, based on the current directory.
`nsinit` is a cli application which demonstrates the use of libcontainer.
It is able to spawn new containers or join existing containers. A root
filesystem must be provided for use along with a container configuration file.
To build `nsinit`, run `make binary`. It will save the binary into
`bundles/nsinit`.
To use `nsinit`, cd into a Linux rootfs and copy a `container.json` file into the directory with your specified configuration. Environment, networking, and different capabilities for the container are specified in this file. The configuration is used for each process executed inside the container.
To use `nsinit`, cd into a Linux rootfs and copy a `container.json` file into
the directory with your specified configuration. Environment, networking,
and different capabilities for the container are specified in this file.
The configuration is used for each process executed inside the container.
See the `sample_configs` folder for examples of what the container configuration should look like.
To execute `/bin/bash` in the current directory as a container just run the following **as root**:
```bash
nsinit exec /bin/bash
nsinit exec --tty /bin/bash
```
If you wish to spawn another process inside the container while your current bash session is running, run the same command again to get another bash shell (or change the command). If the original process (PID 1) dies, all other processes spawned inside the container will be killed and the namespace will be removed.
If you wish to spawn another process inside the container while your
current bash session is running, run the same command again to
get another bash shell (or change the command). If the original
process (PID 1) dies, all other processes spawned inside the container
will be killed and the namespace will be removed.
You can identify if a process is running in a container by looking to see if `state.json` is in the root of the directory.
You can identify if a process is running in a container by
looking to see if `state.json` is in the root of the directory.
You may also specify an alternate root place where the `container.json` file is read and where the `state.json` file will be saved.
You may also specify an alternate root place where
the `container.json` file is read and where the `state.json` file will be saved.
#### Future
See the [roadmap](ROADMAP.md).
......
......@@ -13,4 +13,8 @@ Our goal is to make libcontainer run everywhere, but currently libcontainer requ
## Cross-architecture support
Our goal is to make libcontainer run everywhere. However currently libcontainer only runs on x86_64 systems. We plan on expanding architecture support, so that libcontainer containers can be created and used on more architectures.
Our goal is to make libcontainer run everywhere. Recently libcontainer has
expanded from its initial support for x86_64 systems to include POWER (ppc64
little and big endian variants), IBM System z (s390x 64-bit), and ARM. We plan
to continue expanding architecture support such that libcontainer containers
can be created and used on more architectures.
/*
Temporary API endpoint for libcontainer while the full API is finalized (api.go).
*/
package libcontainer
import (
"github.com/docker/libcontainer/cgroups/fs"
"github.com/docker/libcontainer/network"
)
// TODO(vmarmol): Complete Stats() in final libcontainer API and move users to that.
// DEPRECATED: The below portions are only to be used during the transition to the official API.
// Returns all available stats for the given container.
func GetStats(container *Config, state *State) (stats *ContainerStats, err error) {
stats = &ContainerStats{}
if stats.CgroupStats, err = fs.GetStats(state.CgroupPaths); err != nil {
return stats, err
}
stats.NetworkStats, err = network.GetStats(&state.NetworkState)
return stats, err
}
......@@ -24,7 +24,6 @@ func ApplyProfile(name string) error {
if name == "" {
return nil
}
cName := C.CString(name)
defer C.free(unsafe.Pointer(cName))
......
......@@ -67,12 +67,12 @@ func generateProfile(out io.Writer) error {
data := &data{
Name: "docker-default",
}
if tuntablesExists() {
if tunablesExists() {
data.Imports = append(data.Imports, "#include <tunables/global>")
} else {
data.Imports = append(data.Imports, "@{PROC}=/proc/")
}
if abstrctionsEsists() {
if abstractionsExists() {
data.InnerImports = append(data.InnerImports, "#include <abstractions/base>")
}
if err := compiled.Execute(out, data); err != nil {
......@@ -82,13 +82,13 @@ func generateProfile(out io.Writer) error {
}
// check if the tunables/global exist
func tuntablesExists() bool {
func tunablesExists() bool {
_, err := os.Stat("/etc/apparmor.d/tunables/global")
return err == nil
}
// check if abstractions/base exist
func abstrctionsEsists() bool {
func abstractionsExists() bool {
_, err := os.Stat("/etc/apparmor.d/abstractions/base")
return err == nil
}
// +build linux
package libcontainer
import (
"fmt"
"os"
"github.com/syndtr/gocapability/capability"
)
const allCapabilityTypes = capability.CAPS | capability.BOUNDS
var capabilityList = map[string]capability.Cap{
"SETPCAP": capability.CAP_SETPCAP,
"SYS_MODULE": capability.CAP_SYS_MODULE,
"SYS_RAWIO": capability.CAP_SYS_RAWIO,
"SYS_PACCT": capability.CAP_SYS_PACCT,
"SYS_ADMIN": capability.CAP_SYS_ADMIN,
"SYS_NICE": capability.CAP_SYS_NICE,
"SYS_RESOURCE": capability.CAP_SYS_RESOURCE,
"SYS_TIME": capability.CAP_SYS_TIME,
"SYS_TTY_CONFIG": capability.CAP_SYS_TTY_CONFIG,
"MKNOD": capability.CAP_MKNOD,
"AUDIT_WRITE": capability.CAP_AUDIT_WRITE,
"AUDIT_CONTROL": capability.CAP_AUDIT_CONTROL,
"MAC_OVERRIDE": capability.CAP_MAC_OVERRIDE,
"MAC_ADMIN": capability.CAP_MAC_ADMIN,
"NET_ADMIN": capability.CAP_NET_ADMIN,
"SYSLOG": capability.CAP_SYSLOG,
"CHOWN": capability.CAP_CHOWN,
"NET_RAW": capability.CAP_NET_RAW,
"DAC_OVERRIDE": capability.CAP_DAC_OVERRIDE,
"FOWNER": capability.CAP_FOWNER,
"DAC_READ_SEARCH": capability.CAP_DAC_READ_SEARCH,
"FSETID": capability.CAP_FSETID,
"KILL": capability.CAP_KILL,
"SETGID": capability.CAP_SETGID,
"SETUID": capability.CAP_SETUID,
"LINUX_IMMUTABLE": capability.CAP_LINUX_IMMUTABLE,
"NET_BIND_SERVICE": capability.CAP_NET_BIND_SERVICE,
"NET_BROADCAST": capability.CAP_NET_BROADCAST,
"IPC_LOCK": capability.CAP_IPC_LOCK,
"IPC_OWNER": capability.CAP_IPC_OWNER,
"SYS_CHROOT": capability.CAP_SYS_CHROOT,
"SYS_PTRACE": capability.CAP_SYS_PTRACE,
"SYS_BOOT": capability.CAP_SYS_BOOT,
"LEASE": capability.CAP_LEASE,
"SETFCAP": capability.CAP_SETFCAP,
"WAKE_ALARM": capability.CAP_WAKE_ALARM,
"BLOCK_SUSPEND": capability.CAP_BLOCK_SUSPEND,
}
func newCapWhitelist(caps []string) (*whitelist, error) {
l := []capability.Cap{}
for _, c := range caps {
v, ok := capabilityList[c]
if !ok {
return nil, fmt.Errorf("unknown capability %q", c)
}
l = append(l, v)
}
pid, err := capability.NewPid(os.Getpid())
if err != nil {
return nil, err
}
return &whitelist{
keep: l,
pid: pid,
}, nil
}
type whitelist struct {
pid capability.Capabilities
keep []capability.Cap
}
// dropBoundingSet drops the capability bounding set to those specified in the whitelist.
func (w *whitelist) dropBoundingSet() error {
w.pid.Clear(capability.BOUNDS)
w.pid.Set(capability.BOUNDS, w.keep...)
return w.pid.Apply(capability.BOUNDS)
}
// drop drops all capabilities for the current process except those specified in the whitelist.
func (w *whitelist) drop() error {
w.pid.Clear(allCapabilityTypes)
w.pid.Set(allCapabilityTypes, w.keep...)
return w.pid.Apply(allCapabilityTypes)
}
......@@ -3,16 +3,38 @@ package cgroups
import (
"fmt"
"github.com/docker/libcontainer/devices"
"github.com/docker/libcontainer/configs"
)
type FreezerState string
type Manager interface {
// Apply cgroup configuration to the process with the specified pid
Apply(pid int) error
const (
Undefined FreezerState = ""
Frozen FreezerState = "FROZEN"
Thawed FreezerState = "THAWED"
)
// Returns the PIDs inside the cgroup set
GetPids() ([]int, error)
// Returns statistics for the cgroup set
GetStats() (*Stats, error)
// Toggles the freezer cgroup according with specified state
Freeze(state configs.FreezerState) error
// Destroys the cgroup set
Destroy() error
// NewCgroupManager() and LoadCgroupManager() require following attributes:
// Paths map[string]string
// Cgroups *cgroups.Cgroup
// Paths maps cgroup subsystem to path at which it is mounted.
// Cgroups specifies specific cgroup settings for the various subsystems
// Returns cgroup paths to save in a state file and to be able to
// restore the object later.
GetPaths() map[string]string
// Set the cgroup as configured.
Set(container *configs.Config) error
}
type NotFoundError struct {
Subsystem string
......@@ -32,25 +54,6 @@ func IsNotFound(err error) bool {
if err == nil {
return false
}
_, ok := err.(*NotFoundError)
return ok
}
type Cgroup struct {
Name string `json:"name,omitempty"`
Parent string `json:"parent,omitempty"` // name of parent cgroup or slice
AllowAllDevices bool `json:"allow_all_devices,omitempty"` // If this is true allow access to any kind of device within the container. If false, allow access only to devices explicitly listed in the allowed_devices list.
AllowedDevices []*devices.Device `json:"allowed_devices,omitempty"`
Memory int64 `json:"memory,omitempty"` // Memory limit (in bytes)
MemoryReservation int64 `json:"memory_reservation,omitempty"` // Memory reservation or soft_limit (in bytes)
MemorySwap int64 `json:"memory_swap,omitempty"` // Total memory usage (memory + swap); set `-1' to disable swap
CpuShares int64 `json:"cpu_shares,omitempty"` // CPU shares (relative weight vs. other containers)
CpuQuota int64 `json:"cpu_quota,omitempty"` // CPU hardcap limit (in usecs). Allowed cpu time in a given period.
CpuPeriod int64 `json:"cpu_period,omitempty"` // CPU period to be used for hardcapping (in usecs). 0 to use system default.
CpusetCpus string `json:"cpuset_cpus,omitempty"` // CPU to use
CpusetMems string `json:"cpuset_mems,omitempty"` // MEM to use
Freezer FreezerState `json:"freezer,omitempty"` // set the freeze value for the process
Slice string `json:"slice,omitempty"` // Parent slice to use for systemd
}
package fs
import (
"fmt"
"io/ioutil"
"os"
"path/filepath"
"strconv"
"sync"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
var (
......@@ -24,43 +25,63 @@ var (
CgroupProcesses = "cgroup.procs"
)
type subsystem interface {
// Returns the stats, as 'stats', corresponding to the cgroup under 'path'.
GetStats(path string, stats *cgroups.Stats) error
// Removes the cgroup represented by 'data'.
Remove(*data) error
// Creates and joins the cgroup represented by data.
Apply(*data) error
// Set the cgroup represented by cgroup.
Set(path string, cgroup *configs.Cgroup) error
}
type Manager struct {
Cgroups *configs.Cgroup
Paths map[string]string
}
// The absolute path to the root of the cgroup hierarchies.
var cgroupRootLock sync.Mutex
var cgroupRoot string
// TODO(vmarmol): Report error here, we'll probably need to wait for the new API.
func init() {
// we can pick any subsystem to find the root
cpuRoot, err := cgroups.FindCgroupMountpoint("cpu")
// Gets the cgroupRoot.
func getCgroupRoot() (string, error) {
cgroupRootLock.Lock()
defer cgroupRootLock.Unlock()
if cgroupRoot != "" {
return cgroupRoot, nil
}
root, err := cgroups.FindCgroupMountpointDir()
if err != nil {
return
return "", err
}
cgroupRoot = filepath.Dir(cpuRoot)
if _, err := os.Stat(cgroupRoot); err != nil {
return
if _, err := os.Stat(root); err != nil {
return "", err
}
}
type subsystem interface {
// Returns the stats, as 'stats', corresponding to the cgroup under 'path'.
GetStats(path string, stats *cgroups.Stats) error
// Removes the cgroup represented by 'data'.
Remove(*data) error
// Creates and joins the cgroup represented by data.
Set(*data) error
cgroupRoot = root
return cgroupRoot, nil
}
type data struct {
root string
cgroup string
c *cgroups.Cgroup
c *configs.Cgroup
pid int
}
func Apply(c *cgroups.Cgroup, pid int) (map[string]string, error) {
d, err := getCgroupData(c, pid)
func (m *Manager) Apply(pid int) error {
if m.Cgroups == nil {
return nil
}
d, err := getCgroupData(m.Cgroups, pid)
if err != nil {
return nil, err
return err
}
paths := make(map[string]string)
......@@ -70,10 +91,10 @@ func Apply(c *cgroups.Cgroup, pid int) (map[string]string, error) {
}
}()
for name, sys := range subsystems {
if err := sys.Set(d); err != nil {
return nil, err
if err := sys.Apply(d); err != nil {
return err
}
// FIXME: Apply should, ideally, be reentrant or be broken up into a separate
// TODO: Apply should, ideally, be reentrant or be broken up into a separate
// create and join phase so that the cgroup hierarchy for a container can be
// created then join consists of writing the process pids to cgroup.procs
p, err := d.path(name)
......@@ -81,16 +102,26 @@ func Apply(c *cgroups.Cgroup, pid int) (map[string]string, error) {
if cgroups.IsNotFound(err) {
continue
}
return nil, err
return err
}
paths[name] = p
}
return paths, nil
m.Paths = paths
return nil
}
func (m *Manager) Destroy() error {
return cgroups.RemovePaths(m.Paths)
}
func (m *Manager) GetPaths() map[string]string {
return m.Paths
}
// Symmetrical public function to update device based cgroups. Also available
// in the systemd implementation.
