1======================
2ioctl based interfaces
3======================
4
5ioctl() is the most common way for applications to interface
6with device drivers. It is flexible and easily extended by adding new
7commands and can be passed through character devices, block devices as
8well as sockets and other special file descriptors.
9
10However, it is also very easy to get ioctl command definitions wrong,
11and hard to fix them later without breaking existing applications,
12so this documentation tries to help developers get it right.
13
14Command number definitions
15==========================
16
17The command number, or request number, is the second argument passed to
18the ioctl system call. While this can be any 32-bit number that uniquely
19identifies an action for a particular driver, there are a number of
20conventions around defining them.
21
22``include/uapi/asm-generic/ioctl.h`` provides four macros for defining
23ioctl commands that follow modern conventions: ``_IO``, ``_IOR``,
24``_IOW``, and ``_IOWR``. These should be used for all new commands,
25with the correct parameters:
26
27_IO/_IOR/_IOW/_IOWR
28   The macro name specifies how the argument will be used.  It may be a
29   pointer to data to be passed into the kernel (_IOW), out of the kernel
30   (_IOR), or both (_IOWR).  _IO can indicate either commands with no
31   argument or those passing an integer value instead of a pointer.
32   It is recommended to only use _IO for commands without arguments,
33   and use pointers for passing data.
34
35type
36   An 8-bit number, often a character literal, specific to a subsystem
37   or driver, and listed in Documentation/userspace-api/ioctl/ioctl-number.rst
38
39nr
40  An 8-bit number identifying the specific command, unique for a give
41  value of 'type'
42
43data_type
44  The name of the data type pointed to by the argument, the command number
45  encodes the ``sizeof(data_type)`` value in a 13-bit or 14-bit integer,
46  leading to a limit of 8191 bytes for the maximum size of the argument.
47  Note: do not pass sizeof(data_type) type into _IOR/_IOW/IOWR, as that
48  will lead to encoding sizeof(sizeof(data_type)), i.e. sizeof(size_t).
49  _IO does not have a data_type parameter.
50
51
52Interface versions
53==================
54
55Some subsystems use version numbers in data structures to overload
56commands with different interpretations of the argument.
57
58This is generally a bad idea, since changes to existing commands tend
59to break existing applications.
60
61A better approach is to add a new ioctl command with a new number. The
62old command still needs to be implemented in the kernel for compatibility,
63but this can be a wrapper around the new implementation.
64
65Return code
66===========
67
68ioctl commands can return negative error codes as documented in errno(3);
69these get turned into errno values in user space. On success, the return
70code should be zero. It is also possible but not recommended to return
71a positive 'long' value.
72
73When the ioctl callback is called with an unknown command number, the
74handler returns either -ENOTTY or -ENOIOCTLCMD, which also results in
75-ENOTTY being returned from the system call. Some subsystems return
76-ENOSYS or -EINVAL here for historic reasons, but this is wrong.
77
78Prior to Linux 5.5, compat_ioctl handlers were required to return
79-ENOIOCTLCMD in order to use the fallback conversion into native
80commands. As all subsystems are now responsible for handling compat
81mode themselves, this is no longer needed, but it may be important to
82consider when backporting bug fixes to older kernels.
83
84Timestamps
85==========
86
87Traditionally, timestamps and timeout values are passed as ``struct
88timespec`` or ``struct timeval``, but these are problematic because of
89incompatible definitions of these structures in user space after the
90move to 64-bit time_t.
91
92The ``struct __kernel_timespec`` type can be used instead to be embedded
93in other data structures when separate second/nanosecond values are
94desired, or passed to user space directly. This is still not ideal though,
95as the structure matches neither the kernel's timespec64 nor the user
96space timespec exactly. The get_timespec64() and put_timespec64() helper
97functions can be used to ensure that the layout remains compatible with
98user space and the padding is treated correctly.
99
100As it is cheap to convert seconds to nanoseconds, but the opposite
101requires an expensive 64-bit division, a simple __u64 nanosecond value
102can be simpler and more efficient.
103
104Timeout values and timestamps should ideally use CLOCK_MONOTONIC time,
105as returned by ktime_get_ns() or ktime_get_ts64().  Unlike
106CLOCK_REALTIME, this makes the timestamps immune from jumping backwards
107or forwards due to leap second adjustments and clock_settime() calls.
108
109ktime_get_real_ns() can be used for CLOCK_REALTIME timestamps that
110need to be persistent across a reboot or between multiple machines.
111
11232-bit compat mode
113==================
114
115In order to support 32-bit user space running on a 64-bit machine, each
116subsystem or driver that implements an ioctl callback handler must also
117implement the corresponding compat_ioctl handler.
118
119As long as all the rules for data structures are followed, this is as
120easy as setting the .compat_ioctl pointer to a helper function such as
121compat_ptr_ioctl() or blkdev_compat_ptr_ioctl().
