Files
podman/pkg/specgen/namespaces.go
Paul Holzinger 2805c7353b pkg/specgen: parse default network mode on server
When podman-remote is used we should not resolve the default network
mode on the client. Defaults should be set on the server. In this case
this is important because we have different defaults for root/rootless.
So when the client is rootless and the server is root we must pick the
root default.

Note that this already worked when --network was set since we did not
parsed the flag in this case. To reproduce you need --network=default.

Also removed a unused function.

[NO NEW TESTS NEEDED] I tested it manually but I am not sure how I can
hook a test like this up in CI. The client would need to run as rootless
and the server as root or the other way around.

Fixes #14368

Signed-off-by: Paul Holzinger <pholzing@redhat.com>
2022-06-01 17:13:30 +02:00

539 lines
16 KiB
Go

package specgen
import (
"fmt"
"net"
"os"
"strings"
"github.com/containers/common/libnetwork/types"
"github.com/containers/common/pkg/cgroups"
cutil "github.com/containers/common/pkg/util"
"github.com/containers/podman/v4/libpod/define"
"github.com/containers/podman/v4/pkg/util"
"github.com/containers/storage"
spec "github.com/opencontainers/runtime-spec/specs-go"
"github.com/opencontainers/runtime-tools/generate"
"github.com/pkg/errors"
)
type NamespaceMode string
const (
// Default indicates the spec generator should determine
// a sane default
Default NamespaceMode = "default"
// Host means the the namespace is derived from
// the host
Host NamespaceMode = "host"
// Path is the path to a namespace
Path NamespaceMode = "path"
// FromContainer means namespace is derived from a
// different container
FromContainer NamespaceMode = "container"
// FromPod indicates the namespace is derived from a pod
FromPod NamespaceMode = "pod"
// Private indicates the namespace is private
Private NamespaceMode = "private"
// Shareable indicates the namespace is shareable
Shareable NamespaceMode = "shareable"
// None indicates the IPC namespace is created without mounting /dev/shm
None NamespaceMode = "none"
// NoNetwork indicates no network namespace should
// be joined. loopback should still exists.
// Only used with the network namespace, invalid otherwise.
NoNetwork NamespaceMode = "none"
// Bridge indicates that a CNI network stack
// should be used.
// Only used with the network namespace, invalid otherwise.
Bridge NamespaceMode = "bridge"
// Slirp indicates that a slirp4netns network stack should
// be used.
// Only used with the network namespace, invalid otherwise.
Slirp NamespaceMode = "slirp4netns"
// KeepId indicates a user namespace to keep the owner uid inside
// of the namespace itself.
// Only used with the user namespace, invalid otherwise.
KeepID NamespaceMode = "keep-id"
// NoMap indicates a user namespace to keep the owner uid out
// of the namespace itself.
// Only used with the user namespace, invalid otherwise.
NoMap NamespaceMode = "no-map"
// Auto indicates to automatically create a user namespace.
// Only used with the user namespace, invalid otherwise.
Auto NamespaceMode = "auto"
// DefaultKernelNamespaces is a comma-separated list of default kernel
// namespaces.
DefaultKernelNamespaces = "ipc,net,uts"
)
// Namespace describes the namespace
type Namespace struct {
NSMode NamespaceMode `json:"nsmode,omitempty"`
Value string `json:"value,omitempty"`
}
// IsDefault returns whether the namespace is set to the default setting (which
// also includes the empty string).
func (n *Namespace) IsDefault() bool {
return n.NSMode == Default || n.NSMode == ""
}
// IsHost returns a bool if the namespace is host based
func (n *Namespace) IsHost() bool {
return n.NSMode == Host
}
// IsNone returns a bool if the namespace is set to none
func (n *Namespace) IsNone() bool {
return n.NSMode == None
}
// IsBridge returns a bool if the namespace is a Bridge
func (n *Namespace) IsBridge() bool {
return n.NSMode == Bridge
}
// IsPath indicates via bool if the namespace is based on a path
func (n *Namespace) IsPath() bool {