func ApplyDevices(c *cgroups.Cgroup, pid int) error {
func ApplyDevices(c *configs.Cgroup, pid int) error {
d, err := getCgroupData(c, pid)
if err != nil {
return err
......@@ -98,12 +129,12 @@ func ApplyDevices(c *cgroups.Cgroup, pid int) error {
devices := subsystems["devices"]
return devices.Set(d)
return devices.Apply(d)
}
func GetStats(systemPaths map[string]string) (*cgroups.Stats, error) {
func (m *Manager) GetStats() (*cgroups.Stats, error) {
stats := cgroups.NewStats()
for name, path := range systemPaths {
for name, path := range m.Paths {
sys, ok := subsystems[name]
if !ok || !cgroups.PathExists(path) {
continue
......@@ -116,23 +147,48 @@ func GetStats(systemPaths map[string]string) (*cgroups.Stats, error) {
return stats, nil
}
func (m *Manager) Set(container *configs.Config) error {
for name, path := range m.Paths {
sys, ok := subsystems[name]
if !ok || !cgroups.PathExists(path) {
continue
}
if err := sys.Set(path, container.Cgroups); err != nil {
return err
}
}
return nil
}
// Freeze toggles the container's freezer cgroup depending on the state
// provided
func Freeze(c *cgroups.Cgroup, state cgroups.FreezerState) error {
d, err := getCgroupData(c, 0)
func (m *Manager) Freeze(state configs.FreezerState) error {
d, err := getCgroupData(m.Cgroups, 0)
if err != nil {
return err
}
dir, err := d.path("freezer")
if err != nil {
return err
}
c.Freezer = state
prevState := m.Cgroups.Freezer
m.Cgroups.Freezer = state
freezer := subsystems["freezer"]
err = freezer.Set(dir, m.Cgroups)
if err != nil {
m.Cgroups.Freezer = prevState
return err
}
return freezer.Set(d)
return nil
}
func GetPids(c *cgroups.Cgroup) ([]int, error) {
d, err := getCgroupData(c, 0)
func (m *Manager) GetPids() ([]int, error) {
d, err := getCgroupData(m.Cgroups, 0)
if err != nil {
return nil, err
}
......@@ -145,9 +201,10 @@ func GetPids(c *cgroups.Cgroup) ([]int, error) {
return cgroups.ReadProcsFile(dir)
}
func getCgroupData(c *cgroups.Cgroup, pid int) (*data, error) {
if cgroupRoot == "" {
return nil, fmt.Errorf("failed to find the cgroup root")
func getCgroupData(c *configs.Cgroup, pid int) (*data, error) {
root, err := getCgroupRoot()
if err != nil {
return nil, err
}
cgroup := c.Name
......@@ -156,38 +213,34 @@ func getCgroupData(c *cgroups.Cgroup, pid int) (*data, error) {
}
return &data{
root: cgroupRoot,
root: root,
cgroup: cgroup,
c: c,
pid: pid,
}, nil
}
func (raw *data) parent(subsystem string) (string, error) {
func (raw *data) parent(subsystem, mountpoint string) (string, error) {
initPath, err := cgroups.GetInitCgroupDir(subsystem)
if err != nil {
return "", err
}
return filepath.Join(raw.root, subsystem, initPath), nil
return filepath.Join(mountpoint, initPath), nil
}
func (raw *data) path(subsystem string) (string, error) {
mnt, err := cgroups.FindCgroupMountpoint(subsystem)
// If we didn't mount the subsystem, there is no point we make the path.
if err != nil {
return "", err
}
// If the cgroup name/path is absolute do not look relative to the cgroup of the init process.
if filepath.IsAbs(raw.cgroup) {
path := filepath.Join(raw.root, subsystem, raw.cgroup)
if _, err := os.Stat(path); err != nil {
if os.IsNotExist(err) {
return "", cgroups.NewNotFoundError(subsystem)
}
return "", err
}
return path, nil
return filepath.Join(raw.root, subsystem, raw.cgroup), nil
}
parent, err := raw.parent(subsystem)
parent, err := raw.parent(subsystem, mnt)
if err != nil {
return "", err
}
......
......@@ -9,17 +9,38 @@ import (
"strings"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type BlkioGroup struct {
}
func (s *BlkioGroup) Set(d *data) error {
// we just want to join this group even though we don't set anything
if _, err := d.join("blkio"); err != nil && !cgroups.IsNotFound(err) {
func (s *BlkioGroup) Apply(d *data) error {
dir, err := d.join("blkio")
if err != nil && !cgroups.IsNotFound(err) {
return err
}
if err := s.Set(dir, d.c); err != nil {
return err
}
return nil
}
func (s *BlkioGroup) Set(path string, cgroup *configs.Cgroup) error {
if cgroup.BlkioWeight != 0 {
if err := writeFile(path, "blkio.weight", strconv.FormatInt(cgroup.BlkioWeight, 10)); err != nil {
return err
}
}
if cgroup.BlkioWeightDevice != "" {
if err := writeFile(path, "blkio.weight_device", cgroup.BlkioWeightDevice); err != nil {
return err
}
}
return nil
}
......
package fs
import (
"strconv"
"testing"
"github.com/docker/libcontainer/cgroups"
......@@ -72,6 +73,64 @@ func appendBlkioStatEntry(blkioStatEntries *[]cgroups.BlkioStatEntry, major, min
*blkioStatEntries = append(*blkioStatEntries, cgroups.BlkioStatEntry{Major: major, Minor: minor, Value: value, Op: op})
}
func TestBlkioSetWeight(t *testing.T) {
helper := NewCgroupTestUtil("blkio", t)
defer helper.cleanup()
const (
weightBefore = 100
weightAfter = 200
)
helper.writeFileContents(map[string]string{
"blkio.weight": strconv.Itoa(weightBefore),
})
helper.CgroupData.c.BlkioWeight = weightAfter
blkio := &BlkioGroup{}
if err := blkio.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamUint(helper.CgroupPath, "blkio.weight")
if err != nil {
t.Fatalf("Failed to parse blkio.weight - %s", err)
}
if value != weightAfter {
t.Fatal("Got the wrong value, set blkio.weight failed.")
}
}
func TestBlkioSetWeightDevice(t *testing.T) {
helper := NewCgroupTestUtil("blkio", t)
defer helper.cleanup()
const (
weightDeviceBefore = "8:0 400"
weightDeviceAfter = "8:0 500"
)
helper.writeFileContents(map[string]string{
"blkio.weight_device": weightDeviceBefore,
})
helper.CgroupData.c.BlkioWeightDevice = weightDeviceAfter
blkio := &BlkioGroup{}
if err := blkio.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamString(helper.CgroupPath, "blkio.weight_device")
if err != nil {
t.Fatalf("Failed to parse blkio.weight_device - %s", err)
}
if value != weightDeviceAfter {
t.Fatal("Got the wrong value, set blkio.weight_device failed.")
}
}
func TestBlkioStats(t *testing.T) {
helper := NewCgroupTestUtil("blkio", t)
defer helper.cleanup()
......
......@@ -7,33 +7,44 @@ import (
"strconv"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type CpuGroup struct {
}
func (s *CpuGroup) Set(d *data) error {
func (s *CpuGroup) Apply(d *data) error {
// We always want to join the cpu group, to allow fair cpu scheduling
// on a container basis
dir, err := d.join("cpu")
if err != nil {
return err
}
if d.c.CpuShares != 0 {
if err := writeFile(dir, "cpu.shares", strconv.FormatInt(d.c.CpuShares, 10)); err != nil {
if err := s.Set(dir, d.c); err != nil {
return err
}
return nil
}
func (s *CpuGroup) Set(path string, cgroup *configs.Cgroup) error {
if cgroup.CpuShares != 0 {
if err := writeFile(path, "cpu.shares", strconv.FormatInt(cgroup.CpuShares, 10)); err != nil {
return err
}
}
if d.c.CpuPeriod != 0 {
if err := writeFile(dir, "cpu.cfs_period_us", strconv.FormatInt(d.c.CpuPeriod, 10)); err != nil {
if cgroup.CpuPeriod != 0 {
if err := writeFile(path, "cpu.cfs_period_us", strconv.FormatInt(cgroup.CpuPeriod, 10)); err != nil {
return err
}
}
if d.c.CpuQuota != 0 {
if err := writeFile(dir, "cpu.cfs_quota_us", strconv.FormatInt(d.c.CpuQuota, 10)); err != nil {
if cgroup.CpuQuota != 0 {
if err := writeFile(path, "cpu.cfs_quota_us", strconv.FormatInt(cgroup.CpuQuota, 10)); err != nil {
return err
}
}
return nil
}
......
......@@ -2,11 +2,81 @@ package fs
import (
"fmt"
"strconv"
"testing"
"github.com/docker/libcontainer/cgroups"
)
func TestCpuSetShares(t *testing.T) {
helper := NewCgroupTestUtil("cpu", t)
defer helper.cleanup()
const (
sharesBefore = 1024
sharesAfter = 512
)
helper.writeFileContents(map[string]string{
"cpu.shares": strconv.Itoa(sharesBefore),
})
helper.CgroupData.c.CpuShares = sharesAfter
cpu := &CpuGroup{}
if err := cpu.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamUint(helper.CgroupPath, "cpu.shares")
if err != nil {
t.Fatalf("Failed to parse cpu.shares - %s", err)
}
if value != sharesAfter {
t.Fatal("Got the wrong value, set cpu.shares failed.")
}
}
func TestCpuSetBandWidth(t *testing.T) {
helper := NewCgroupTestUtil("cpu", t)
defer helper.cleanup()
const (
quotaBefore = 8000
quotaAfter = 5000
periodBefore = 10000
periodAfter = 7000
)
helper.writeFileContents(map[string]string{
"cpu.cfs_quota_us": strconv.Itoa(quotaBefore),
"cpu.cfs_period_us": strconv.Itoa(periodBefore),
})
helper.CgroupData.c.CpuQuota = quotaAfter
helper.CgroupData.c.CpuPeriod = periodAfter
cpu := &CpuGroup{}
if err := cpu.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
quota, err := getCgroupParamUint(helper.CgroupPath, "cpu.cfs_quota_us")
if err != nil {
t.Fatalf("Failed to parse cpu.cfs_quota_us - %s", err)
}
if quota != quotaAfter {
t.Fatal("Got the wrong value, set cpu.cfs_quota_us failed.")
}
period, err := getCgroupParamUint(helper.CgroupPath, "cpu.cfs_period_us")
if err != nil {
t.Fatalf("Failed to parse cpu.cfs_period_us - %s", err)
}
if period != periodAfter {
t.Fatal("Got the wrong value, set cpu.cfs_period_us failed.")
}
}
func TestCpuStats(t *testing.T) {
helper := NewCgroupTestUtil("cpu", t)
defer helper.cleanup()
......
......@@ -8,6 +8,7 @@ import (
"strings"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
"github.com/docker/libcontainer/system"
)
......@@ -21,7 +22,7 @@ var clockTicks = uint64(system.GetClockTicks())
type CpuacctGroup struct {
}
func (s *CpuacctGroup) Set(d *data) error {
func (s *CpuacctGroup) Apply(d *data) error {
// we just want to join this group even though we don't set anything
if _, err := d.join("cpuacct"); err != nil && !cgroups.IsNotFound(err) {
return err
......@@ -30,6 +31,10 @@ func (s *CpuacctGroup) Set(d *data) error {
return nil
}
func (s *CpuacctGroup) Set(path string, cgroup *configs.Cgroup) error {
return nil
}
func (s *CpuacctGroup) Remove(d *data) error {
return removePath(d.path("cpuacct"))
}
......
......@@ -8,17 +8,35 @@ import (
"strconv"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type CpusetGroup struct {
}
func (s *CpusetGroup) Set(d *data) error {
func (s *CpusetGroup) Apply(d *data) error {
dir, err := d.path("cpuset")
if err != nil {
return err
}
return s.SetDir(dir, d.c.CpusetCpus, d.c.CpusetMems, d.pid)
return s.ApplyDir(dir, d.c, d.pid)
}
func (s *CpusetGroup) Set(path string, cgroup *configs.Cgroup) error {
if cgroup.CpusetCpus != "" {
if err := writeFile(path, "cpuset.cpus", cgroup.CpusetCpus); err != nil {
return err
}
}
if cgroup.CpusetMems != "" {
if err := writeFile(path, "cpuset.mems", cgroup.CpusetMems); err != nil {
return err
}
}
return nil
}
func (s *CpusetGroup) Remove(d *data) error {
......@@ -29,7 +47,7 @@ func (s *CpusetGroup) GetStats(path string, stats *cgroups.Stats) error {
return nil
}
func (s *CpusetGroup) SetDir(dir, cpus string, mems string, pid int) error {
func (s *CpusetGroup) ApplyDir(dir string, cgroup *configs.Cgroup, pid int) error {
if err := s.ensureParent(dir); err != nil {
return err
}
......@@ -40,17 +58,10 @@ func (s *CpusetGroup) SetDir(dir, cpus string, mems string, pid int) error {
return err
}
// If we don't use --cpuset-xxx, the default value inherit from parent cgroup
// is set in s.ensureParent, otherwise, use the value we set
if cpus != "" {
if err := writeFile(dir, "cpuset.cpus", cpus); err != nil {
return err
}
}
if mems != "" {
if err := writeFile(dir, "cpuset.mems", mems); err != nil {
return err
}
// the default values inherit from parent cgroup are already set in
// s.ensureParent, cover these if we have our own
if err := s.Set(dir, cgroup); err != nil {
return err
}
return nil
......
package fs
import (
"testing"
)
func TestCpusetSetCpus(t *testing.T) {
helper := NewCgroupTestUtil("cpuset", t)
defer helper.cleanup()
const (
cpusBefore = "0"
cpusAfter = "1-3"
)
helper.writeFileContents(map[string]string{
"cpuset.cpus": cpusBefore,
})
helper.CgroupData.c.CpusetCpus = cpusAfter
cpuset := &CpusetGroup{}
if err := cpuset.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamString(helper.CgroupPath, "cpuset.cpus")
if err != nil {
t.Fatalf("Failed to parse cpuset.cpus - %s", err)
}
if value != cpusAfter {
t.Fatal("Got the wrong value, set cpuset.cpus failed.")
}
}
func TestCpusetSetMems(t *testing.T) {
helper := NewCgroupTestUtil("cpuset", t)
defer helper.cleanup()
const (
memsBefore = "0"
memsAfter = "1"
)
helper.writeFileContents(map[string]string{
"cpuset.mems": memsBefore,
})
helper.CgroupData.c.CpusetMems = memsAfter
cpuset := &CpusetGroup{}
if err := cpuset.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamString(helper.CgroupPath, "cpuset.mems")
if err != nil {
t.Fatalf("Failed to parse cpuset.mems - %s", err)
}
if value != memsAfter {
t.Fatal("Got the wrong value, set cpuset.mems failed.")
}
}
package fs
import "github.com/docker/libcontainer/cgroups"
import (
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type DevicesGroup struct {
}
func (s *DevicesGroup) Set(d *data) error {
func (s *DevicesGroup) Apply(d *data) error {
dir, err := d.join("devices")
if err != nil {
return err
}
if !d.c.AllowAllDevices {
if err := writeFile(dir, "devices.deny", "a"); err != nil {
if err := s.Set(dir, d.c); err != nil {
return err
}
return nil
}
func (s *DevicesGroup) Set(path string, cgroup *configs.Cgroup) error {
if !cgroup.AllowAllDevices {
if err := writeFile(path, "devices.deny", "a"); err != nil {
return err
}
for _, dev := range d.c.AllowedDevices {
if err := writeFile(dir, "devices.allow", dev.GetCgroupAllowString()); err != nil {
for _, dev := range cgroup.AllowedDevices {
if err := writeFile(path, "devices.allow", dev.CgroupString()); err != nil {
return err
}
}
}
return nil
}
......
package fs
import (
"testing"
"github.com/docker/libcontainer/configs"
)
var (
allowedDevices = []*configs.Device{
{
Path: "/dev/zero",
Type: 'c',
Major: 1,
Minor: 5,
Permissions: "rwm",
FileMode: 0666,
},
}
allowedList = "c 1:5 rwm"
)
func TestDevicesSetAllow(t *testing.T) {
helper := NewCgroupTestUtil("devices", t)
defer helper.cleanup()
helper.writeFileContents(map[string]string{
"devices.deny": "a",
})
helper.CgroupData.c.AllowAllDevices = false
helper.CgroupData.c.AllowedDevices = allowedDevices
devices := &DevicesGroup{}
if err := devices.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamString(helper.CgroupPath, "devices.allow")
if err != nil {
t.Fatalf("Failed to parse devices.allow - %s", err)
}
if value != allowedList {
t.Fatal("Got the wrong value, set devices.allow failed.")