122
123compat_ptr()
124------------
125
126On the s390 architecture, 31-bit user space has ambiguous representations
127for data pointers, with the upper bit being ignored. When running such
128a process in compat mode, the compat_ptr() helper must be used to
129clear the upper bit of a compat_uptr_t and turn it into a valid 64-bit
130pointer.  On other architectures, this macro only performs a cast to a
131``void __user *`` pointer.
132
133In an compat_ioctl() callback, the last argument is an unsigned long,
134which can be interpreted as either a pointer or a scalar depending on
135the command. If it is a scalar, then compat_ptr() must not be used, to
136ensure that the 64-bit kernel behaves the same way as a 32-bit kernel
137for arguments with the upper bit set.
138
139The compat_ptr_ioctl() helper can be used in place of a custom
140compat_ioctl file operation for drivers that only take arguments that
141are pointers to compatible data structures.
142
143Structure layout
144----------------
145
146Compatible data structures have the same layout on all architectures,
147avoiding all problematic members:
148
149* ``long`` and ``unsigned long`` are the size of a register, so
150  they can be either 32-bit or 64-bit wide and cannot be used in portable
151  data structures. Fixed-length replacements are ``__s32``, ``__u32``,
152  ``__s64`` and ``__u64``.
153
154* Pointers have the same problem, in addition to requiring the
155  use of compat_ptr(). The best workaround is to use ``__u64``
156  in place of pointers, which requires a cast to ``uintptr_t`` in user
157  space, and the use of u64_to_user_ptr() in the kernel to convert
158  it back into a user pointer.
159
160* On the x86-32 (i386) architecture, the alignment of 64-bit variables
161  is only 32-bit, but they are naturally aligned on most other
162  architectures including x86-64. This means a structure like::
163
164    struct foo {
165        __u32 a;
166        __u64 b;
167        __u32 c;
168    };
169
170  has four bytes of padding between a and b on x86-64, plus another four
171  bytes of padding at the end, but no padding on i386, and it needs a
172  compat_ioctl conversion handler to translate between the two formats.
173
174  To avoid this problem, all structures should have their members
175  naturally aligned, or explicit reserved fields added in place of the
176  implicit padding. The ``pahole`` tool can be used for checking the
177  alignment.
178
179* On ARM OABI user space, structures are padded to multiples of 32-bit,
180  making some structs incompatible with modern EABI kernels if they
181  do not end on a 32-bit boundary.
182
183* On the m68k architecture, struct members are not guaranteed to have an
184  alignment greater than 16-bit, which is a problem when relying on
185  implicit padding.
186
187* Bitfields and enums generally work as one would expect them to,
188  but some properties of them are implementation-defined, so it is better
189  to avoid them completely in ioctl interfaces.
190
191* ``char`` members can be either signed or unsigned, depending on
192  the architecture, so the __u8 and __s8 types should be used for 8-bit
193  integer values, though char arrays are clearer for fixed-length strings.
194
195Information leaks
196=================
197
198Uninitialized data must not be copied back to user space, as this can
199cause an information leak, which can be used to defeat kernel address
200space layout randomization (KASLR), helping in an attack.
201
202For this reason (and for compat support) it is best to avoid any
203implicit padding in data structures.  Where there is implicit padding
204in an existing structure, kernel drivers must be careful to fully
205initialize an instance of the structure before copying it to user
206space.  This is usually done by calling memset() before assigning to
207individual members.
208
209Subsystem abstractions
210======================
211
212While some device drivers implement their own ioctl function, most
213subsystems implement the same command for multiple drivers.  Ideally the
214subsystem has an .ioctl() handler that copies the arguments from and
215to user space, passing them into subsystem specific callback functions
216through normal kernel pointers.
217
218This helps in various ways:
219
220* Applications written for one driver are more likely to work for
221  another one in the same subsystem if there are no subtle differences
222  in the user space ABI.
223
224* The complexity of user space access and data structure layout is done
225  in one place, reducing the potential for implementation bugs.
226
227* It is more likely to be reviewed by experienced developers
228  that can spot problems in the interface when the ioctl is shared
229  between multiple drivers than when it is only used in a single driver.
230
231Alternatives to ioctl
232=====================
233
234There are many cases in which ioctl is not the best solution for a
235problem. Alternatives include:
236
237* System calls are a better choice for a system-wide feature that
238  is not tied to a physical device or constrained by the file system
239  permissions of a character device node
240
241* netlink is the preferred way of configuring any network related
242  objects through sockets.
243
244* debugfs is used for ad-hoc interfaces for debugging functionality
245  that does not need to be exposed as a stable interface to applications.
246
247* sysfs is a good way to expose the state of an in-kernel object
248  that is not tied to a file descriptor.
249
250* configfs can be used for more complex configuration than sysfs
251
252* A custom file system can provide extra flexibility with a simple
253  user interface but adds a lot of complexity to the implementation.
254