return n.NSMode == Path
}
// IsContainer indicates via bool if the namespace is based on a container
func (n *Namespace) IsContainer() bool {
return n.NSMode == FromContainer
}
// IsPod indicates via bool if the namespace is based on a pod
func (n *Namespace) IsPod() bool {
return n.NSMode == FromPod
}
// IsPrivate indicates the namespace is private
func (n *Namespace) IsPrivate() bool {
return n.NSMode == Private
}
// IsAuto indicates the namespace is auto
func (n *Namespace) IsAuto() bool {
return n.NSMode == Auto
}
// IsKeepID indicates the namespace is KeepID
func (n *Namespace) IsKeepID() bool {
return n.NSMode == KeepID
}
// IsNoMap indicates the namespace is NoMap
func (n *Namespace) IsNoMap() bool {
return n.NSMode == NoMap
}
func (n *Namespace) String() string {
if n.Value != "" {
return fmt.Sprintf("%s:%s", n.NSMode, n.Value)
}
return string(n.NSMode)
}
func validateUserNS(n *Namespace) error {
if n == nil {
return nil
}
switch n.NSMode {
case Auto, KeepID, NoMap:
return nil
}
return n.validate()
}
func validateNetNS(n *Namespace) error {
if n == nil {
return nil
}
switch n.NSMode {
case Slirp:
break
case "", Default, Host, Path, FromContainer, FromPod, Private, NoNetwork, Bridge:
break
default:
return errors.Errorf("invalid network %q", n.NSMode)
}
// Path and From Container MUST have a string value set
if n.NSMode == Path || n.NSMode == FromContainer {
if len(n.Value) < 1 {
return errors.Errorf("namespace mode %s requires a value", n.NSMode)
}
} else if n.NSMode != Slirp {
// All others except must NOT set a string value
if len(n.Value) > 0 {
return errors.Errorf("namespace value %s cannot be provided with namespace mode %s", n.Value, n.NSMode)
}
}
return nil
}
func validateIPCNS(n *Namespace) error {
if n == nil {
return nil
}
switch n.NSMode {
case Shareable, None:
return nil
}
return n.validate()
}
// Validate perform simple validation on the namespace to make sure it is not
// invalid from the get-go
func (n *Namespace) validate() error {
if n == nil {
return nil
}
switch n.NSMode {
case "", Default, Host, Path, FromContainer, FromPod, Private:
// Valid, do nothing
case NoNetwork, Bridge, Slirp:
return errors.Errorf("cannot use network modes with non-network namespace")
default:
return errors.Errorf("invalid namespace type %s specified", n.NSMode)
}
// Path and From Container MUST have a string value set
if n.NSMode == Path || n.NSMode == FromContainer {
if len(n.Value) < 1 {
return errors.Errorf("namespace mode %s requires a value", n.NSMode)
}
} else {
// All others must NOT set a string value
if len(n.Value) > 0 {
return errors.Errorf("namespace value %s cannot be provided with namespace mode %s", n.Value, n.NSMode)
}
}
return nil
}
// ParseNamespace parses a namespace in string form.
// This is not intended for the network namespace, which has a separate
// function.
func ParseNamespace(ns string) (Namespace, error) {
toReturn := Namespace{}
switch {
case ns == "pod":
toReturn.NSMode = FromPod
case ns == "host":
toReturn.NSMode = Host
case ns == "private", ns == "":
toReturn.NSMode = Private
case strings.HasPrefix(ns, "ns:"):
split := strings.SplitN(ns, ":", 2)
if len(split) != 2 {
return toReturn, errors.Errorf("must provide a path to a namespace when specifying \"ns:\"")
}
toReturn.NSMode = Path
toReturn.Value = split[1]
case strings.HasPrefix(ns, "container:"):
split := strings.SplitN(ns, ":", 2)
if len(split) != 2 {
return toReturn, errors.Errorf("must provide name or ID or a container when specifying \"container:\"")
}
toReturn.NSMode = FromContainer
toReturn.Value = split[1]
default:
return toReturn, errors.Errorf("unrecognized namespace mode %s passed", ns)
}
return toReturn, nil
}
// ParseCgroupNamespace parses a cgroup namespace specification in string
// form.
func ParseCgroupNamespace(ns string) (Namespace, error) {
toReturn := Namespace{}
// Cgroup is host for v1, private for v2.
// We can't trust c/common for this, as it only assumes private.
cgroupsv2, err := cgroups.IsCgroup2UnifiedMode()
if err != nil {
return toReturn, err
}
if cgroupsv2 {