}
}
package fs
import (
"fmt"
"strings"
"time"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type FreezerGroup struct {
}
func (s *FreezerGroup) Set(d *data) error {
switch d.c.Freezer {
case cgroups.Frozen, cgroups.Thawed:
dir, err := d.path("freezer")
if err != nil {
return err
}
func (s *FreezerGroup) Apply(d *data) error {
dir, err := d.join("freezer")
if err != nil && !cgroups.IsNotFound(err) {
return err
}
if err := s.Set(dir, d.c); err != nil {
return err
}
if err := writeFile(dir, "freezer.state", string(d.c.Freezer)); err != nil {
return nil
}
func (s *FreezerGroup) Set(path string, cgroup *configs.Cgroup) error {
switch cgroup.Freezer {
case configs.Frozen, configs.Thawed:
if err := writeFile(path, "freezer.state", string(cgroup.Freezer)); err != nil {
return err
}
for {
state, err := readFile(dir, "freezer.state")
state, err := readFile(path, "freezer.state")
if err != nil {
return err
}
if strings.TrimSpace(state) == string(d.c.Freezer) {
if strings.TrimSpace(state) == string(cgroup.Freezer) {
break
}
time.Sleep(1 * time.Millisecond)
}
case configs.Undefined:
return nil
default:
if _, err := d.join("freezer"); err != nil && !cgroups.IsNotFound(err) {
return err
}
return fmt.Errorf("Invalid argument '%s' to freezer.state", string(cgroup.Freezer))
}
return nil
......
package fs
import (
"testing"
"github.com/docker/libcontainer/configs"
)
func TestFreezerSetState(t *testing.T) {
helper := NewCgroupTestUtil("freezer", t)
defer helper.cleanup()
helper.writeFileContents(map[string]string{
"freezer.state": string(configs.Frozen),
})
helper.CgroupData.c.Freezer = configs.Thawed
freezer := &FreezerGroup{}
if err := freezer.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamString(helper.CgroupPath, "freezer.state")
if err != nil {
t.Fatalf("Failed to parse freezer.state - %s", err)
}
if value != string(configs.Thawed) {
t.Fatal("Got the wrong value, set freezer.state failed.")
}
}
func TestFreezerSetInvalidState(t *testing.T) {
helper := NewCgroupTestUtil("freezer", t)
defer helper.cleanup()
const (
invalidArg configs.FreezerState = "Invalid"
)
helper.CgroupData.c.Freezer = invalidArg
freezer := &FreezerGroup{}
if err := freezer.Set(helper.CgroupPath, helper.CgroupData.c); err == nil {
t.Fatal("Failed to return invalid argument error")
}
}
......@@ -8,12 +8,13 @@ import (
"strconv"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type MemoryGroup struct {
}
func (s *MemoryGroup) Set(d *data) error {
func (s *MemoryGroup) Apply(d *data) error {
dir, err := d.join("memory")
// only return an error for memory if it was specified
if err != nil && (d.c.Memory != 0 || d.c.MemoryReservation != 0 || d.c.MemorySwap != 0) {
......@@ -25,31 +26,42 @@ func (s *MemoryGroup) Set(d *data) error {
}
}()
// Only set values if some config was specified.
if d.c.Memory != 0 || d.c.MemoryReservation != 0 || d.c.MemorySwap != 0 {
if d.c.Memory != 0 {
if err := writeFile(dir, "memory.limit_in_bytes", strconv.FormatInt(d.c.Memory, 10)); err != nil {
return err
}
if err := s.Set(dir, d.c); err != nil {
return err
}
return nil
}
func (s *MemoryGroup) Set(path string, cgroup *configs.Cgroup) error {
if cgroup.Memory != 0 {
if err := writeFile(path, "memory.limit_in_bytes", strconv.FormatInt(cgroup.Memory, 10)); err != nil {
return err
}
if d.c.MemoryReservation != 0 {
if err := writeFile(dir, "memory.soft_limit_in_bytes", strconv.FormatInt(d.c.MemoryReservation, 10)); err != nil {
return err
}
}
if cgroup.MemoryReservation != 0 {
if err := writeFile(path, "memory.soft_limit_in_bytes", strconv.FormatInt(cgroup.MemoryReservation, 10)); err != nil {
return err
}
// By default, MemorySwap is set to twice the size of RAM.
// If you want to omit MemorySwap, set it to '-1'.
if d.c.MemorySwap == 0 {
if err := writeFile(dir, "memory.memsw.limit_in_bytes", strconv.FormatInt(d.c.Memory*2, 10)); err != nil {
return err
}
}
// By default, MemorySwap is set to twice the size of Memory.
if cgroup.MemorySwap == 0 && cgroup.Memory != 0 {
if err := writeFile(path, "memory.memsw.limit_in_bytes", strconv.FormatInt(cgroup.Memory*2, 10)); err != nil {
return err
}
}
if cgroup.MemorySwap > 0 {
if err := writeFile(path, "memory.memsw.limit_in_bytes", strconv.FormatInt(cgroup.MemorySwap, 10)); err != nil {
return err
}
if d.c.MemorySwap > 0 {
if err := writeFile(dir, "memory.memsw.limit_in_bytes", strconv.FormatInt(d.c.MemorySwap, 10)); err != nil {
return err
}
}
if cgroup.OomKillDisable {
if err := writeFile(path, "memory.oom_control", "1"); err != nil {
return err
}
}
return nil
}
......@@ -83,6 +95,7 @@ func (s *MemoryGroup) GetStats(path string, stats *cgroups.Stats) error {
return fmt.Errorf("failed to parse memory.usage_in_bytes - %v", err)
}
stats.MemoryStats.Usage = value
stats.MemoryStats.Cache = stats.MemoryStats.Stats["cache"]
value, err = getCgroupParamUint(path, "memory.max_usage_in_bytes")
if err != nil {
return fmt.Errorf("failed to parse memory.max_usage_in_bytes - %v", err)
......
package fs
import (
"strconv"
"testing"
"github.com/docker/libcontainer/cgroups"
......@@ -14,6 +15,103 @@ rss 1024`
memoryFailcnt = "100\n"
)
func TestMemorySetMemory(t *testing.T) {
helper := NewCgroupTestUtil("memory", t)
defer helper.cleanup()
const (
memoryBefore = 314572800 // 300M
memoryAfter = 524288000 // 500M
reservationBefore = 209715200 // 200M
reservationAfter = 314572800 // 300M
)
helper.writeFileContents(map[string]string{
"memory.limit_in_bytes": strconv.Itoa(memoryBefore),
"memory.soft_limit_in_bytes": strconv.Itoa(reservationBefore),
})
helper.CgroupData.c.Memory = memoryAfter
helper.CgroupData.c.MemoryReservation = reservationAfter
memory := &MemoryGroup{}
if err := memory.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamUint(helper.CgroupPath, "memory.limit_in_bytes")
if err != nil {
t.Fatalf("Failed to parse memory.limit_in_bytes - %s", err)
}
if value != memoryAfter {
t.Fatal("Got the wrong value, set memory.limit_in_bytes failed.")
}
value, err = getCgroupParamUint(helper.CgroupPath, "memory.soft_limit_in_bytes")
if err != nil {
t.Fatalf("Failed to parse memory.soft_limit_in_bytes - %s", err)
}
if value != reservationAfter {
t.Fatal("Got the wrong value, set memory.soft_limit_in_bytes failed.")
}
}
func TestMemorySetMemoryswap(t *testing.T) {
helper := NewCgroupTestUtil("memory", t)
defer helper.cleanup()
const (
memoryswapBefore = 314572800 // 300M
memoryswapAfter = 524288000 // 500M
)
helper.writeFileContents(map[string]string{
"memory.memsw.limit_in_bytes": strconv.Itoa(memoryswapBefore),
})
helper.CgroupData.c.MemorySwap = memoryswapAfter
memory := &MemoryGroup{}
if err := memory.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamUint(helper.CgroupPath, "memory.memsw.limit_in_bytes")
if err != nil {
t.Fatalf("Failed to parse memory.memsw.limit_in_bytes - %s", err)
}
if value != memoryswapAfter {
t.Fatal("Got the wrong value, set memory.memsw.limit_in_bytes failed.")
}
}
func TestMemorySetMemoryswapDefault(t *testing.T) {
helper := NewCgroupTestUtil("memory", t)
defer helper.cleanup()
const (
memoryBefore = 209715200 // 200M
memoryAfter = 314572800 // 300M
memoryswapAfter = 629145600 // 300M*2
)
helper.writeFileContents(map[string]string{
"memory.limit_in_bytes": strconv.Itoa(memoryBefore),
})
helper.CgroupData.c.Memory = memoryAfter
memory := &MemoryGroup{}
if err := memory.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamUint(helper.CgroupPath, "memory.memsw.limit_in_bytes")
if err != nil {
t.Fatalf("Failed to parse memory.memsw.limit_in_bytes - %s", err)
}
if value != memoryswapAfter {
t.Fatal("Got the wrong value, set memory.memsw.limit_in_bytes failed.")
}
}
func TestMemoryStats(t *testing.T) {
helper := NewCgroupTestUtil("memory", t)
defer helper.cleanup()
......@@ -30,7 +128,7 @@ func TestMemoryStats(t *testing.T) {
if err != nil {
t.Fatal(err)
}
expectedStats := cgroups.MemoryStats{Usage: 2048, MaxUsage: 4096, Failcnt: 100, Stats: map[string]uint64{"cache": 512, "rss": 1024}}
expectedStats := cgroups.MemoryStats{Usage: 2048, Cache: 512, MaxUsage: 4096, Failcnt: 100, Stats: map[string]uint64{"cache": 512, "rss": 1024}}
expectMemoryStatEquals(t, expectedStats, actualStats.MemoryStats)
}
......@@ -132,3 +230,30 @@ func TestMemoryStatsBadMaxUsageFile(t *testing.T) {
t.Fatal("Expected failure")
}
}
func TestMemorySetOomControl(t *testing.T) {
helper := NewCgroupTestUtil("memory", t)
defer helper.cleanup()
const (
oom_kill_disable = 1 // disable oom killer, default is 0
)
helper.writeFileContents(map[string]string{
"memory.oom_control": strconv.Itoa(oom_kill_disable),
})
memory := &MemoryGroup{}
if err := memory.Set(helper.CgroupPath, helper.CgroupData.c); err != nil {
t.Fatal(err)
}
value, err := getCgroupParamUint(helper.CgroupPath, "memory.oom_control")
if err != nil {
t.Fatalf("Failed to parse memory.oom_control - %s", err)
}
if value != oom_kill_disable {
t.Fatalf("Got the wrong value, set memory.oom_control failed.")
}
}
......@@ -2,12 +2,13 @@ package fs
import (
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type PerfEventGroup struct {
}
func (s *PerfEventGroup) Set(d *data) error {
func (s *PerfEventGroup) Apply(d *data) error {
// we just want to join this group even though we don't set anything
if _, err := d.join("perf_event"); err != nil && !cgroups.IsNotFound(err) {
return err
......@@ -15,6 +16,10 @@ func (s *PerfEventGroup) Set(d *data) error {
return nil
}
func (s *PerfEventGroup) Set(path string, cgroup *configs.Cgroup) error {
return nil
}
func (s *PerfEventGroup) Remove(d *data) error {
return removePath(d.path("perf_event"))
}
......
......@@ -6,10 +6,12 @@ Creates a mock of the cgroup filesystem for the duration of the test.
package fs
import (
"fmt"
"io/ioutil"
"os"
"path/filepath"
"testing"
"github.com/docker/libcontainer/configs"
)
type cgroupTestUtil struct {
......@@ -26,13 +28,15 @@ type cgroupTestUtil struct {
// Creates a new test util for the specified subsystem
func NewCgroupTestUtil(subsystem string, t *testing.T) *cgroupTestUtil {
d := &data{}
tempDir, err := ioutil.TempDir("", fmt.Sprintf("%s_cgroup_test", subsystem))
d := &data{
c: &configs.Cgroup{},
}
tempDir, err := ioutil.TempDir("", "cgroup_test")
if err != nil {
t.Fatal(err)
}
d.root = tempDir
testCgroupPath, err := d.path(subsystem)
testCgroupPath := filepath.Join(d.root, subsystem)
if err != nil {
t.Fatal(err)
}
......
......@@ -60,3 +60,13 @@ func getCgroupParamUint(cgroupPath, cgroupFile string) (uint64, error) {
return parseUint(strings.TrimSpace(string(contents)), 10, 64)
}
// Gets a string value from the specified cgroup file
func getCgroupParamString(cgroupPath, cgroupFile string) (string, error) {
contents, err := ioutil.ReadFile(filepath.Join(cgroupPath, cgroupFile))
if err != nil {
return "", err
}
return strings.TrimSpace(string(contents)), nil
}
......@@ -33,6 +33,8 @@ type CpuStats struct {
type MemoryStats struct {
// current res_counter usage for memory
Usage uint64 `json:"usage,omitempty"`
// memory used for cache
Cache uint64 `json:"cache,omitempty"`
// maximum usage ever recorded.
MaxUsage uint64 `json:"max_usage,omitempty"`
// TODO(vishh): Export these as stronger types.
......
......@@ -6,24 +6,50 @@ import (
"fmt"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type Manager struct {
Cgroups *configs.Cgroup
Paths map[string]string
}
func UseSystemd() bool {
return false
}
func Apply(c *cgroups.Cgroup, pid int) (map[string]string, error) {
func (m *Manager) Apply(pid int) error {
return fmt.Errorf("Systemd not supported")
}
func (m *Manager) GetPids() ([]int, error) {
return nil, fmt.Errorf("Systemd not supported")
}
func (m *Manager) Destroy() error {
return fmt.Errorf("Systemd not supported")
}
func (m *Manager) GetPaths() map[string]string {
return nil
}
func (m *Manager) GetStats() (*cgroups.Stats, error) {
return nil, fmt.Errorf("Systemd not supported")
}
func GetPids(c *cgroups.Cgroup) ([]int, error) {
func (m *Manager) Set(container *configs.Config) error {
return nil, fmt.Errorf("Systemd not supported")
}
func ApplyDevices(c *cgroups.Cgroup, pid int) error {
func (m *Manager) Freeze(state configs.FreezerState) error {
return fmt.Errorf("Systemd not supported")
}
func ApplyDevices(c *configs.Cgroup, pid int) error {
return fmt.Errorf("Systemd not supported")
}
func Freeze(c *cgroups.Cgroup, state cgroups.FreezerState) error {
func Freeze(c *configs.Cgroup, state configs.FreezerState) error {
return fmt.Errorf("Systemd not supported")
}
......@@ -34,6 +34,21 @@ func FindCgroupMountpoint(subsystem string) (string, error) {
return "", NewNotFoundError(subsystem)
}
func FindCgroupMountpointDir() (string, error) {
mounts, err := mount.GetMounts()
if err != nil {
return "", err
}
for _, mount := range mounts {
if mount.Fstype == "cgroup" {
return filepath.Dir(mount.Mountpoint), nil
}
}
return "", NewNotFoundError("cgroup")
}
type Mount struct {
Mountpoint string
Subsystems []string
......
package libcontainer
import (
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/mount"
"github.com/docker/libcontainer/network"
)
type MountConfig mount.MountConfig
type Network network.Network
type NamespaceType string
const (
NEWNET NamespaceType = "NEWNET"
NEWPID NamespaceType = "NEWPID"
NEWNS NamespaceType = "NEWNS"
NEWUTS NamespaceType = "NEWUTS"
NEWIPC NamespaceType = "NEWIPC"
NEWUSER NamespaceType = "NEWUSER"
)
// Namespace defines configuration for each namespace. It specifies an
// alternate path that is able to be joined via setns.
type Namespace struct {
Type NamespaceType `json:"type"`
Path string `json:"path,omitempty"`
}
type Namespaces []Namespace
func (n *Namespaces) Remove(t NamespaceType) bool {
i := n.index(t)
if i == -1 {
return false
}
*n = append((*n)[:i], (*n)[i+1:]...)
return true
}
func (n *Namespaces) Add(t NamespaceType, path string) {
i := n.index(t)
if i == -1 {
*n = append(*n, Namespace{Type: t, Path: path})
return
}
(*n)[i].Path = path
}
func (n *Namespaces) index(t NamespaceType) int {
for i, ns := range *n {
if ns.Type == t {
return i
}
}
return -1
}
func (n *Namespaces) Contains(t NamespaceType) bool {
return n.index(t) != -1
}
// Config defines configuration options for executing a process inside a contained environment.
type Config struct {
// Mount specific options.
