1 .. SPDX-License-Identifier: GPL-2.0
7 (Taken almost verbatim from Eric Biederman's netns tagging patch
10 The problem. Network devices show up in sysfs and with the network
11 namespace active multiple devices with the same name can show up in
12 the same directory, ouch!
14 To avoid that problem and allow existing applications in network
15 namespaces to see the same interface that is currently presented in
16 sysfs, sysfs now has tagging directory support.
18 By using the network namespace pointers as tags to separate out
19 the sysfs directory entries we ensure that we don't have conflicts
20 in the directories and applications only see a limited set of
23 Each sysfs directory entry may be tagged with a namespace via the
24 ``void *ns member`` of its ``kernfs_node``. If a directory entry is tagged,
25 then ``kernfs_node->flags`` will have a flag between KOBJ_NS_TYPE_NONE
26 and KOBJ_NS_TYPES, and ns will point to the namespace to which it
29 Each sysfs superblock's kernfs_super_info contains an array
30 ``void *ns[KOBJ_NS_TYPES]``. When a task in a tagging namespace
31 kobj_nstype first mounts sysfs, a new superblock is created. It
32 will be differentiated from other sysfs mounts by having its
33 ``s_fs_info->ns[kobj_nstype]`` set to the new namespace. Note that
34 through bind mounting and mounts propagation, a task can easily view
35 the contents of other namespaces' sysfs mounts. Therefore, when a
36 namespace exits, it will call kobj_ns_exit() to invalidate any
37 kernfs_node->ns pointers pointing to it.
39 Users of this interface:
41 - define a type in the ``kobj_ns_type`` enumeration.
42 - call kobj_ns_type_register() with its ``kobj_ns_type_operations`` which has
44 - current_ns() which returns current's namespace
45 - netlink_ns() which returns a socket's namespace
46 - initial_ns() which returns the initial namesapce
48 - call kobj_ns_exit() when an individual tag is no longer valid