switch ns {
case "host":
toReturn.NSMode = Host
case "private", "":
toReturn.NSMode = Private
default:
return toReturn, errors.Errorf("unrecognized cgroup namespace mode %s passed", ns)
}
} else {
toReturn.NSMode = Host
}
return toReturn, nil
}
// ParseIPCNamespace parses a ipc namespace specification in string
// form.
func ParseIPCNamespace(ns string) (Namespace, error) {
toReturn := Namespace{}
switch {
case ns == "shareable", ns == "":
toReturn.NSMode = Shareable
return toReturn, nil
case ns == "none":
toReturn.NSMode = None
return toReturn, nil
}
return ParseNamespace(ns)
}
// ParseUserNamespace parses a user namespace specification in string
// form.
func ParseUserNamespace(ns string) (Namespace, error) {
toReturn := Namespace{}
switch {
case ns == "auto":
toReturn.NSMode = Auto
return toReturn, nil
case strings.HasPrefix(ns, "auto:"):
split := strings.SplitN(ns, ":", 2)
if len(split) != 2 {
return toReturn, errors.Errorf("invalid setting for auto: mode")
}
toReturn.NSMode = Auto
toReturn.Value = split[1]
return toReturn, nil
case ns == "keep-id":
toReturn.NSMode = KeepID
return toReturn, nil
case ns == "nomap":
toReturn.NSMode = NoMap
return toReturn, nil
case ns == "":
toReturn.NSMode = Host
return toReturn, nil
}
return ParseNamespace(ns)
}
// ParseNetworkFlag parses a network string slice into the network options
// If the input is nil or empty it will use the default setting from containers.conf
func ParseNetworkFlag(networks []string) (Namespace, map[string]types.PerNetworkOptions, map[string][]string, error) {
var networkOptions map[string][]string
// by default we try to use the containers.conf setting
// if we get at least one value use this instead
ns := containerConfig.Containers.NetNS
if len(networks) > 0 {
ns = networks[0]
}
toReturn := Namespace{}
podmanNetworks := make(map[string]types.PerNetworkOptions)
switch {
case ns == string(Slirp), strings.HasPrefix(ns, string(Slirp)+":"):
parts := strings.SplitN(ns, ":", 2)
if len(parts) > 1 {
networkOptions = make(map[string][]string)
networkOptions[parts[0]] = strings.Split(parts[1], ",")
}
toReturn.NSMode = Slirp
case ns == string(FromPod):
toReturn.NSMode = FromPod
case ns == "" || ns == string(Default) || ns == string(Private):
toReturn.NSMode = Private
case ns == string(Bridge), strings.HasPrefix(ns, string(Bridge)+":"):
toReturn.NSMode = Bridge
parts := strings.SplitN(ns, ":", 2)
netOpts := types.PerNetworkOptions{}
if len(parts) > 1 {
var err error
netOpts, err = parseBridgeNetworkOptions(parts[1])
if err != nil {
return toReturn, nil, nil, err
}
}
// we have to set the special default network name here
podmanNetworks["default"] = netOpts
case ns == string(NoNetwork):
toReturn.NSMode = NoNetwork
case ns == string(Host):
toReturn.NSMode = Host
case strings.HasPrefix(ns, "ns:"):
split := strings.SplitN(ns, ":", 2)
if len(split) != 2 {
return toReturn, nil, nil, errors.Errorf("must provide a path to a namespace when specifying \"ns:\"")
}
toReturn.NSMode = Path
toReturn.Value = split[1]
case strings.HasPrefix(ns, string(FromContainer)+":"):
split := strings.SplitN(ns, ":", 2)
if len(split) != 2 {
return toReturn, nil, nil, errors.Errorf("must provide name or ID or a container when specifying \"container:\"")
}
toReturn.NSMode = FromContainer
toReturn.Value = split[1]
default:
// we should have a normal network
parts := strings.SplitN(ns, ":", 2)
if len(parts) == 1 {
// Assume we have been given a comma separated list of networks for backwards compat.
networkList := strings.Split(ns, ",")
for _, net := range networkList {
podmanNetworks[net] = types.PerNetworkOptions{}
}
} else {
if parts[0] == "" {
return toReturn, nil, nil, errors.New("network name cannot be empty")
}
netOpts, err := parseBridgeNetworkOptions(parts[1])
if err != nil {
return toReturn, nil, nil, errors.Wrapf(err, "invalid option for network %s", parts[0])
}
podmanNetworks[parts[0]] = netOpts
}
// networks need bridge mode
toReturn.NSMode = Bridge
}
if len(networks) > 1 {
if !toReturn.IsBridge() {