MountConfig *MountConfig `json:"mount_config,omitempty"`
// Pathname to container's root filesystem
RootFs string `json:"root_fs,omitempty"`
// Hostname optionally sets the container's hostname if provided
Hostname string `json:"hostname,omitempty"`
// User will set the uid and gid of the executing process running inside the container
User string `json:"user,omitempty"`
// WorkingDir will change the processes current working directory inside the container's rootfs
WorkingDir string `json:"working_dir,omitempty"`
// Env will populate the processes environment with the provided values
// Any values from the parent processes will be cleared before the values
// provided in Env are provided to the process
Env []string `json:"environment,omitempty"`
// Tty when true will allocate a pty slave on the host for access by the container's process
// and ensure that it is mounted inside the container's rootfs
Tty bool `json:"tty,omitempty"`
// Namespaces specifies the container's namespaces that it should setup when cloning the init process
// If a namespace is not provided that namespace is shared from the container's parent process
Namespaces Namespaces `json:"namespaces,omitempty"`
// Capabilities specify the capabilities to keep when executing the process inside the container
// All capbilities not specified will be dropped from the processes capability mask
Capabilities []string `json:"capabilities,omitempty"`
// Networks specifies the container's network setup to be created
Networks []*Network `json:"networks,omitempty"`
// Routes can be specified to create entries in the route table as the container is started
Routes []*Route `json:"routes,omitempty"`
// Cgroups specifies specific cgroup settings for the various subsystems that the container is
// placed into to limit the resources the container has available
Cgroups *cgroups.Cgroup `json:"cgroups,omitempty"`
// AppArmorProfile specifies the profile to apply to the process running in the container and is
// change at the time the process is execed
AppArmorProfile string `json:"apparmor_profile,omitempty"`
// ProcessLabel specifies the label to apply to the process running in the container. It is
// commonly used by selinux
ProcessLabel string `json:"process_label,omitempty"`
// RestrictSys will remount /proc/sys, /sys, and mask over sysrq-trigger as well as /proc/irq and
// /proc/bus
RestrictSys bool `json:"restrict_sys,omitempty"`
// Rlimits specifies the resource limits, such as max open files, to set in the container
// If Rlimits are not set, the container will inherit rlimits from the parent process
Rlimits []Rlimit `json:"rlimits,omitempty"`
}
// Routes can be specified to create entries in the route table as the container is started
//
// All of destination, source, and gateway should be either IPv4 or IPv6.
// One of the three options must be present, and ommitted entries will use their
// IP family default for the route table. For IPv4 for example, setting the
// gateway to 1.2.3.4 and the interface to eth0 will set up a standard
// destination of 0.0.0.0(or *) when viewed in the route table.
type Route struct {
// Sets the destination and mask, should be a CIDR. Accepts IPv4 and IPv6
Destination string `json:"destination,omitempty"`
// Sets the source and mask, should be a CIDR. Accepts IPv4 and IPv6
Source string `json:"source,omitempty"`
// Sets the gateway. Accepts IPv4 and IPv6
Gateway string `json:"gateway,omitempty"`
// The device to set this route up for, for example: eth0
InterfaceName string `json:"interface_name,omitempty"`
}
type Rlimit struct {
Type int `json:"type,omitempty"`
Hard uint64 `json:"hard,omitempty"`
Soft uint64 `json:"soft,omitempty"`
}
package configs
type FreezerState string
const (
Undefined FreezerState = ""
Frozen FreezerState = "FROZEN"
Thawed FreezerState = "THAWED"
)
type Cgroup struct {
Name string `json:"name"`
// name of parent cgroup or slice
Parent string `json:"parent"`
// If this is true allow access to any kind of device within the container. If false, allow access only to devices explicitly listed in the allowed_devices list.
AllowAllDevices bool `json:"allow_all_devices"`
AllowedDevices []*Device `json:"allowed_devices"`
// Memory limit (in bytes)
Memory int64 `json:"memory"`
// Memory reservation or soft_limit (in bytes)
MemoryReservation int64 `json:"memory_reservation"`
// Total memory usage (memory + swap); set `-1' to disable swap
MemorySwap int64 `json:"memory_swap"`
// CPU shares (relative weight vs. other containers)
CpuShares int64 `json:"cpu_shares"`
// CPU hardcap limit (in usecs). Allowed cpu time in a given period.
CpuQuota int64 `json:"cpu_quota"`
// CPU period to be used for hardcapping (in usecs). 0 to use system default.
CpuPeriod int64 `json:"cpu_period"`
// CPU to use
CpusetCpus string `json:"cpuset_cpus"`
// MEM to use
CpusetMems string `json:"cpuset_mems"`
// Specifies per cgroup weight, range is from 10 to 1000.
BlkioWeight int64 `json:"blkio_weight"`
// Weight per cgroup per device, can override BlkioWeight.
BlkioWeightDevice string `json:"blkio_weight_device"`
// set the freeze value for the process
Freezer FreezerState `json:"freezer"`
// Parent slice to use for systemd TODO: remove in favor or parent
Slice string `json:"slice"`
// Whether to disable OOM Killer
OomKillDisable bool `json:"oom_kill_disable"`
}
package configs
import "fmt"
type Rlimit struct {
Type int `json:"type"`
Hard uint64 `json:"hard"`
Soft uint64 `json:"soft"`
}
// IDMap represents UID/GID Mappings for User Namespaces.
type IDMap struct {
ContainerID int `json:"container_id"`
HostID int `json:"host_id"`
Size int `json:"size"`
}
// Config defines configuration options for executing a process inside a contained environment.
type Config struct {
// NoPivotRoot will use MS_MOVE and a chroot to jail the process into the container's rootfs
// This is a common option when the container is running in ramdisk
NoPivotRoot bool `json:"no_pivot_root"`
// ParentDeathSignal specifies the signal that is sent to the container's process in the case
// that the parent process dies.
ParentDeathSignal int `json:"parent_death_signal"`
// PivotDir allows a custom directory inside the container's root filesystem to be used as pivot, when NoPivotRoot is not set.
// When a custom PivotDir not set, a temporary dir inside the root filesystem will be used. The pivot dir needs to be writeable.
// This is required when using read only root filesystems. In these cases, a read/writeable path can be (bind) mounted somewhere inside the root filesystem to act as pivot.
PivotDir string `json:"pivot_dir"`
// Path to a directory containing the container's root filesystem.
Rootfs string `json:"rootfs"`
// Readonlyfs will remount the container's rootfs as readonly where only externally mounted
// bind mounts are writtable.
Readonlyfs bool `json:"readonlyfs"`
// Mounts specify additional source and destination paths that will be mounted inside the container's
// rootfs and mount namespace if specified
Mounts []*Mount `json:"mounts"`
// The device nodes that should be automatically created within the container upon container start. Note, make sure that the node is marked as allowed in the cgroup as well!
Devices []*Device `json:"devices"`
MountLabel string `json:"mount_label"`
// Hostname optionally sets the container's hostname if provided
Hostname string `json:"hostname"`
// Namespaces specifies the container's namespaces that it should setup when cloning the init process
// If a namespace is not provided that namespace is shared from the container's parent process
Namespaces Namespaces `json:"namespaces"`
// Capabilities specify the capabilities to keep when executing the process inside the container
// All capbilities not specified will be dropped from the processes capability mask
Capabilities []string `json:"capabilities"`
// Networks specifies the container's network setup to be created
Networks []*Network `json:"networks"`
// Routes can be specified to create entries in the route table as the container is started
Routes []*Route `json:"routes"`
// Cgroups specifies specific cgroup settings for the various subsystems that the container is
// placed into to limit the resources the container has available
Cgroups *Cgroup `json:"cgroups"`
// AppArmorProfile specifies the profile to apply to the process running in the container and is
// change at the time the process is execed
AppArmorProfile string `json:"apparmor_profile"`
// ProcessLabel specifies the label to apply to the process running in the container. It is
// commonly used by selinux
ProcessLabel string `json:"process_label"`
// Rlimits specifies the resource limits, such as max open files, to set in the container
// If Rlimits are not set, the container will inherit rlimits from the parent process
Rlimits []Rlimit `json:"rlimits"`
// AdditionalGroups specifies the gids that should be added to supplementary groups
// in addition to those that the user belongs to.
AdditionalGroups []int `json:"additional_groups"`
// UidMappings is an array of User ID mappings for User Namespaces
UidMappings []IDMap `json:"uid_mappings"`
// GidMappings is an array of Group ID mappings for User Namespaces
GidMappings []IDMap `json:"gid_mappings"`
// MaskPaths specifies paths within the container's rootfs to mask over with a bind
// mount pointing to /dev/null as to prevent reads of the file.
MaskPaths []string `json:"mask_paths"`
// ReadonlyPaths specifies paths within the container's rootfs to remount as read-only
// so that these files prevent any writes.
ReadonlyPaths []string `json:"readonly_paths"`
}
// Gets the root uid for the process on host which could be non-zero
// when user namespaces are enabled.
func (c Config) HostUID() (int, error) {
if c.Namespaces.Contains(NEWUSER) {
if c.UidMappings == nil {
return -1, fmt.Errorf("User namespaces enabled, but no user mappings found.")
}
id, found := c.hostIDFromMapping(0, c.UidMappings)
if !found {
return -1, fmt.Errorf("User namespaces enabled, but no root user mapping found.")
}
return id, nil
}
// Return default root uid 0
return 0, nil
}
// Gets the root uid for the process on host which could be non-zero
// when user namespaces are enabled.
func (c Config) HostGID() (int, error) {
if c.Namespaces.Contains(NEWUSER) {
if c.GidMappings == nil {
return -1, fmt.Errorf("User namespaces enabled, but no gid mappings found.")
}
id, found := c.hostIDFromMapping(0, c.GidMappings)
if !found {
return -1, fmt.Errorf("User namespaces enabled, but no root user mapping found.")
}
return id, nil
}
// Return default root uid 0
return 0, nil
}
// Utility function that gets a host ID for a container ID from user namespace map
// if that ID is present in the map.
func (c Config) hostIDFromMapping(containerID int, uMap []IDMap) (int, bool) {
for _, m := range uMap {
if (containerID >= m.ContainerID) && (containerID <= (m.ContainerID + m.Size - 1)) {
hostID := m.HostID + (containerID - m.ContainerID)
return hostID, true
}
}
return -1, false
}
package libcontainer
package configs
import (
"encoding/json"
"fmt"
"os"
"path/filepath"
"testing"
"github.com/docker/libcontainer/devices"
)
// Checks whether the expected capability is specified in the capabilities.
......@@ -19,13 +18,13 @@ func contains(expected string, values []string) bool {
return false
}
func containsDevice(expected *devices.Device, values []*devices.Device) bool {
func containsDevice(expected *Device, values []*Device) bool {
for _, d := range values {
if d.Path == expected.Path &&
d.CgroupPermissions == expected.CgroupPermissions &&
d.Permissions == expected.Permissions &&
d.FileMode == expected.FileMode &&
d.MajorNumber == expected.MajorNumber &&
d.MinorNumber == expected.MinorNumber &&
d.Major == expected.Major &&
d.Minor == expected.Minor &&
d.Type == expected.Type {
return true
}
......@@ -34,7 +33,7 @@ func containsDevice(expected *devices.Device, values []*devices.Device) bool {
}
func loadConfig(name string) (*Config, error) {
f, err := os.Open(filepath.Join("sample_configs", name))
f, err := os.Open(filepath.Join("../sample_configs", name))
if err != nil {
return nil, err
}
......@@ -45,6 +44,34 @@ func loadConfig(name string) (*Config, error) {
return nil, err
}
// Check that a config doesn't contain extra fields
var configMap, abstractMap map[string]interface{}
if _, err := f.Seek(0, 0); err != nil {
return nil, err
}
if err := json.NewDecoder(f).Decode(&abstractMap); err != nil {
return nil, err
}
configData, err := json.Marshal(&container)
if err != nil {
return nil, err
}
if err := json.Unmarshal(configData, &configMap); err != nil {
return nil, err
}
for k := range configMap {
delete(abstractMap, k)
}
if len(abstractMap) != 0 {
return nil, fmt.Errorf("unknown fields: %s", abstractMap)
}
return container, nil
}
......@@ -59,11 +86,6 @@ func TestConfigJsonFormat(t *testing.T) {
t.Fail()
}
if !container.Tty {
t.Log("tty should be set to true")
t.Fail()
}
if !container.Namespaces.Contains(NEWNET) {
t.Log("namespaces should contain NEWNET")
t.Fail()
......@@ -101,11 +123,6 @@ func TestConfigJsonFormat(t *testing.T) {
t.Fail()
}
if n.VethPrefix != "veth" {
t.Logf("veth prefix should be veth but received %q", n.VethPrefix)
t.Fail()
}
if n.Gateway != "172.17.42.1" {
t.Logf("veth gateway should be 172.17.42.1 but received %q", n.Gateway)
t.Fail()
......@@ -119,18 +136,12 @@ func TestConfigJsonFormat(t *testing.T) {
break
}
}
for _, d := range devices.DefaultSimpleDevices {
if !containsDevice(d, container.MountConfig.DeviceNodes) {
for _, d := range DefaultSimpleDevices {
if !containsDevice(d, container.Devices) {
t.Logf("expected device configuration for %s", d.Path)
t.Fail()
}
}
if !container.RestrictSys {
t.Log("expected restrict sys to be true")
t.Fail()
}
}
func TestApparmorProfile(t *testing.T) {
......@@ -154,8 +165,8 @@ func TestSelinuxLabels(t *testing.T) {
if container.ProcessLabel != label {
t.Fatalf("expected process label %q but received %q", label, container.ProcessLabel)
}
if container.MountConfig.MountLabel != label {
t.Fatalf("expected mount label %q but received %q", label, container.MountConfig.MountLabel)
if container.MountLabel != label {
t.Fatalf("expected mount label %q but received %q", label, container.MountLabel)
}
}
......@@ -170,3 +181,69 @@ func TestRemoveNamespace(t *testing.T) {
t.Fatalf("namespaces should have 0 items but reports %d", len(ns))
}
}
func TestHostUIDNoUSERNS(t *testing.T) {
config := &Config{
Namespaces: Namespaces{},
}
uid, err := config.HostUID()
if err != nil {
t.Fatal(err)
}
if uid != 0 {
t.Fatalf("expected uid 0 with no USERNS but received %d", uid)
}
}
func TestHostUIDWithUSERNS(t *testing.T) {
config := &Config{
Namespaces: Namespaces{{Type: NEWUSER}},
UidMappings: []IDMap{
{
ContainerID: 0,
HostID: 1000,
Size: 1,
},
},
}
uid, err := config.HostUID()
if err != nil {
t.Fatal(err)
}
if uid != 1000 {
t.Fatalf("expected uid 1000 with no USERNS but received %d", uid)
}
}
func TestHostGIDNoUSERNS(t *testing.T) {
config := &Config{
Namespaces: Namespaces{},
}
uid, err := config.HostGID()
if err != nil {
t.Fatal(err)
}
if uid != 0 {
t.Fatalf("expected gid 0 with no USERNS but received %d", uid)
}
}
func TestHostGIDWithUSERNS(t *testing.T) {
config := &Config{
Namespaces: Namespaces{{Type: NEWUSER}},
GidMappings: []IDMap{
{
ContainerID: 0,
HostID: 1000,
Size: 1,
},
},
}
uid, err := config.HostGID()
if err != nil {
t.Fatal(err)
}
if uid != 1000 {
t.Fatalf("expected gid 1000 with no USERNS but received %d", uid)
}
}
package configs
import (
"fmt"
"os"
)
const (
Wildcard = -1
)
type Device struct {
// Device type, block, char, etc.
