diff --git a/vendor/src/github.com/containernetworking/cni/pkg/ns/README.md b/vendor/src/github.com/containernetworking/cni/pkg/ns/README.md new file mode 100644 index 00000000..99aed9c8 --- /dev/null +++ b/vendor/src/github.com/containernetworking/cni/pkg/ns/README.md @@ -0,0 +1,34 @@ +### Namespaces, Threads, and Go +On Linux each OS thread can have a different network namespace. Go's thread scheduling model switches goroutines between OS threads based on OS thread load and whether the goroutine would block other goroutines. This can result in a goroutine switching network namespaces without notice and lead to errors in your code. + +### Namespace Switching +Switching namespaces with the `ns.Set()` method is not recommended without additional strategies to prevent unexpected namespace changes when your goroutines switch OS threads. + +Go provides the `runtime.LockOSThread()` function to ensure a specific goroutine executes on its current OS thread and prevents any other goroutine from running in that thread until the locked one exits. Careful usage of `LockOSThread()` and goroutines can provide good control over which network namespace a given goroutine executes in. + +For example, you cannot rely on the `ns.Set()` namespace being the current namespace after the `Set()` call unless you do two things. First, the goroutine calling `Set()` must have previously called `LockOSThread()`. Second, you must ensure `runtime.UnlockOSThread()` is not called somewhere in-between. You also cannot rely on the initial network namespace remaining the current network namespace if any other code in your program switches namespaces, unless you have already called `LockOSThread()` in that goroutine. Note that `LockOSThread()` prevents the Go scheduler from optimally scheduling goroutines for best performance, so `LockOSThread()` should only be used in small, isolated goroutines that release the lock quickly. + +### Do() The Recommended Thing +The `ns.Do()` method provides control over network namespaces for you by implementing these strategies. All code dependent on a particular network namespace (including the root namespace) should be wrapped in the `ns.Do()` method to ensure the correct namespace is selected for the duration of your code. For example: + +```go +targetNs, err := ns.NewNS() +if err != nil { + return err +} +err = targetNs.Do(func(hostNs ns.NetNS) error { + dummy := &netlink.Dummy{ + LinkAttrs: netlink.LinkAttrs{ + Name: "dummy0", + }, + } + return netlink.LinkAdd(dummy) +}) +``` + +Note this requirement to wrap every network call is very onerous - any libraries you call might call out to network services such as DNS, and all such calls need to be protected after you call `ns.Do()`. The CNI plugins all exit very soon after calling `ns.Do()` which helps to minimize the problem. + +### Further Reading + - https://github.com/golang/go/wiki/LockOSThread + - http://morsmachine.dk/go-scheduler + - https://github.com/containernetworking/cni/issues/262 diff --git a/vendor/src/github.com/containernetworking/cni/pkg/ns/ns.go b/vendor/src/github.com/containernetworking/cni/pkg/ns/ns.go new file mode 100644 index 00000000..3246ebf3 --- /dev/null +++ b/vendor/src/github.com/containernetworking/cni/pkg/ns/ns.go @@ -0,0 +1,302 @@ +// Copyright 2015 CNI authors +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. + +package ns + +import ( + "crypto/rand" + "fmt" + "os" + "path" + "runtime" + "sync" + "syscall" + + "golang.org/x/sys/unix" +) + +type NetNS interface { + // Executes the passed closure in this object's network namespace, + // attempting to restore the original namespace before returning. + // However, since each OS thread can have a different network namespace, + // and Go's thread scheduling is highly variable, callers cannot + // guarantee any specific namespace is set unless operations that + // require that namespace are wrapped with Do(). Also, no code called + // from Do() should call runtime.UnlockOSThread(), or the risk + // of executing code in an incorrect namespace will be greater. See + // https://github.com/golang/go/wiki/LockOSThread for further details. + Do(toRun func(NetNS) error) error + + // Sets the current network namespace to this object's network namespace. + // Note that since Go's thread scheduling is highly variable, callers + // cannot guarantee the requested namespace will be the current namespace + // after this function is called; to ensure this wrap operations that + // require the namespace with Do() instead. + Set() error + + // Returns the filesystem path representing this object's network namespace + Path() string + + // Returns a file descriptor representing this object's network namespace + Fd() uintptr + + // Cleans up this instance of the network namespace; if this instance + // is the last user the namespace will be destroyed + Close() error +} + +type netNS struct { + file *os.File + mounted bool + closed bool +} + +func getCurrentThreadNetNSPath() string { + // /proc/self/ns/net returns the namespace of the main thread, not + // of whatever thread this goroutine is running on. Make sure we + // use the thread's net namespace since the thread is switching around + return fmt.Sprintf("/proc/%d/task/%d/ns/net", os.Getpid(), unix.Gettid()) +} + +// Returns an object representing the current OS thread's network namespace +func GetCurrentNS() (NetNS, error) { + return GetNS(getCurrentThreadNetNSPath()) +} + +const ( + // https://github.com/torvalds/linux/blob/master/include/uapi/linux/magic.h + NSFS_MAGIC = 0x6e736673 + PROCFS_MAGIC = 0x9fa0 +) + +type NSPathNotExistErr struct{ msg string } + +func (e NSPathNotExistErr) Error() string { return e.msg } + +type NSPathNotNSErr struct{ msg string } + +func (e NSPathNotNSErr) Error() string { return e.msg } + +func IsNSorErr(nspath string) error { + stat := syscall.Statfs_t{} + if err := syscall.Statfs(nspath, &stat); err != nil { + if os.IsNotExist(err) { + err = NSPathNotExistErr{msg: fmt.Sprintf("failed to Statfs %q: %v", nspath, err)} + } else { + err = fmt.Errorf("failed to Statfs %q: %v", nspath, err) + } + return err + } + + switch stat.Type { + case PROCFS_MAGIC, NSFS_MAGIC: + return nil + default: + return NSPathNotNSErr{msg: fmt.Sprintf("unknown FS magic on %q: %x", nspath, stat.Type)} + } +} + +// Returns an object representing the namespace referred to by @path +func GetNS(nspath string) (NetNS, error) { + err := IsNSorErr(nspath) + if err != nil { + return nil, err + } + + fd, err := os.Open(nspath) + if err != nil { + return nil, err + } + + return &netNS{file: fd}, nil +} + +// Creates a new persistent network namespace and returns an object +// representing that namespace, without switching to it +func NewNS() (NetNS, error) { + const nsRunDir = "/var/run/netns" + + b := make([]byte, 16) + _, err := rand.Reader.Read(b) + if err != nil { + return nil, fmt.Errorf("failed to generate random netns name: %v", err) + } + + err = os.MkdirAll(nsRunDir, 0755) + if err != nil { + return nil, err + } + + // create an empty file at the mount point + nsName := fmt.Sprintf("cni-%x-%x-%x-%x-%x", b[0:4], b[4:6], b[6:8], b[8:10], b[10:]) + nsPath := path.Join(nsRunDir, nsName) + mountPointFd, err := os.Create(nsPath) + if err != nil { + return nil, err + } + mountPointFd.Close() + + // Ensure the mount point is cleaned up on errors; if the namespace + // was successfully mounted this will have no effect because the file + // is in-use + defer os.RemoveAll(nsPath) + + var wg sync.WaitGroup + wg.Add(1) + + // do namespace work in a dedicated goroutine, so that we can safely + // Lock/Unlock OSThread without upsetting the lock/unlock state of + // the caller of this function + var fd *os.File + go (func() { + defer wg.Done() + runtime.LockOSThread() + + var origNS NetNS + origNS, err = GetNS(getCurrentThreadNetNSPath()) + if err != nil { + return + } + defer origNS.Close() + + // create a new netns on the current thread + err = unix.Unshare(unix.CLONE_NEWNET) + if err != nil { + return + } + defer origNS.Set() + + // bind mount the new netns from the current thread onto the mount point + err = unix.Mount(getCurrentThreadNetNSPath(), nsPath, "none", unix.MS_BIND, "") + if err != nil { + return + } + + fd, err = os.Open(nsPath) + if err != nil { + return + } + })() + wg.Wait() + + if err != nil { + unix.Unmount(nsPath, unix.MNT_DETACH) + return nil, fmt.Errorf("failed to create namespace: %v", err) + } + + return &netNS{file: fd, mounted: true}, nil +} + +func (ns *netNS) Path() string { + return ns.file.Name() +} + +func (ns *netNS) Fd() uintptr { + return ns.file.Fd() +} + +func (ns *netNS) errorIfClosed() error { + if ns.closed { + return fmt.Errorf("%q has already been closed", ns.file.Name()) + } + return nil +} + +func (ns *netNS) Close() error { + if err := ns.errorIfClosed(); err != nil { + return err + } + + if err := ns.file.Close(); err != nil { + return fmt.Errorf("Failed to close %q: %v", ns.file.Name(), err) + } + ns.closed = true + + if ns.mounted { + if err := unix.Unmount(ns.file.Name(), unix.MNT_DETACH); err != nil { + return fmt.Errorf("Failed to unmount namespace %s: %v", ns.file.Name(), err) + } + if err := os.RemoveAll(ns.file.Name()); err != nil { + return fmt.Errorf("Failed to clean up namespace %s: %v", ns.file.Name(), err) + } + ns.mounted = false + } + + return nil +} + +func (ns *netNS) Do(toRun func(NetNS) error) error { + if err := ns.errorIfClosed(); err != nil { + return err + } + + containedCall := func(hostNS NetNS) error { + threadNS, err := GetNS(getCurrentThreadNetNSPath()) + if err != nil { + return fmt.Errorf("failed to open current netns: %v", err) + } + defer threadNS.Close() + + // switch to target namespace + if err = ns.Set(); err != nil { + return fmt.Errorf("error switching to ns %v: %v", ns.file.Name(), err) + } + defer threadNS.Set() // switch back + + return toRun(hostNS) + } + + // save a handle to current network namespace + hostNS, err := GetNS(getCurrentThreadNetNSPath()) + if err != nil { + return fmt.Errorf("Failed to open current namespace: %v", err) + } + defer hostNS.Close() + + var wg sync.WaitGroup + wg.Add(1) + + var innerError error + go func() { + defer wg.Done() + runtime.LockOSThread() + innerError = containedCall(hostNS) + }() + wg.Wait() + + return innerError +} + +func (ns *netNS) Set() error { + if err := ns.errorIfClosed(); err != nil { + return err + } + + if _, _, err := unix.Syscall(unix.SYS_SETNS, ns.Fd(), uintptr(unix.CLONE_NEWNET), 0); err != 0 { + return fmt.Errorf("Error switching to ns %v: %v", ns.file.Name(), err) + } + + return nil +} + +// WithNetNSPath executes the passed closure under the given network +// namespace, restoring the original namespace afterwards. +func WithNetNSPath(nspath string, toRun func(NetNS) error) error { + ns, err := GetNS(nspath) + if err != nil { + return err + } + defer ns.Close() + return ns.Do(toRun) +}