return toReturn, nil, nil, errors.Wrapf(define.ErrInvalidArg, "cannot set multiple networks without bridge network mode, selected mode %s", toReturn.NSMode)
}
for _, network := range networks[1:] {
parts := strings.SplitN(network, ":", 2)
if parts[0] == "" {
return toReturn, nil, nil, errors.Wrapf(define.ErrInvalidArg, "network name cannot be empty")
}
if cutil.StringInSlice(parts[0], []string{string(Bridge), string(Slirp), string(FromPod), string(NoNetwork),
string(Default), string(Private), string(Path), string(FromContainer), string(Host)}) {
return toReturn, nil, nil, errors.Wrapf(define.ErrInvalidArg, "can only set extra network names, selected mode %s conflicts with bridge", parts[0])
}
netOpts := types.PerNetworkOptions{}
if len(parts) > 1 {
var err error
netOpts, err = parseBridgeNetworkOptions(parts[1])
if err != nil {
return toReturn, nil, nil, errors.Wrapf(err, "invalid option for network %s", parts[0])
}
}
podmanNetworks[parts[0]] = netOpts
}
}
return toReturn, podmanNetworks, networkOptions, nil
}
func parseBridgeNetworkOptions(opts string) (types.PerNetworkOptions, error) {
netOpts := types.PerNetworkOptions{}
if len(opts) == 0 {
return netOpts, nil
}
allopts := strings.Split(opts, ",")
for _, opt := range allopts {
split := strings.SplitN(opt, "=", 2)
switch split[0] {
case "ip", "ip6":
ip := net.ParseIP(split[1])
if ip == nil {
return netOpts, errors.Errorf("invalid ip address %q", split[1])
}
netOpts.StaticIPs = append(netOpts.StaticIPs, ip)
case "mac":
mac, err := net.ParseMAC(split[1])
if err != nil {
return netOpts, err
}
netOpts.StaticMAC = types.HardwareAddr(mac)
case "alias":
if split[1] == "" {
return netOpts, errors.New("alias cannot be empty")
}
netOpts.Aliases = append(netOpts.Aliases, split[1])
case "interface_name":
if split[1] == "" {
return netOpts, errors.New("interface_name cannot be empty")
}
netOpts.InterfaceName = split[1]
default:
return netOpts, errors.Errorf("unknown bridge network option: %s", split[0])
}
}
return netOpts, nil
}
func SetupUserNS(idmappings *storage.IDMappingOptions, userns Namespace, g *generate.Generator) (string, error) {
// User
var user string
switch userns.NSMode {
case Path:
if _, err := os.Stat(userns.Value); err != nil {
return user, errors.Wrap(err, "cannot find specified user namespace path")
}
if err := g.AddOrReplaceLinuxNamespace(string(spec.UserNamespace), userns.Value); err != nil {
return user, err
}
// runc complains if no mapping is specified, even if we join another ns. So provide a dummy mapping
g.AddLinuxUIDMapping(uint32(0), uint32(0), uint32(1))
g.AddLinuxGIDMapping(uint32(0), uint32(0), uint32(1))
case Host:
if err := g.RemoveLinuxNamespace(string(spec.UserNamespace)); err != nil {
return user, err
}
case KeepID:
mappings, uid, gid, err := util.GetKeepIDMapping()
if err != nil {
return user, err
}
idmappings = mappings
g.SetProcessUID(uint32(uid))
g.SetProcessGID(uint32(gid))
user = fmt.Sprintf("%d:%d", uid, gid)
if err := privateUserNamespace(idmappings, g); err != nil {
return user, err
}
case NoMap:
mappings, uid, gid, err := util.GetNoMapMapping()
if err != nil {
return user, err
}
idmappings = mappings
g.SetProcessUID(uint32(uid))
g.SetProcessGID(uint32(gid))
user = fmt.Sprintf("%d:%d", uid, gid)
if err := privateUserNamespace(idmappings, g); err != nil {
return user, err
}
case Private:
if err := privateUserNamespace(idmappings, g); err != nil {
return user, err
}
}
return user, nil
}
func privateUserNamespace(idmappings *storage.IDMappingOptions, g *generate.Generator) error {
if err := g.AddOrReplaceLinuxNamespace(string(spec.UserNamespace), ""); err != nil {
return err
}
if idmappings == nil || (len(idmappings.UIDMap) == 0 && len(idmappings.GIDMap) == 0) {
return errors.Errorf("must provide at least one UID or GID mapping to configure a user namespace")
}
for _, uidmap := range idmappings.UIDMap {
g.AddLinuxUIDMapping(uint32(uidmap.HostID), uint32(uidmap.ContainerID), uint32(uidmap.Size))
}
for _, gidmap := range idmappings.GIDMap {
g.AddLinuxGIDMapping(uint32(gidmap.HostID), uint32(gidmap.ContainerID), uint32(gidmap.Size))
}
return nil
}