Type rune `json:"type"`
// Path to the device.
Path string `json:"path"`
// Major is the device's major number.
Major int64 `json:"major"`
// Minor is the device's minor number.
Minor int64 `json:"minor"`
// Cgroup permissions format, rwm.
Permissions string `json:"permissions"`
// FileMode permission bits for the device.
FileMode os.FileMode `json:"file_mode"`
// Uid of the device.
Uid uint32 `json:"uid"`
// Gid of the device.
Gid uint32 `json:"gid"`
}
func (d *Device) CgroupString() string {
return fmt.Sprintf("%c %s:%s %s", d.Type, deviceNumberString(d.Major), deviceNumberString(d.Minor), d.Permissions)
}
func (d *Device) Mkdev() int {
return int((d.Major << 8) | (d.Minor & 0xff) | ((d.Minor & 0xfff00) << 12))
}
// deviceNumberString converts the device number to a string return result.
func deviceNumberString(number int64) string {
if number == Wildcard {
return "*"
}
return fmt.Sprint(number)
}
package configs
var (
// These are devices that are to be both allowed and created.
DefaultSimpleDevices = []*Device{
// /dev/null and zero
{
Path: "/dev/null",
Type: 'c',
Major: 1,
Minor: 3,
Permissions: "rwm",
FileMode: 0666,
},
{
Path: "/dev/zero",
Type: 'c',
Major: 1,
Minor: 5,
Permissions: "rwm",
FileMode: 0666,
},
{
Path: "/dev/full",
Type: 'c',
Major: 1,
Minor: 7,
Permissions: "rwm",
FileMode: 0666,
},
// consoles and ttys
{
Path: "/dev/tty",
Type: 'c',
Major: 5,
Minor: 0,
Permissions: "rwm",
FileMode: 0666,
},
// /dev/urandom,/dev/random
{
Path: "/dev/urandom",
Type: 'c',
Major: 1,
Minor: 9,
Permissions: "rwm",
FileMode: 0666,
},
{
Path: "/dev/random",
Type: 'c',
Major: 1,
Minor: 8,
Permissions: "rwm",
FileMode: 0666,
},
}
DefaultAllowedDevices = append([]*Device{
// allow mknod for any device
{
Type: 'c',
Major: Wildcard,
Minor: Wildcard,
Permissions: "m",
},
{
Type: 'b',
Major: Wildcard,
Minor: Wildcard,
Permissions: "m",
},
{
Path: "/dev/console",
Type: 'c',
Major: 5,
Minor: 1,
Permissions: "rwm",
},
{
Path: "/dev/tty0",
Type: 'c',
Major: 4,
Minor: 0,
Permissions: "rwm",
},
{
Path: "/dev/tty1",
Type: 'c',
Major: 4,
Minor: 1,
Permissions: "rwm",
},
// /dev/pts/ - pts namespaces are "coming soon"
{
Path: "",
Type: 'c',
Major: 136,
Minor: Wildcard,
Permissions: "rwm",
},
{
Path: "",
Type: 'c',
Major: 5,
Minor: 2,
Permissions: "rwm",
},
// tuntap
{
Path: "",
Type: 'c',
Major: 10,
Minor: 200,
Permissions: "rwm",
},
}, DefaultSimpleDevices...)
DefaultAutoCreatedDevices = append([]*Device{
{
// /dev/fuse is created but not allowed.
// This is to allow java to work. Because java
// Insists on there being a /dev/fuse
// https://github.com/docker/docker/issues/514
// https://github.com/docker/docker/issues/2393
//
Path: "/dev/fuse",
Type: 'c',
Major: 10,
Minor: 229,
Permissions: "rwm",
},
}, DefaultSimpleDevices...)
)
package configs
type Mount struct {
// Source path for the mount.
Source string `json:"source"`
// Destination path for the mount inside the container.
Destination string `json:"destination"`
// Device the mount is for.
Device string `json:"device"`
// Mount flags.
Flags int `json:"flags"`
// Mount data applied to the mount.
Data string `json:"data"`
// Relabel source if set, "z" indicates shared, "Z" indicates unshared.
Relabel string `json:"relabel"`
}
package configs
import (
"fmt"
"syscall"
)
type NamespaceType string
const (
NEWNET NamespaceType = "NEWNET"
NEWPID NamespaceType = "NEWPID"
NEWNS NamespaceType = "NEWNS"
NEWUTS NamespaceType = "NEWUTS"
NEWIPC NamespaceType = "NEWIPC"
NEWUSER NamespaceType = "NEWUSER"
)
func NamespaceTypes() []NamespaceType {
return []NamespaceType{
NEWNET,
NEWPID,
NEWNS,
NEWUTS,
NEWIPC,
NEWUSER,
}
}
// Namespace defines configuration for each namespace. It specifies an
// alternate path that is able to be joined via setns.
type Namespace struct {
Type NamespaceType `json:"type"`
Path string `json:"path"`
}
func (n *Namespace) Syscall() int {
return namespaceInfo[n.Type]
}
func (n *Namespace) GetPath(pid int) string {
if n.Path != "" {
return n.Path
}
return fmt.Sprintf("/proc/%d/ns/%s", pid, n.file())
}
func (n *Namespace) file() string {
file := ""
switch n.Type {
case NEWNET:
file = "net"
case NEWNS:
file = "mnt"
case NEWPID:
file = "pid"
case NEWIPC:
file = "ipc"
case NEWUSER:
file = "user"
case NEWUTS:
file = "uts"
}
return file
}
type Namespaces []Namespace
func (n *Namespaces) Remove(t NamespaceType) bool {
i := n.index(t)
if i == -1 {
return false
}
*n = append((*n)[:i], (*n)[i+1:]...)
return true
}
func (n *Namespaces) Add(t NamespaceType, path string) {
i := n.index(t)
if i == -1 {
*n = append(*n, Namespace{Type: t, Path: path})
return
}
(*n)[i].Path = path
}
func (n *Namespaces) index(t NamespaceType) int {
for i, ns := range *n {
if ns.Type == t {
return i
}
}
return -1
}
func (n *Namespaces) Contains(t NamespaceType) bool {
return n.index(t) != -1
}
var namespaceInfo = map[NamespaceType]int{
NEWNET: syscall.CLONE_NEWNET,
NEWNS: syscall.CLONE_NEWNS,
NEWUSER: syscall.CLONE_NEWUSER,
NEWIPC: syscall.CLONE_NEWIPC,
NEWUTS: syscall.CLONE_NEWUTS,
NEWPID: syscall.CLONE_NEWPID,
}
// CloneFlags parses the container's Namespaces options to set the correct
// flags on clone, unshare. This functions returns flags only for new namespaces.
func (n *Namespaces) CloneFlags() uintptr {
var flag int
for _, v := range *n {
if v.Path != "" {
continue
}
flag |= namespaceInfo[v.Type]
}
return uintptr(flag)
}
package configs
// Network defines configuration for a container's networking stack
//
// The network configuration can be omited from a container causing the
// container to be setup with the host's networking stack
type Network struct {
// Type sets the networks type, commonly veth and loopback
Type string `json:"type"`
// Name of the network interface
Name string `json:"name"`
// The bridge to use.
Bridge string `json:"bridge"`
// MacAddress contains the MAC address to set on the network interface
MacAddress string `json:"mac_address"`
// Address contains the IPv4 and mask to set on the network interface
Address string `json:"address"`
// Gateway sets the gateway address that is used as the default for the interface
Gateway string `json:"gateway"`
// IPv6Address contains the IPv6 and mask to set on the network interface
IPv6Address string `json:"ipv6_address"`
// IPv6Gateway sets the ipv6 gateway address that is used as the default for the interface
IPv6Gateway string `json:"ipv6_gateway"`
// Mtu sets the mtu value for the interface and will be mirrored on both the host and
// container's interfaces if a pair is created, specifically in the case of type veth
// Note: This does not apply to loopback interfaces.
Mtu int `json:"mtu"`
// TxQueueLen sets the tx_queuelen value for the interface and will be mirrored on both the host and
// container's interfaces if a pair is created, specifically in the case of type veth
// Note: This does not apply to loopback interfaces.
TxQueueLen int `json:"txqueuelen"`
// HostInterfaceName is a unique name of a veth pair that resides on in the host interface of the
// container.
HostInterfaceName string `json:"host_interface_name"`
// HairpinMode specifies if hairpin NAT should be enabled on the virtual interface
// bridge port in the case of type veth
// Note: This is unsupported on some systems.
// Note: This does not apply to loopback interfaces.
HairpinMode bool `json:"hairpin_mode"`
}
// Routes can be specified to create entries in the route table as the container is started
//
// All of destination, source, and gateway should be either IPv4 or IPv6.
// One of the three options must be present, and ommitted entries will use their
// IP family default for the route table. For IPv4 for example, setting the
// gateway to 1.2.3.4 and the interface to eth0 will set up a standard
// destination of 0.0.0.0(or *) when viewed in the route table.
type Route struct {
// Sets the destination and mask, should be a CIDR. Accepts IPv4 and IPv6
Destination string `json:"destination"`
// Sets the source and mask, should be a CIDR. Accepts IPv4 and IPv6
Source string `json:"source"`
// Sets the gateway. Accepts IPv4 and IPv6
Gateway string `json:"gateway"`
// The device to set this route up for, for example: eth0
InterfaceName string `json:"interface_name"`
}
package validate
import (
"fmt"
"os"
"path/filepath"
"github.com/docker/libcontainer/configs"
)
type Validator interface {
Validate(*configs.Config) error
}
func New() Validator {
return &ConfigValidator{}
}
type ConfigValidator struct {
}
func (v *ConfigValidator) Validate(config *configs.Config) error {
if err := v.rootfs(config); err != nil {
return err
}
if err := v.network(config); err != nil {
return err
}
if err := v.hostname(config); err != nil {
return err
}
if err := v.security(config); err != nil {
return err
}
if err := v.usernamespace(config); err != nil {
return err
}
return nil
}
// rootfs validates the the rootfs is an absolute path and is not a symlink
// to the container's root filesystem.
func (v *ConfigValidator) rootfs(config *configs.Config) error {
cleaned, err := filepath.Abs(config.Rootfs)
if err != nil {
return err
}
if cleaned, err = filepath.EvalSymlinks(cleaned); err != nil {
return err
}
if config.Rootfs != cleaned {
return fmt.Errorf("%s is not an absolute path or is a symlink", config.Rootfs)
}
return nil
}
func (v *ConfigValidator) network(config *configs.Config) error {
if !config.Namespaces.Contains(configs.NEWNET) {
if len(config.Networks) > 0 || len(config.Routes) > 0 {
return fmt.Errorf("unable to apply network settings without a private NET namespace")
}
}
return nil
}
func (v *ConfigValidator) hostname(config *configs.Config) error {
if config.Hostname != "" && !config.Namespaces.Contains(configs.NEWUTS) {
return fmt.Errorf("unable to set hostname without a private UTS namespace")
}
return nil
}
func (v *ConfigValidator) security(config *configs.Config) error {
// restrict sys without mount namespace
if (len(config.MaskPaths) > 0 || len(config.ReadonlyPaths) > 0) &&
!config.Namespaces.Contains(configs.NEWNS) {
return fmt.Errorf("unable to restrict sys entries without a private MNT namespace")
}
return nil
}
func (v *ConfigValidator) usernamespace(config *configs.Config) error {
if config.Namespaces.Contains(configs.NEWUSER) {
if _, err := os.Stat("/proc/self/ns/user"); os.IsNotExist(err) {
return fmt.Errorf("USER namespaces aren't enabled in the kernel")
}
} else {
if config.UidMappings != nil || config.GidMappings != nil {
return fmt.Errorf("User namespace mappings specified, but USER namespace isn't enabled in the config")
}
}
return nil
}
package libcontainer
import "io"
// Console represents a pseudo TTY.
type Console interface {
io.ReadWriter
io.Closer
// Path returns the filesystem path to the slave side of the pty.
Path() string
// Fd returns the fd for the master of the pty.
Fd() uintptr
}
// +build linux
package console
import (
"fmt"
"os"
"path/filepath"
"syscall"
"unsafe"
"github.com/docker/libcontainer/label"
)
// Setup initializes the proper /dev/console inside the rootfs path
func Setup(rootfs, consolePath, mountLabel string) error {
oldMask := syscall.Umask(0000)
defer syscall.Umask(oldMask)
if err := os.Chmod(consolePath, 0600); err != nil {
return err
}
if err := os.Chown(consolePath, 0, 0); err != nil {
return err
}
if err := label.SetFileLabel(consolePath, mountLabel); err != nil {
return fmt.Errorf("set file label %s %s", consolePath, err)
}
dest := filepath.Join(rootfs, "dev/console")
f, err := os.Create(dest)
if err != nil && !os.IsExist(err) {
return fmt.Errorf("create %s %s", dest, err)
}
if f != nil {
f.Close()
}
if err := syscall.Mount(consolePath, dest, "bind", syscall.MS_BIND, ""); err != nil {
return fmt.Errorf("bind %s to %s %s", consolePath, dest, err)
}
return nil
}
func OpenAndDup(consolePath string) error {
slave, err := OpenTerminal(consolePath, syscall.O_RDWR)
if err != nil {
return fmt.Errorf("open terminal %s", err)
}
if err := syscall.Dup2(int(slave.Fd()), 0); err != nil {
return err
}
if err := syscall.Dup2(int(slave.Fd()), 1); err != nil {
return err
}
return syscall.Dup2(int(slave.Fd()), 2)
}
// Unlockpt unlocks the slave pseudoterminal device corresponding to the master pseudoterminal referred to by f.
