1.. SPDX-License-Identifier: GPL-2.0
2
3The Android binderfs Filesystem
4===============================
5
6Android binderfs is a filesystem for the Android binder IPC mechanism.  It
7allows to dynamically add and remove binder devices at runtime.  Binder devices
8located in a new binderfs instance are independent of binder devices located in
9other binderfs instances.  Mounting a new binderfs instance makes it possible
10to get a set of private binder devices.
11
12Mounting binderfs
13-----------------
14
15Android binderfs can be mounted with::
16
17  mkdir /dev/binderfs
18  mount -t binder binder /dev/binderfs
19
20at which point a new instance of binderfs will show up at ``/dev/binderfs``.
21In a fresh instance of binderfs no binder devices will be present.  There will
22only be a ``binder-control`` device which serves as the request handler for
23binderfs. Mounting another binderfs instance at a different location will
24create a new and separate instance from all other binderfs mounts.  This is
25identical to the behavior of e.g. ``devpts`` and ``tmpfs``. The Android
26binderfs filesystem can be mounted in user namespaces.
27
28Options
29-------
30max
31  binderfs instances can be mounted with a limit on the number of binder
32  devices that can be allocated. The ``max=<count>`` mount option serves as
33  a per-instance limit. If ``max=<count>`` is set then only ``<count>`` number
34  of binder devices can be allocated in this binderfs instance.
35
36stats
37  Using ``stats=global`` enables global binder statistics.
38  ``stats=global`` is only available for a binderfs instance mounted in the
39  initial user namespace. An attempt to use the option to mount a binderfs
40  instance in another user namespace will return a permission error.
41
42Allocating binder Devices
43-------------------------
44
45.. _ioctl: http://man7.org/linux/man-pages/man2/ioctl.2.html
46
47To allocate a new binder device in a binderfs instance a request needs to be
48sent through the ``binder-control`` device node.  A request is sent in the form
49of an `ioctl() <ioctl_>`_.
50
51What a program needs to do is to open the ``binder-control`` device node and
52send a ``BINDER_CTL_ADD`` request to the kernel.  Users of binderfs need to
53tell the kernel which name the new binder device should get.  By default a name
54can only contain up to ``BINDERFS_MAX_NAME`` chars including the terminating
55zero byte.
56
57Once the request is made via an `ioctl() <ioctl_>`_ passing a ``struct
58binder_device`` with the name to the kernel it will allocate a new binder
59device and return the major and minor number of the new device in the struct
60(This is necessary because binderfs allocates a major device number
61dynamically.).  After the `ioctl() <ioctl_>`_ returns there will be a new
62binder device located under /dev/binderfs with the chosen name.
63
64Deleting binder Devices
65-----------------------
66
67.. _unlink: http://man7.org/linux/man-pages/man2/unlink.2.html
68.. _rm: http://man7.org/linux/man-pages/man1/rm.1.html
69
70Binderfs binder devices can be deleted via `unlink() <unlink_>`_.  This means
71that the `rm() <rm_>`_ tool can be used to delete them. Note that the
72``binder-control`` device cannot be deleted since this would make the binderfs
73instance unusable.  The ``binder-control`` device will be deleted when the
74binderfs instance is unmounted and all references to it have been dropped.
75