vbatts/pkg
1
0
Fork 0
Code Pull requests Releases Activity

Mount /proc and /sys read-only, except in privileged containers.

Browse source 
It has been pointed out that some files in /proc and /sys can be used
to break out of containers. However, if those filesystems are mounted
read-only, most of the known exploits are mitigated, since they rely
on writing some file in those filesystems.

This does not replace security modules (like SELinux or AppArmor), it
is just another layer of security. Likewise, it doesn't mean that the
other mitigations (shadowing parts of /proc or /sys with bind mounts)
are useless. Those measures are still useful. As such, the shadowing
of /proc/kcore is still enabled with both LXC and native drivers.

Special care has to be taken with /proc/1/attr, which still needs to
be mounted read-write in order to enable the AppArmor profile. It is
bind-mounted from a private read-write mount of procfs.

All that enforcement is done in dockerinit. The code doing the real
work is in libcontainer. The init function for the LXC driver calls
the function from libcontainer to avoid code duplication.

Docker-DCO-1.1-Signed-off-by: Jérôme Petazzoni <jerome@docker.com> (github: jpetazzo)
This commit is contained in:
Jérôme Petazzoni 2014-04-30 18:00:42 -07:00 • committed by Michael Crosby
parent ecb2b00021
commit a5364236a7
3 changed files with 68 additions and 44 deletions
Download patch file Download diff file Expand all files Collapse all files

12
libcontainer/mount/init.go
View file

@ -11,7 +11,6 @@ import (
"github.com/dotcloud/docker/pkg/label"
"github.com/dotcloud/docker/pkg/libcontainer"
"github.com/dotcloud/docker/pkg/libcontainer/mount/nodes"
"github.com/dotcloud/docker/pkg/libcontainer/security/restrict"
"github.com/dotcloud/docker/pkg/system"
)
@ -51,11 +50,6 @@ func InitializeMountNamespace(rootfs, console string, container *libcontainer.Co
if err := nodes.CopyN(rootfs, nodes.DefaultNodes); err != nil {
return fmt.Errorf("copy dev nodes %s", err)
}
if restrictionPath := container.Context["restriction_path"]; restrictionPath != "" {
if err := restrict.Restrict(rootfs, restrictionPath); err != nil {
return fmt.Errorf("restrict %s", err)
}
}
if err := SetupPtmx(rootfs, console, container.Context["mount_label"]); err != nil {
return err
}
@ -124,10 +118,11 @@ func setupBindmounts(rootfs string, bindMounts libcontainer.Mounts) error {
}
// TODO: this is crappy right now and should be cleaned up with a better way of handling system and
// standard bind mounts allowing them to be more dymanic
// standard bind mounts allowing them to be more dynamic
func newSystemMounts(rootfs, mountLabel string, mounts libcontainer.Mounts) []mount {
systemMounts := []mount{
{source: "proc", path: filepath.Join(rootfs, "proc"), device: "proc", flags: defaultMountFlags},
{source: "sysfs", path: filepath.Join(rootfs, "sys"), device: "sysfs", flags: defaultMountFlags},
}
if len(mounts.OfType("devtmpfs")) == 1 {
@ -138,8 +133,5 @@ func newSystemMounts(rootfs, mountLabel string, mounts libcontainer.Mounts) []mo
mount{source: "devpts", path: filepath.Join(rootfs, "dev", "pts"), device: "devpts", flags: syscall.MS_NOSUID | syscall.MS_NOEXEC, data: label.FormatMountLabel("newinstance,ptmxmode=0666,mode=620,gid=5", mountLabel)},
)
if len(mounts.OfType("sysfs")) == 1 {
systemMounts = append(systemMounts, mount{source: "sysfs", path: filepath.Join(rootfs, "sys"), device: "sysfs", flags: defaultMountFlags})
}
return systemMounts
}