// Unlockpt should be called before opening the slave side of a pseudoterminal.
func Unlockpt(f *os.File) error {
var u int32
return Ioctl(f.Fd(), syscall.TIOCSPTLCK, uintptr(unsafe.Pointer(&u)))
}
// Ptsname retrieves the name of the first available pts for the given master.
func Ptsname(f *os.File) (string, error) {
var n int32
if err := Ioctl(f.Fd(), syscall.TIOCGPTN, uintptr(unsafe.Pointer(&n))); err != nil {
return "", err
}
return fmt.Sprintf("/dev/pts/%d", n), nil
}
// CreateMasterAndConsole will open /dev/ptmx on the host and retreive the
// pts name for use as the pty slave inside the container
func CreateMasterAndConsole() (*os.File, string, error) {
master, err := os.OpenFile("/dev/ptmx", syscall.O_RDWR|syscall.O_NOCTTY|syscall.O_CLOEXEC, 0)
if err != nil {
return nil, "", err
}
console, err := Ptsname(master)
if err != nil {
return nil, "", err
}
if err := Unlockpt(master); err != nil {
return nil, "", err
}
return master, console, nil
}
// OpenPtmx opens /dev/ptmx, i.e. the PTY master.
func OpenPtmx() (*os.File, error) {
// O_NOCTTY and O_CLOEXEC are not present in os package so we use the syscall's one for all.
return os.OpenFile("/dev/ptmx", syscall.O_RDONLY|syscall.O_NOCTTY|syscall.O_CLOEXEC, 0)
}
// OpenTerminal is a clone of os.OpenFile without the O_CLOEXEC
// used to open the pty slave inside the container namespace
func OpenTerminal(name string, flag int) (*os.File, error) {
r, e := syscall.Open(name, flag, 0)
if e != nil {
return nil, &os.PathError{Op: "open", Path: name, Err: e}
}
return os.NewFile(uintptr(r), name), nil
}
func Ioctl(fd uintptr, flag, data uintptr) error {
if _, _, err := syscall.Syscall(syscall.SYS_IOCTL, fd, flag, data); err != 0 {
return err
}
return nil
}
// +build linux
package libcontainer
import (
"fmt"
"os"
"path/filepath"
"syscall"
"unsafe"
"github.com/docker/libcontainer/label"
)
// newConsole returns an initalized console that can be used within a container by copying bytes
// from the master side to the slave that is attached as the tty for the container's init process.
func newConsole(uid, gid int) (Console, error) {
master, err := os.OpenFile("/dev/ptmx", syscall.O_RDWR|syscall.O_NOCTTY|syscall.O_CLOEXEC, 0)
if err != nil {
return nil, err
}
console, err := ptsname(master)
if err != nil {
return nil, err
}
if err := unlockpt(master); err != nil {
return nil, err
}
if err := os.Chmod(console, 0600); err != nil {
return nil, err
}
if err := os.Chown(console, uid, gid); err != nil {
return nil, err
}
return &linuxConsole{
slavePath: console,
master: master,
}, nil
}
// newConsoleFromPath is an internal fucntion returning an initialzied console for use inside
// a container's MNT namespace.
func newConsoleFromPath(slavePath string) *linuxConsole {
return &linuxConsole{
slavePath: slavePath,
}
}
// linuxConsole is a linux psuedo TTY for use within a container.
type linuxConsole struct {
master *os.File
slavePath string
}
func (c *linuxConsole) Fd() uintptr {
return c.master.Fd()
}
func (c *linuxConsole) Path() string {
return c.slavePath
}
func (c *linuxConsole) Read(b []byte) (int, error) {
return c.master.Read(b)
}
func (c *linuxConsole) Write(b []byte) (int, error) {
return c.master.Write(b)
}
func (c *linuxConsole) Close() error {
if m := c.master; m != nil {
return m.Close()
}
return nil
}
// mount initializes the console inside the rootfs mounting with the specified mount label
// and applying the correct ownership of the console.
func (c *linuxConsole) mount(rootfs, mountLabel string, uid, gid int) error {
oldMask := syscall.Umask(0000)
defer syscall.Umask(oldMask)
if err := label.SetFileLabel(c.slavePath, mountLabel); err != nil {
return err
}
dest := filepath.Join(rootfs, "/dev/console")
f, err := os.Create(dest)
if err != nil && !os.IsExist(err) {
return err
}
if f != nil {
f.Close()
}
return syscall.Mount(c.slavePath, dest, "bind", syscall.MS_BIND, "")
}
// dupStdio opens the slavePath for the console and dup2s the fds to the current
// processes stdio, fd 0,1,2.
func (c *linuxConsole) dupStdio() error {
slave, err := c.open(syscall.O_RDWR)
if err != nil {
return err
}
fd := int(slave.Fd())
for _, i := range []int{0, 1, 2} {
if err := syscall.Dup2(fd, i); err != nil {
return err
}
}
return nil
}
// open is a clone of os.OpenFile without the O_CLOEXEC used to open the pty slave.
func (c *linuxConsole) open(flag int) (*os.File, error) {
r, e := syscall.Open(c.slavePath, flag, 0)
if e != nil {
return nil, &os.PathError{
Op: "open",
Path: c.slavePath,
Err: e,
}
}
return os.NewFile(uintptr(r), c.slavePath), nil
}
func ioctl(fd uintptr, flag, data uintptr) error {
if _, _, err := syscall.Syscall(syscall.SYS_IOCTL, fd, flag, data); err != 0 {
return err
}
return nil
}
// unlockpt unlocks the slave pseudoterminal device corresponding to the master pseudoterminal referred to by f.
// unlockpt should be called before opening the slave side of a pty.
func unlockpt(f *os.File) error {
var u int32
return ioctl(f.Fd(), syscall.TIOCSPTLCK, uintptr(unsafe.Pointer(&u)))
}
// ptsname retrieves the name of the first available pts for the given master.
func ptsname(f *os.File) (string, error) {
var n int32
if err := ioctl(f.Fd(), syscall.TIOCGPTN, uintptr(unsafe.Pointer(&n))); err != nil {
return "", err
}
return fmt.Sprintf("/dev/pts/%d", n), nil
}
/*
NOTE: The API is in flux and mainly not implemented. Proceed with caution until further notice.
*/
// Libcontainer provides a native Go implementation for creating containers
// with namespaces, cgroups, capabilities, and filesystem access controls.
// It allows you to manage the lifecycle of the container performing additional operations
// after the container is created.
package libcontainer
import (
"github.com/docker/libcontainer/configs"
)
// The status of a container.
type Status int
const (
// The container exists and is running.
Running Status = iota + 1
// The container exists, it is in the process of being paused.
Pausing
// The container exists, but all its processes are paused.
Paused
// The container does not exist.
Destroyed
)
// State represents a running container's state
type State struct {
// ID is the container ID.
ID string `json:"id"`
// InitProcessPid is the init process id in the parent namespace.
InitProcessPid int `json:"init_process_pid"`
// InitProcessStartTime is the init process start time.
InitProcessStartTime string `json:"init_process_start"`
// Path to all the cgroups setup for a container. Key is cgroup subsystem name
// with the value as the path.
CgroupPaths map[string]string `json:"cgroup_paths"`
// NamespacePaths are filepaths to the container's namespaces. Key is the namespace type
// with the value as the path.
NamespacePaths map[configs.NamespaceType]string `json:"namespace_paths"`
// Config is the container's configuration.
Config configs.Config `json:"config"`
}
// A libcontainer container object.
//
// Each container is thread-safe within the same process. Since a container can
......@@ -12,67 +57,88 @@ type Container interface {
// Returns the ID of the container
ID() string
// Returns the current run state of the container.
// Returns the current status of the container.
//
// Errors:
// errors:
// ContainerDestroyed - Container no longer exists,
// SystemError - System error.
RunState() (*RunState, Error)
// Returns the current config of the container.
Config() *Config
// Systemerror - System error.
Status() (Status, error)
// Start a process inside the container. Returns the PID of the new process (in the caller process's namespace) and a channel that will return the exit status of the process whenever it dies.
// State returns the current container's state information.
//
// Errors:
// ContainerDestroyed - Container no longer exists,
// ConfigInvalid - config is invalid,
// ContainerPaused - Container is paused,
// SystemError - System error.
Start(config *ProcessConfig) (pid int, exitChan chan int, err Error)
// errors:
// Systemerror - System erroor.
State() (*State, error)
// Destroys the container after killing all running processes.
//
// Any event registrations are removed before the container is destroyed.
// No error is returned if the container is already destroyed.
//
// Errors:
// SystemError - System error.
Destroy() Error
// Returns the current config of the container.
Config() configs.Config
// Returns the PIDs inside this container. The PIDs are in the namespace of the calling process.
//
// Errors:
// errors:
// ContainerDestroyed - Container no longer exists,
// SystemError - System error.
// Systemerror - System error.
//
// Some of the returned PIDs may no longer refer to processes in the Container, unless
// the Container state is PAUSED in which case every PID in the slice is valid.
Processes() ([]int, Error)
Processes() ([]int, error)
// Returns statistics for the container.
//
// Errors:
// errors:
// ContainerDestroyed - Container no longer exists,
// Systemerror - System error.
Stats() (*Stats, error)
// Set cgroup resources of container as configured
//
// We can use this to change resources when containers are running.
//
// errors:
// Systemerror - System error.
Set(config configs.Config) error
// Start a process inside the container. Returns error if process fails to
// start. You can track process lifecycle with passed Process structure.
//
// errors:
// ContainerDestroyed - Container no longer exists,
// SystemError - System error.
Stats() (*ContainerStats, Error)
// ConfigInvalid - config is invalid,
// ContainerPaused - Container is paused,
// Systemerror - System error.
Start(process *Process) (err error)
// Destroys the container after killing all running processes.
//
// Any event registrations are removed before the container is destroyed.
// No error is returned if the container is already destroyed.
//
// errors:
// Systemerror - System error.
Destroy() error
// If the Container state is RUNNING or PAUSING, sets the Container state to PAUSING and pauses
// the execution of any user processes. Asynchronously, when the container finished being paused the
// state is changed to PAUSED.
// If the Container state is PAUSED, do nothing.
//
// Errors:
// errors:
// ContainerDestroyed - Container no longer exists,
// SystemError - System error.
Pause() Error
// Systemerror - System error.
Pause() error
// If the Container state is PAUSED, resumes the execution of any user processes in the
// Container before setting the Container state to RUNNING.
// If the Container state is RUNNING, do nothing.
//
// Errors:
// errors:
// ContainerDestroyed - Container no longer exists,
// SystemError - System error.
Resume() Error
// Systemerror - System error.
Resume() error
// NotifyOOM returns a read-only channel signaling when the container receives an OOM notification.
//
// errors:
// Systemerror - System error.
NotifyOOM() (<-chan struct{}, error)
}
// +build linux
package libcontainer
import (
"encoding/json"
"fmt"
"os"
"os/exec"
"path/filepath"
"sync"
"syscall"
log "github.com/Sirupsen/logrus"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
const stdioFdCount = 3
type linuxContainer struct {
id string
root string
config *configs.Config
cgroupManager cgroups.Manager
initPath string
initArgs []string
initProcess parentProcess
m sync.Mutex
}
// ID returns the container's unique ID
func (c *linuxContainer) ID() string {
return c.id
}
// Config returns the container's configuration
func (c *linuxContainer) Config() configs.Config {
return *c.config
}
func (c *linuxContainer) Status() (Status, error) {
c.m.Lock()
defer c.m.Unlock()
return c.currentStatus()
}
func (c *linuxContainer) State() (*State, error) {
c.m.Lock()
defer c.m.Unlock()
return c.currentState()
}
func (c *linuxContainer) Processes() ([]int, error) {
pids, err := c.cgroupManager.GetPids()
if err != nil {
return nil, newSystemError(err)
}
return pids, nil
}
func (c *linuxContainer) Stats() (*Stats, error) {
var (
err error
stats = &Stats{}
)
if stats.CgroupStats, err = c.cgroupManager.GetStats(); err != nil {
return stats, newSystemError(err)
}
for _, iface := range c.config.Networks {
switch iface.Type {
case "veth":
istats, err := getNetworkInterfaceStats(iface.HostInterfaceName)
if err != nil {
return stats, newSystemError(err)
}
stats.Interfaces = append(stats.Interfaces, istats)
}
}
return stats, nil
}
func (c *linuxContainer) Set(config configs.Config) error {
c.m.Lock()
defer c.m.Unlock()
c.config = &config
return c.cgroupManager.Set(c.config)
}
func (c *linuxContainer) Start(process *Process) error {
c.m.Lock()
defer c.m.Unlock()
status, err := c.currentStatus()
if err != nil {
return err
}
doInit := status == Destroyed
parent, err := c.newParentProcess(process, doInit)
if err != nil {
return newSystemError(err)
}
if err := parent.start(); err != nil {
// terminate the process to ensure that it properly is reaped.
if err := parent.terminate(); err != nil {
log.Warn(err)
}
return newSystemError(err)
}
process.ops = parent
if doInit {
c.updateState(parent)
}
return nil
}
func (c *linuxContainer) newParentProcess(p *Process, doInit bool) (parentProcess, error) {
parentPipe, childPipe, err := newPipe()
if err != nil {
return nil, newSystemError(err)
}
cmd, err := c.commandTemplate(p, childPipe)
if err != nil {
return nil, newSystemError(err)
}
if !doInit {
return c.newSetnsProcess(p, cmd, parentPipe, childPipe), nil
}
return c.newInitProcess(p, cmd, parentPipe, childPipe)
}
func (c *linuxContainer) commandTemplate(p *Process, childPipe *os.File) (*exec.Cmd, error) {
cmd := &exec.Cmd{
Path: c.initPath,
Args: c.initArgs,
}
cmd.Stdin = p.Stdin
cmd.Stdout = p.Stdout
cmd.Stderr = p.Stderr
cmd.Dir = c.config.Rootfs
if cmd.SysProcAttr == nil {
cmd.SysProcAttr = &syscall.SysProcAttr{}
}
cmd.ExtraFiles = append(p.ExtraFiles, childPipe)
cmd.Env = append(cmd.Env, fmt.Sprintf("_LIBCONTAINER_INITPIPE=%d", stdioFdCount+len(cmd.ExtraFiles)-1))
// NOTE: when running a container with no PID namespace and the parent process spawning the container is
// PID1 the pdeathsig is being delivered to the container's init process by the kernel for some reason
// even with the parent still running.
if c.config.ParentDeathSignal > 0 {
cmd.SysProcAttr.Pdeathsig = syscall.Signal(c.config.ParentDeathSignal)
}
return cmd, nil
}
func (c *linuxContainer) newInitProcess(p *Process, cmd *exec.Cmd, parentPipe, childPipe *os.File) (*initProcess, error) {
t := "_LIBCONTAINER_INITTYPE=standard"
cloneFlags := c.config.Namespaces.CloneFlags()
if cloneFlags&syscall.CLONE_NEWUSER != 0 {
if err := c.addUidGidMappings(cmd.SysProcAttr); err != nil {
// user mappings are not supported
return nil, err
}
// Default to root user when user namespaces are enabled.
if cmd.SysProcAttr.Credential == nil {
cmd.SysProcAttr.Credential = &syscall.Credential{}
}
}
cmd.Env = append(cmd.Env, t)
cmd.SysProcAttr.Cloneflags = cloneFlags
return &initProcess{
cmd: cmd,
childPipe: childPipe,
parentPipe: parentPipe,
manager: c.cgroupManager,
config: c.newInitConfig(p),
}, nil
}
func (c *linuxContainer) newSetnsProcess(p *Process, cmd *exec.Cmd, parentPipe, childPipe *os.File) *setnsProcess {
cmd.Env = append(cmd.Env,
fmt.Sprintf("_LIBCONTAINER_INITPID=%d", c.initProcess.pid()),
"_LIBCONTAINER_INITTYPE=setns",
)
if p.consolePath != "" {
cmd.Env = append(cmd.Env, "_LIBCONTAINER_CONSOLE_PATH="+p.consolePath)
}
// TODO: set on container for process management
return &setnsProcess{
cmd: cmd,
cgroupPaths: c.cgroupManager.GetPaths(),
childPipe: childPipe,
parentPipe: parentPipe,
config: c.newInitConfig(p),
}
}
func (c *linuxContainer) newInitConfig(process *Process) *initConfig {
return &initConfig{
Config: c.config,
Args: process.Args,
Env: process.Env,
User: process.User,
Cwd: process.Cwd,
Console: process.consolePath,
Capabilities: process.Capabilities,
PassedFilesCount: len(process.ExtraFiles),
}
}
func newPipe() (parent *os.File, child *os.File, err error) {
fds, err := syscall.Socketpair(syscall.AF_LOCAL, syscall.SOCK_STREAM|syscall.SOCK_CLOEXEC, 0)
if err != nil {
return nil, nil, err
}
return os.NewFile(uintptr(fds[1]), "parent"), os.NewFile(uintptr(fds[0]), "child"), nil
}
func (c *linuxContainer) Destroy() error {
c.m.Lock()
defer c.m.Unlock()
status, err := c.currentStatus()
if err != nil {
return err
}
if status != Destroyed {
return newGenericError(fmt.Errorf("container is not destroyed"), ContainerNotStopped)
}
if !c.config.Namespaces.Contains(configs.NEWPID) {
if err := killCgroupProcesses(c.cgroupManager); err != nil {
log.Warn(err)
}
}
err = c.cgroupManager.Destroy()
if rerr := os.RemoveAll(c.root); err == nil {
err = rerr
}
c.initProcess = nil
return err
}
func (c *linuxContainer) Pause() error {
c.m.Lock()
defer c.m.Unlock()
return c.cgroupManager.Freeze(configs.Frozen)
}
func (c *linuxContainer) Resume() error {
c.m.Lock()
defer c.m.Unlock()
return c.cgroupManager.Freeze(configs.Thawed)
}
func (c *linuxContainer) NotifyOOM() (<-chan struct{}, error) {
return notifyOnOOM(c.cgroupManager.GetPaths())
}
func (c *linuxContainer) updateState(process parentProcess) error {
c.initProcess = process
state, err := c.currentState()
if err != nil {
return err
}
f, err := os.Create(filepath.Join(c.root, stateFilename))
if err != nil {
return err
}
defer f.Close()
return json.NewEncoder(f).Encode(state)
}
func (c *linuxContainer) currentStatus() (Status, error) {
if c.initProcess == nil {
return Destroyed, nil
}
// return Running if the init process is alive
if err := syscall.Kill(c.initProcess.pid(), 0); err != nil {
if err == syscall.ESRCH {
return Destroyed, nil
}
return 0, newSystemError(err)
}
if c.config.Cgroups != nil && c.config.Cgroups.Freezer == configs.Frozen {
return Paused, nil
}
return Running, nil
}
func (c *linuxContainer) currentState() (*State, error) {
status, err := c.currentStatus()
if err != nil {
return nil, err
}
if status == Destroyed {
return nil, newGenericError(fmt.Errorf("container destroyed"), ContainerNotExists)
}
startTime, err := c.initProcess.startTime()
if err != nil {
return nil, newSystemError(err)
}
state := &State{
ID: c.ID(),
Config: *c.config,
InitProcessPid: c.initProcess.pid(),
InitProcessStartTime: startTime,
CgroupPaths: c.cgroupManager.GetPaths(),
NamespacePaths: make(map[configs.NamespaceType]string),
}
for _, ns := range c.config.Namespaces {
state.NamespacePaths[ns.Type] = ns.GetPath(c.initProcess.pid())
}
for _, nsType := range configs.NamespaceTypes() {
if _, ok := state.NamespacePaths[nsType]; !ok {
ns := configs.Namespace{Type: nsType}
state.NamespacePaths[ns.Type] = ns.GetPath(c.initProcess.pid())
}
}
return state, nil
}
// +build linux
package libcontainer
import (
"fmt"
"os"
"testing"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/configs"
)
type mockCgroupManager struct {
pids []int
stats *cgroups.Stats
paths map[string]string
}
func (m *mockCgroupManager) GetPids() ([]int, error) {
return m.pids, nil
}
func (m *mockCgroupManager) GetStats() (*cgroups.Stats, error) {
return m.stats, nil
}
func (m *mockCgroupManager) Apply(pid int) error {
return nil
}
func (m *mockCgroupManager) Set(container *configs.Config) error {
return nil
}
func (m *mockCgroupManager) Destroy() error {
return nil
}
func (m *mockCgroupManager) GetPaths() map[string]string {
return m.paths
}
func (m *mockCgroupManager) Freeze(state configs.FreezerState) error {
return nil
}
type mockProcess struct {
_pid int
started string
}
func (m *mockProcess) terminate() error {
return nil
}
func (m *mockProcess) pid() int {
return m._pid
}
func (m *mockProcess) startTime() (string, error) {
return m.started, nil
}
func (m *mockProcess) start() error {
return nil
}
func (m *mockProcess) wait() (*os.ProcessState, error) {
return nil, nil
}
func (m *mockProcess) signal(_ os.Signal) error {
return nil
}
func TestGetContainerPids(t *testing.T) {
container := &linuxContainer{
id: "myid",
config: &configs.Config{},
cgroupManager: &mockCgroupManager{pids: []int{1, 2, 3}},
}
pids, err := container.Processes()
if err != nil {
t.Fatal(err)
}
for i, expected := range []int{1, 2, 3} {
if pids[i] != expected {
t.Fatalf("expected pid %d but received %d", expected, pids[i])
}
}
}
func TestGetContainerStats(t *testing.T) {
container := &linuxContainer{
id: "myid",
config: &configs.Config{},
cgroupManager: &mockCgroupManager{
pids: []int{1, 2, 3},
stats: &cgroups.Stats{
MemoryStats: cgroups.MemoryStats{
Usage: 1024,
},
},
},
}
stats, err := container.Stats()
if err != nil {
t.Fatal(err)
}
if stats.CgroupStats == nil {
t.Fatal("cgroup stats are nil")
}
if stats.CgroupStats.MemoryStats.Usage != 1024 {
t.Fatalf("expected memory usage 1024 but recevied %d", stats.CgroupStats.MemoryStats.Usage)
}
}
func TestGetContainerState(t *testing.T) {
var (
pid = os.Getpid()
expectedMemoryPath = "/sys/fs/cgroup/memory/myid"
expectedNetworkPath = "/networks/fd"
)
container := &linuxContainer{
id: "myid",
config: &configs.Config{
Namespaces: []configs.Namespace{
{Type: configs.NEWPID},
{Type: configs.NEWNS},
{Type: configs.NEWNET, Path: expectedNetworkPath},
{Type: configs.NEWUTS},
// emulate host for IPC
//{Type: configs.NEWIPC},
},
},
initProcess: &mockProcess{
_pid: pid,
started: "010",
},
cgroupManager: &mockCgroupManager{
pids: []int{1, 2, 3},
stats: &cgroups.Stats{
MemoryStats: cgroups.MemoryStats{
Usage: 1024,
},
},
paths: map[string]string{
"memory": expectedMemoryPath,
},
},
}
state, err := container.State()
if err != nil {
t.Fatal(err)
}
if state.InitProcessPid != pid {
t.Fatalf("expected pid %d but received %d", pid, state.InitProcessPid)
}
if state.InitProcessStartTime != "010" {
t.Fatalf("expected process start time 010 but received %s", state.InitProcessStartTime)
}
paths := state.CgroupPaths
if paths == nil {
t.Fatal("cgroup paths should not be nil")
}
if memPath := paths["memory"]; memPath != expectedMemoryPath {
t.Fatalf("expected memory path %q but received %q", expectedMemoryPath, memPath)
}
for _, ns := range container.config.Namespaces {
path := state.NamespacePaths[ns.Type]
if path == "" {
t.Fatalf("expected non nil namespace path for %s", ns.Type)
}
if ns.Type == configs.NEWNET {
if path != expectedNetworkPath {
t.Fatalf("expected path %q but received %q", expectedNetworkPath, path)
}
} else {
file := ""
switch ns.Type {
case configs.NEWNET:
file = "net"
case configs.NEWNS:
file = "mnt"
case configs.NEWPID:
file = "pid"
case configs.NEWIPC:
file = "ipc"
case configs.NEWUSER:
file = "user"
case configs.NEWUTS:
file = "uts"
}
expected := fmt.Sprintf("/proc/%d/ns/%s", pid, file)
if expected != path {
t.Fatalf("expected path %q but received %q", expected, path)
}
}
}
}
// +build !go1.4
package libcontainer
import (
"fmt"
"syscall"
)
// not available before go 1.4
func (c *linuxContainer) addUidGidMappings(sys *syscall.SysProcAttr) error {
return fmt.Errorf("User namespace is not supported in golang < 1.4")
}
// +build go1.4
package libcontainer
import "syscall"
// Converts IDMap to SysProcIDMap array and adds it to SysProcAttr.
func (c *linuxContainer) addUidGidMappings(sys *syscall.SysProcAttr) error {
if c.config.UidMappings != nil {
sys.UidMappings = make([]syscall.SysProcIDMap, len(c.config.UidMappings))
for i, um := range c.config.UidMappings {
sys.UidMappings[i].ContainerID = um.ContainerID
sys.UidMappings[i].HostID = um.HostID
sys.UidMappings[i].Size = um.Size
}
}
if c.config.GidMappings != nil {
sys.GidMappings = make([]syscall.SysProcIDMap, len(c.config.GidMappings))
for i, gm := range c.config.GidMappings {
sys.GidMappings[i].ContainerID = gm.ContainerID
sys.GidMappings[i].HostID = gm.HostID
sys.GidMappings[i].Size = gm.Size
}
}
return nil
}
package devices
var (
// These are devices that are to be both allowed and created.
DefaultSimpleDevices = []*Device{
// /dev/null and zero
{
Path: "/dev/null",
Type: 'c',
MajorNumber: 1,
MinorNumber: 3,
CgroupPermissions: "rwm",
FileMode: 0666,
},
{
Path: "/dev/zero",
Type: 'c',
MajorNumber: 1,
MinorNumber: 5,
CgroupPermissions: "rwm",
FileMode: 0666,
},
{
Path: "/dev/full",
Type: 'c',
MajorNumber: 1,
MinorNumber: 7,
CgroupPermissions: "rwm",
FileMode: 0666,
},
// consoles and ttys
{
Path: "/dev/tty",
Type: 'c',
MajorNumber: 5,
MinorNumber: 0,
CgroupPermissions: "rwm",
FileMode: 0666,
},
// /dev/urandom,/dev/random
{
Path: "/dev/urandom",
Type: 'c',
MajorNumber: 1,
MinorNumber: 9,
CgroupPermissions: "rwm",
FileMode: 0666,
},
{
Path: "/dev/random",
Type: 'c',
MajorNumber: 1,
MinorNumber: 8,
CgroupPermissions: "rwm",
FileMode: 0666,
},
}
DefaultAllowedDevices = append([]*Device{
// allow mknod for any device
{
Type: 'c',
MajorNumber: Wildcard,
MinorNumber: Wildcard,
CgroupPermissions: "m",
},
{
Type: 'b',
MajorNumber: Wildcard,
MinorNumber: Wildcard,
CgroupPermissions: "m",
},
{
Path: "/dev/console",
Type: 'c',
MajorNumber: 5,
MinorNumber: 1,
CgroupPermissions: "rwm",
},
{
Path: "/dev/tty0",
Type: 'c',
MajorNumber: 4,
MinorNumber: 0,
CgroupPermissions: "rwm",
},
{
Path: "/dev/tty1",
Type: 'c',
MajorNumber: 4,
MinorNumber: 1,
CgroupPermissions: "rwm",
},
// /dev/pts/ - pts namespaces are "coming soon"
{
Path: "",
Type: 'c',
MajorNumber: 136,
MinorNumber: Wildcard,
CgroupPermissions: "rwm",
},
{
Path: "",
Type: 'c',
MajorNumber: 5,
MinorNumber: 2,
CgroupPermissions: "rwm",
},
// tuntap
{
Path: "",
Type: 'c',
MajorNumber: 10,
MinorNumber: 200,
CgroupPermissions: "rwm",
},
/*// fuse
{
Path: "",
Type: 'c',
MajorNumber: 10,
MinorNumber: 229,
CgroupPermissions: "rwm",
},
// rtc
{
Path: "",
Type: 'c',
MajorNumber: 254,
MinorNumber: 0,
CgroupPermissions: "rwm",
},
*/
}, DefaultSimpleDevices...)
DefaultAutoCreatedDevices = append([]*Device{
{
// /dev/fuse is created but not allowed.
// This is to allow java to work. Because java
// Insists on there being a /dev/fuse
// https://github.com/docker/docker/issues/514
// https://github.com/docker/docker/issues/2393
//
Path: "/dev/fuse",
Type: 'c',
MajorNumber: 10,
MinorNumber: 229,
CgroupPermissions: "rwm",
},
}, DefaultSimpleDevices...)
)
......@@ -7,14 +7,12 @@ import (
"os"
"path/filepath"
"syscall"
)
const (
Wildcard = -1
"github.com/docker/libcontainer/configs"
)
var (
ErrNotADeviceNode = errors.New("not a device node")
ErrNotADevice = errors.New("not a device node")
)
// Testing dependencies
......@@ -23,45 +21,20 @@ var (
ioutilReadDir = ioutil.ReadDir
)
type Device struct {
Type rune `json:"type,omitempty"`
Path string `json:"path,omitempty"` // It is fine if this is an empty string in the case that you are using Wildcards
MajorNumber int64 `json:"major_number,omitempty"` // Use the wildcard constant for wildcards.
MinorNumber int64 `json:"minor_number,omitempty"` // Use the wildcard constant for wildcards.
CgroupPermissions string `json:"cgroup_permissions,omitempty"` // Typically just "rwm"
FileMode os.FileMode `json:"file_mode,omitempty"` // The permission bits of the file's mode
Uid uint32 `json:"uid,omitempty"`
Gid uint32 `json:"gid,omitempty"`
}
func GetDeviceNumberString(deviceNumber int64) string {
if deviceNumber == Wildcard {
return "*"
} else {
return fmt.Sprintf("%d", deviceNumber)
}
}
func (device *Device) GetCgroupAllowString() string {
return fmt.Sprintf("%c %s:%s %s", device.Type, GetDeviceNumberString(device.MajorNumber), GetDeviceNumberString(device.MinorNumber), device.CgroupPermissions)
}
// Given the path to a device and it's cgroup_permissions(which cannot be easilly queried) look up the information about a linux device and return that information as a Device struct.
func GetDevice(path, cgroupPermissions string) (*Device, error) {
func DeviceFromPath(path, permissions string) (*configs.Device, error) {
fileInfo, err := osLstat(path)
if err != nil {
return nil, err
}
var (
devType rune
mode = fileInfo.Mode()
fileModePermissionBits = os.FileMode.Perm(mode)
)
switch {
case mode&os.ModeDevice == 0:
return nil, ErrNotADeviceNode
return nil, ErrNotADevice
case mode&os.ModeCharDevice != 0:
fileModePermissionBits |= syscall.S_IFCHR
devType = 'c'
......@@ -69,36 +42,33 @@ func GetDevice(path, cgroupPermissions string) (*Device, error) {
fileModePermissionBits |= syscall.S_IFBLK
devType = 'b'
}
stat_t, ok := fileInfo.Sys().(*syscall.Stat_t)
if !ok {
return nil, fmt.Errorf("cannot determine the device number for device %s", path)
}
devNumber := int(stat_t.Rdev)
return &Device{
Type: devType,
Path: path,
MajorNumber: Major(devNumber),
MinorNumber: Minor(devNumber),
CgroupPermissions: cgroupPermissions,
FileMode: fileModePermissionBits,
Uid: stat_t.Uid,
Gid: stat_t.Gid,
return &configs.Device{
Type: devType,
Path: path,
Major: Major(devNumber),
Minor: Minor(devNumber),
Permissions: permissions,
FileMode: fileModePermissionBits,
Uid: stat_t.Uid,
Gid: stat_t.Gid,
}, nil
}
func GetHostDeviceNodes() ([]*Device, error) {
return getDeviceNodes("/dev")
func HostDevices() ([]*configs.Device, error) {
return getDevices("/dev")
}
func getDeviceNodes(path string) ([]*Device, error) {
func getDevices(path string) ([]*configs.Device, error) {
files, err := ioutilReadDir(path)
if err != nil {
return nil, err
}
out := []*Device{}
out := []*configs.Device{}
for _, f := range files {
switch {
case f.IsDir():
......@@ -106,7 +76,7 @@ func getDeviceNodes(path string) ([]*Device, error) {
case "pts", "shm", "fd", "mqueue":
continue
default:
sub, err := getDeviceNodes(filepath.Join(path, f.Name()))
sub, err := getDevices(filepath.Join(path, f.Name()))
if err != nil {
return nil, err
}
......@@ -117,16 +87,14 @@ func getDeviceNodes(path string) ([]*Device, error) {
case f.Name() == "console":
continue
}
device, err := GetDevice(filepath.Join(path, f.Name()), "rwm")
device, err := DeviceFromPath(filepath.Join(path, f.Name()), "rwm")
if err != nil {
if err == ErrNotADeviceNode {
if err == ErrNotADevice {
continue
}
return nil, err
}
out = append(out, device)
}
return out, nil
}
......@@ -6,7 +6,7 @@ import (
"testing"
)
func TestGetDeviceLstatFailure(t *testing.T) {
func TestDeviceFromPathLstatFailure(t *testing.T) {
testError := errors.New("test error")
// Override os.Lstat to inject error.
......@@ -14,13 +14,13 @@ func TestGetDeviceLstatFailure(t *testing.T) {
return nil, testError
}
_, err := GetDevice("", "")
_, err := DeviceFromPath("", "")
if err != testError {
t.Fatalf("Unexpected error %v, expected %v", err, testError)
}
}
func TestGetHostDeviceNodesIoutilReadDirFailure(t *testing.T) {
func TestHostDevicesIoutilReadDirFailure(t *testing.T) {
testError := errors.New("test error")
// Override ioutil.ReadDir to inject error.
......@@ -28,13 +28,13 @@ func TestGetHostDeviceNodesIoutilReadDirFailure(t *testing.T) {
return nil, testError
}
_, err := GetHostDeviceNodes()
_, err := HostDevices()
if err != testError {
t.Fatalf("Unexpected error %v, expected %v", err, testError)
}
}
func TestGetHostDeviceNodesIoutilReadDirDeepFailure(t *testing.T) {
func TestHostDevicesIoutilReadDirDeepFailure(t *testing.T) {
testError := errors.New("test error")
called := false
......@@ -54,7 +54,7 @@ func TestGetHostDeviceNodesIoutilReadDirDeepFailure(t *testing.T) {
return []os.FileInfo{fi}, nil
}
_, err := GetHostDeviceNodes()
_, err := HostDevices()
if err != testError {
t.Fatalf("Unexpected error %v, expected %v", err, testError)
}
......
......@@ -20,7 +20,3 @@ func Major(devNumber int) int64 {
func Minor(devNumber int) int64 {
return int64((devNumber & 0xff) | ((devNumber >> 12) & 0xfff00))
}
func Mkdev(majorNumber int64, minorNumber int64) int {
return int((majorNumber << 8) | (minorNumber & 0xff) | ((minorNumber & 0xfff00) << 12))
}
% nsinit User Manual
% docker/libcontainer
% JAN 2015
NAME:
nsinit - A low-level utility for managing containers.
It is used to spawn new containers or join existing containers.
USAGE:
nsinit [global options] command [command options] [arguments...]
VERSION:
0.1
COMMANDS:
config display the container configuration
exec execute a new command inside a container
init runs the init process inside the namespace
oom display oom notifications for a container
pause pause the container's processes
stats display statistics for the container
unpause unpause the container's processes
help, h shows a list of commands or help for one command
EXAMPLES:
Get the <container_id> of an already running docker container.
`sudo docker ps` will return the list of all the running containers.
take the <container_id> (e.g. 4addb0b2d307) and go to its config directory
`/var/lib/docker/execdriver/native/4addb0b2d307` and here you can run the nsinit
command line utility.
e.g. `nsinit exec /bin/bash` will start a shell on the already running container.
# HISTORY
Jan 2015, Originally compiled by Shishir Mahajan (shishir dot mahajan at redhat dot com)
based on nsinit source material and internal work.
package libcontainer
import "io"
// API error code type.
type ErrorCode int
......@@ -8,29 +10,52 @@ const (
// Factory errors
IdInUse ErrorCode = iota
InvalidIdFormat
// TODO: add Load errors
// Container errors
ContainerDestroyed
ContainerNotExists
ContainerPaused
ContainerNotStopped
ContainerNotRunning
// Process errors
ProcessNotExecuted
// Common errors
ConfigInvalid
SystemError
)
func (c ErrorCode) String() string {
switch c {
case IdInUse:
return "Id already in use"
case InvalidIdFormat:
return "Invalid format"
case ContainerPaused:
return "Container paused"
case ConfigInvalid:
return "Invalid configuration"
case SystemError:
return "System error"
case ContainerNotExists:
return "Container does not exist"
case ContainerNotStopped:
return "Container is not stopped"
case ContainerNotRunning:
return "Container is not running"
default:
return "Unknown error"
}
}
// API Error type.
type Error interface {
error
// Returns the stack trace, if any, which identifies the
// point at which the error occurred.
Stack() []byte
// Returns a verbose string including the error message
// and a representation of the stack trace suitable for
// printing.
Detail() string
Detail(w io.Writer) error
// Returns the error code for this error.
Code() ErrorCode
......
package libcontainer
import "testing"
func TestErrorCode(t *testing.T) {
codes := map[ErrorCode]string{
IdInUse: "Id already in use",
InvalidIdFormat: "Invalid format",
ContainerPaused: "Container paused",
ConfigInvalid: "Invalid configuration",
SystemError: "System error",
ContainerNotExists: "Container does not exist",
}
for code, expected := range codes {
if actual := code.String(); actual != expected {
t.Fatalf("expected string %q but received %q", expected, actual)
}
}
}
package libcontainer
type Factory interface {
import (
"github.com/docker/libcontainer/configs"
)
type Factory interface {
// Creates a new container with the given id and starts the initial process inside it.
// id must be a string containing only letters, digits and underscores and must contain
// between 1 and 1024 characters, inclusive.
......@@ -11,22 +14,32 @@ type Factory interface {
//
// Returns the new container with a running process.
//
// Errors:
// errors:
// IdInUse - id is already in use by a container
// InvalidIdFormat - id has incorrect format
// ConfigInvalid - config is invalid
// SystemError - System error
// Systemerror - System error
//
// On error, any partially created container parts are cleaned up (the operation is atomic).
Create(id string, config *Config) (Container, Error)
Create(id string, config *configs.Config) (Container, error)
// Load takes an ID for an existing container and reconstructs the container
// from the state.
// Load takes an ID for an existing container and returns the container information
// from the state. This presents a read only view of the container.
//
// Errors:
// errors:
// Path does not exist
// Container is stopped
// System error
// TODO: fix description
Load(id string) (Container, Error)
Load(id string) (Container, error)
// StartInitialization is an internal API to libcontainer used during the rexec of the
// container.
//
// Errors:
// Pipe connection error
// System error
StartInitialization() error
// Type returns info string about factory type (e.g. lxc, libcontainer...)
Type() string
}
// +build linux
package libcontainer
import (
"encoding/json"
"fmt"
"io/ioutil"
"os"
"os/exec"
"path/filepath"
"regexp"
"strconv"
"syscall"
"github.com/docker/docker/pkg/mount"
"github.com/docker/libcontainer/cgroups"
"github.com/docker/libcontainer/cgroups/fs"
"github.com/docker/libcontainer/cgroups/systemd"
"github.com/docker/libcontainer/configs"
"github.com/docker/libcontainer/configs/validate"
)
const (
stateFilename = "state.json"
)
var (
idRegex = regexp.MustCompile(`^[\w_]+$`)
maxIdLen = 1024
)
// InitArgs returns an options func to configure a LinuxFactory with the
// provided init arguments.
func InitArgs(args ...string) func(*LinuxFactory) error {
return func(l *LinuxFactory) error {
name := args[0]
if filepath.Base(name) == name {
if lp, err := exec.LookPath(name); err == nil {
name = lp
}
}
l.InitPath = name
l.InitArgs = append([]string{name}, args[1:]...)
return nil
}
}
// InitPath returns an options func to configure a LinuxFactory with the
// provided absolute path to the init binary and arguements.
func InitPath(path string, args ...string) func(*LinuxFactory) error {
return func(l *LinuxFactory) error {
l.InitPath = path
l.InitArgs = args
return nil
}
}
// SystemdCgroups is an options func to configure a LinuxFactory to return
// containers that use systemd to create and manage cgroups.
func SystemdCgroups(l *LinuxFactory) error {
l.NewCgroupsManager = func(config *configs.Cgroup, paths map[string]string) cgroups.Manager {
return &systemd.Manager{
Cgroups: config,
Paths: paths,
}
}
return nil
}
// Cgroupfs is an options func to configure a LinuxFactory to return
// containers that use the native cgroups filesystem implementation to
// create and manage cgroups.
func Cgroupfs(l *LinuxFactory) error {
l.NewCgroupsManager = func(config *configs.Cgroup, paths map[string]string) cgroups.Manager {
return &fs.Manager{
Cgroups: config,
Paths: paths,
}
}
return nil
}
// TmpfsRoot is an option func to mount LinuxFactory.Root to tmpfs.
func TmpfsRoot(l *LinuxFactory) error {
mounted, err := mount.Mounted(l.Root)
if err != nil {
return err
}
if !mounted {
if err := syscall.Mount("tmpfs", l.Root, "tmpfs", 0, ""); err != nil {
return err
}
}
return nil
}
// New returns a linux based container factory based in the root directory and
// configures the factory with the provided option funcs.
func New(root string, options ...func(*LinuxFactory) error) (Factory, error) {
if root != "" {
if err := os.MkdirAll(root, 0700); err != nil {
return nil, newGenericError(err, SystemError)
}
}
l := &LinuxFactory{
Root: root,
Validator: validate.New(),
}
InitArgs(os.Args[0], "init")(l)
Cgroupfs(l)
for _, opt := range options {
if err := opt(l); err != nil {
return nil, err
}
}
return l, nil
}
// LinuxFactory implements the default factory interface for linux based systems.
type LinuxFactory struct {
// Root directory for the factory to store state.
Root string
// InitPath is the absolute path to the init binary.
InitPath string
// InitArgs are arguments for calling the init responsibilities for spawning
// a container.
InitArgs []string
// Validator provides validation to container configurations.
Validator validate.Validator
// NewCgroupsManager returns an initialized cgroups manager for a single container.
NewCgroupsManager func(config *configs.Cgroup, paths map[string]string) cgroups.Manager
}
func (l *LinuxFactory) Create(id string, config *configs.Config) (Container, error) {
if l.Root == "" {
return nil, newGenericError(fmt.Errorf("invalid root"), ConfigInvalid)
}
if err := l.validateID(id); err != nil {
return nil, err
}
if err := l.Validator.Validate(config); err != nil {
return nil, newGenericError(err, ConfigInvalid)
}
containerRoot := filepath.Join(l.Root, id)
if _, err := os.Stat(containerRoot); err == nil {
return nil, newGenericError(fmt.Errorf("Container with id exists: %v", id), IdInUse)
} else if !os.IsNotExist(err) {
return nil, newGenericError(err, SystemError)
}
if err := os.MkdirAll(containerRoot, 0700); err != nil {
return nil, newGenericError(err, SystemError)
}
return &linuxContainer{
id: id,
root: containerRoot,
config: config,
initPath: l.InitPath,
initArgs: l.InitArgs,
cgroupManager: l.NewCgroupsManager(config.Cgroups, nil),
}, nil
}
func (l *LinuxFactory) Load(id string) (Container, error) {
if l.Root == "" {
return nil, newGenericError(fmt.Errorf("invalid root"), ConfigInvalid)
}
containerRoot := filepath.Join(l.Root, id)
state, err := l.loadState(containerRoot)
if err != nil {
return nil, err
}
r := &restoredProcess{
processPid: state.InitProcessPid,
processStartTime: state.InitProcessStartTime,
}
return &linuxContainer{
initProcess: r,
id: id,
config: &state.Config,
initPath: l.InitPath,
initArgs: l.InitArgs,
cgroupManager: l.NewCgroupsManager(state.Config.Cgroups, state.CgroupPaths),
root: containerRoot,
}, nil
}
func (l *LinuxFactory) Type() string {
return "libcontainer"
}
// StartInitialization loads a container by opening the pipe fd from the parent to read the configuration and state
// This is a low level implementation detail of the reexec and should not be consumed externally
func (l *LinuxFactory) StartInitialization() (err error) {
pipefd, err := strconv.Atoi(os.Getenv("_LIBCONTAINER_INITPIPE"))
if err != nil {
return err
}
var (
pipe = os.NewFile(uintptr(pipefd), "pipe")
it = initType(os.Getenv("_LIBCONTAINER_INITTYPE"))
)
// clear the current process's environment to clean any libcontainer
// specific env vars.
os.Clearenv()
defer func() {
// if we have an error during the initialization of the container's init then send it back to the
// parent process in the form of an initError.
if err != nil {
// ensure that any data sent from the parent is consumed so it doesn't
// receive ECONNRESET when the child writes to the pipe.
ioutil.ReadAll(pipe)
if err := json.NewEncoder(pipe).Encode(newSystemError(err)); err != nil {
panic(err)
}
}
// ensure that this pipe is always closed
pipe.Close()
}()
i, err := newContainerInit(it, pipe)
if err != nil {
return err
}
return i.Init()
}
func (l *LinuxFactory) loadState(root string) (*State, error) {
f, err := os.Open(filepath.Join(root, stateFilename))
if err != nil {
if os.IsNotExist(err) {
return nil, newGenericError(err, ContainerNotExists)
}
return nil, newGenericError(err, SystemError)
}
defer f.Close()
var state *State
if err := json.NewDecoder(f).Decode(&state); err != nil {
return nil, newGenericError(err, SystemError)
}
return state, nil
}
func (l *LinuxFactory) validateID(id string) error {
if !idRegex.MatchString(id) {
return newGenericError(fmt.Errorf("Invalid id format: %v", id), InvalidIdFormat)
}
if len(id) > maxIdLen {
return newGenericError(fmt.Errorf("Invalid id format: %v", id), InvalidIdFormat)
}
return nil
}
// restoredProcess represents a process where the calling process may or may not be
// the parent process. This process is created when a factory loads a container from
// a persisted state.
type restoredProcess struct {
processPid int
processStartTime string
}
func (p *restoredProcess) start() error {
return newGenericError(fmt.Errorf("restored process cannot be started"), SystemError)
}
func (p *restoredProcess) pid() int {
return p.processPid
}
func (p *restoredProcess) terminate() error {
return newGenericError(fmt.Errorf("restored process cannot be terminated"), SystemError)
}
func (p *restoredProcess) wait() (*os.ProcessState, error) {
return nil, newGenericError(fmt.Errorf("restored process cannot be waited on"), SystemError)
}
func (p *restoredProcess) startTime() (string, error) {
return p.processStartTime, nil
}
func (p *restoredProcess) signal(s os.Signal) error {
return newGenericError(fmt.Errorf("restored process cannot be signaled"), SystemError)
}
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment