1============================
2Kernel Key Retention Service
3============================
4
5This service allows cryptographic keys, authentication tokens, cross-domain
6user mappings, and similar to be cached in the kernel for the use of
7filesystems and other kernel services.
8
9Keyrings are permitted; these are a special type of key that can hold links to
10other keys. Processes each have three standard keyring subscriptions that a
11kernel service can search for relevant keys.
12
13The key service can be configured on by enabling:
14
15	"Security options"/"Enable access key retention support" (CONFIG_KEYS)
16
17This document has the following sections:
18
19.. contents:: :local:
20
21
22Key Overview
23============
24
25In this context, keys represent units of cryptographic data, authentication
26tokens, keyrings, etc.. These are represented in the kernel by struct key.
27
28Each key has a number of attributes:
29
30	- A serial number.
31	- A type.
32	- A description (for matching a key in a search).
33	- Access control information.
34	- An expiry time.
35	- A payload.
36	- State.
37
38
39  *  Each key is issued a serial number of type key_serial_t that is unique for
40     the lifetime of that key. All serial numbers are positive non-zero 32-bit
41     integers.
42
43     Userspace programs can use a key's serial numbers as a way to gain access
44     to it, subject to permission checking.
45
46  *  Each key is of a defined "type". Types must be registered inside the
47     kernel by a kernel service (such as a filesystem) before keys of that type
48     can be added or used. Userspace programs cannot define new types directly.
49
50     Key types are represented in the kernel by struct key_type. This defines a
51     number of operations that can be performed on a key of that type.
52
53     Should a type be removed from the system, all the keys of that type will
54     be invalidated.
55
56  *  Each key has a description. This should be a printable string. The key
57     type provides an operation to perform a match between the description on a
58     key and a criterion string.
59
60  *  Each key has an owner user ID, a group ID and a permissions mask. These
61     are used to control what a process may do to a key from userspace, and
62     whether a kernel service will be able to find the key.
63
64  *  Each key can be set to expire at a specific time by the key type's
65     instantiation function. Keys can also be immortal.
66
67  *  Each key can have a payload. This is a quantity of data that represent the
68     actual "key". In the case of a keyring, this is a list of keys to which
69     the keyring links; in the case of a user-defined key, it's an arbitrary
70     blob of data.
71
72     Having a payload is not required; and the payload can, in fact, just be a
73     value stored in the struct key itself.
74
75     When a key is instantiated, the key type's instantiation function is
76     called with a blob of data, and that then creates the key's payload in
77     some way.
78
79     Similarly, when userspace wants to read back the contents of the key, if
80     permitted, another key type operation will be called to convert the key's
81     attached payload back into a blob of data.
82
83  *  Each key can be in one of a number of basic states:
84
85      *  Uninstantiated. The key exists, but does not have any data attached.
86     	 Keys being requested from userspace will be in this state.
87
88      *  Instantiated. This is the normal state. The key is fully formed, and
89	 has data attached.
90
91      *  Negative. This is a relatively short-lived state. The key acts as a
92	 note saying that a previous call out to userspace failed, and acts as
93	 a throttle on key lookups. A negative key can be updated to a normal
94	 state.
95
96      *  Expired. Keys can have lifetimes set. If their lifetime is exceeded,
97	 they traverse to this state. An expired key can be updated back to a
98	 normal state.
99
100      *  Revoked. A key is put in this state by userspace action. It can't be
101	 found or operated upon (apart from by unlinking it).
102
103      *  Dead. The key's type was unregistered, and so the key is now useless.
104
105Keys in the last three states are subject to garbage collection.  See the
106section on "Garbage collection".
107
108
109Key Service Overview
110====================
111
112The key service provides a number of features besides keys:
113
114  *  The key service defines three special key types:
115
116     (+) "keyring"
117
118	 Keyrings are special keys that contain a list of other keys. Keyring
119	 lists can be modified using various system calls. Keyrings should not
120	 be given a payload when created.
121
122     (+) "user"
123
124	 A key of this type has a description and a payload that are arbitrary
125	 blobs of data. These can be created, updated and read by userspace,
126	 and aren't intended for use by kernel services.
127
128     (+) "logon"
129
130	 Like a "user" key, a "logon" key has a payload that is an arbitrary
131	 blob of data. It is intended as a place to store secrets which are
132	 accessible to the kernel but not to userspace programs.
133
134	 The description can be arbitrary, but must be prefixed with a non-zero
135	 length string that describes the key "subclass". The subclass is
136	 separated from the rest of the description by a ':'. "logon" keys can
137	 be created and updated from userspace, but the payload is only
138	 readable from kernel space.
139
140  *  Each process subscribes to three keyrings: a thread-specific keyring, a
141     process-specific keyring, and a session-specific keyring.
142
143     The thread-specific keyring is discarded from the child when any sort of
144     clone, fork, vfork or execve occurs. A new keyring is created only when
145     required.
146
147     The process-specific keyring is replaced with an empty one in the child on
148     clone, fork, vfork unless CLONE_THREAD is supplied, in which case it is
149     shared. execve also discards the process's process keyring and creates a
150     new one.
151
152     The session-specific keyring is persistent across clone, fork, vfork and
153     execve, even when the latter executes a set-UID or set-GID binary. A
154     process can, however, replace its current session keyring with a new one
155     by using PR_JOIN_SESSION_KEYRING. It is permitted to request an anonymous
156     new one, or to attempt to create or join one of a specific name.
157
158     The ownership of the thread keyring changes when the real UID and GID of
159     the thread changes.
160
161  *  Each user ID resident in the system holds two special keyrings: a user
162     specific keyring and a default user session keyring. The default session
163     keyring is initialised with a link to the user-specific keyring.
164
165     When a process changes its real UID, if it used to have no session key, it
166     will be subscribed to the default session key for the new UID.
167
168     If a process attempts to access its session key when it doesn't have one,
169     it will be subscribed to the default for its current UID.
170
171  *  Each user has two quotas against which the keys they own are tracked. One
172     limits the total number of keys and keyrings, the other limits the total
173     amount of description and payload space that can be consumed.
174
175     The user can view information on this and other statistics through procfs
176     files.  The root user may also alter the quota limits through sysctl files
177     (see the section "New procfs files").
178
179     Process-specific and thread-specific keyrings are not counted towards a
180     user's quota.
181
182     If a system call that modifies a key or keyring in some way would put the
183     user over quota, the operation is refused and error EDQUOT is returned.
184
185  *  There's a system call interface by which userspace programs can create and
186     manipulate keys and keyrings.
187
188  *  There's a kernel interface by which services can register types and search
189     for keys.
190
191  *  There's a way for the a search done from the kernel to call back to
192     userspace to request a key that can't be found in a process's keyrings.
193
194  *  An optional filesystem is available through which the key database can be
195     viewed and manipulated.
196
197
198Key Access Permissions
199======================
200
201Keys have an owner user ID, a group access ID, and a permissions mask. The mask
202has up to eight bits each for possessor, user, group and other access. Only
203six of each set of eight bits are defined. These permissions granted are:
204
205  *  View
206
207     This permits a key or keyring's attributes to be viewed - including key
208     type and description.
209
210  *  Read
211
212     This permits a key's payload to be viewed or a keyring's list of linked
213     keys.
214
215  *  Write
216
217     This permits a key's payload to be instantiated or updated, or it allows a
218     link to be added to or removed from a keyring.
219
220  *  Search
221
222     This permits keyrings to be searched and keys to be found. Searches can
223     only recurse into nested keyrings that have search permission set.
224
225  *  Link
226
227     This permits a key or keyring to be linked to. To create a link from a
228     keyring to a key, a process must have Write permission on the keyring and
229     Link permission on the key.
230
231  *  Set Attribute
232
233     This permits a key's UID, GID and permissions mask to be changed.
234
235For changing the ownership, group ID or permissions mask, being the owner of
236the key or having the sysadmin capability is sufficient.
237
238
239SELinux Support
240===============
241
242The security class "key" has been added to SELinux so that mandatory access
243controls can be applied to keys created within various contexts.  This support
244is preliminary, and is likely to change quite significantly in the near future.
245Currently, all of the basic permissions explained above are provided in SELinux
246as well; SELinux is simply invoked after all basic permission checks have been
247performed.
248
249The value of the file /proc/self/attr/keycreate influences the labeling of
250newly-created keys.  If the contents of that file correspond to an SELinux
251security context, then the key will be assigned that context.  Otherwise, the
252key will be assigned the current context of the task that invoked the key
253creation request.  Tasks must be granted explicit permission to assign a
254particular context to newly-created keys, using the "create" permission in the
255key security class.
256
257The default keyrings associated with users will be labeled with the default
258context of the user if and only if the login programs have been instrumented to
259properly initialize keycreate during the login process.  Otherwise, they will
260be labeled with the context of the login program itself.
261
262Note, however, that the default keyrings associated with the root user are
263labeled with the default kernel context, since they are created early in the
264boot process, before root has a chance to log in.
265
266The keyrings associated with new threads are each labeled with the context of
267their associated thread, and both session and process keyrings are handled
268similarly.
269
270
271New ProcFS Files
272================
273
274Two files have been added to procfs by which an administrator can find out
275about the status of the key service:
276
277  *  /proc/keys
278
279     This lists the keys that are currently viewable by the task reading the
280     file, giving information about their type, description and permissions.
281     It is not possible to view the payload of the key this way, though some
282     information about it may be given.
283
284     The only keys included in the list are those that grant View permission to
285     the reading process whether or not it possesses them.  Note that LSM
286     security checks are still performed, and may further filter out keys that
287     the current process is not authorised to view.
288
289     The contents of the file look like this::
290
291	SERIAL   FLAGS  USAGE EXPY PERM     UID   GID   TYPE      DESCRIPTION: SUMMARY
292	00000001 I-----    39 perm 1f3f0000     0     0 keyring   _uid_ses.0: 1/4
293	00000002 I-----     2 perm 1f3f0000     0     0 keyring   _uid.0: empty
294	00000007 I-----     1 perm 1f3f0000     0     0 keyring   _pid.1: empty
295	0000018d I-----     1 perm 1f3f0000     0     0 keyring   _pid.412: empty
296	000004d2 I--Q--     1 perm 1f3f0000    32    -1 keyring   _uid.32: 1/4
297	000004d3 I--Q--     3 perm 1f3f0000    32    -1 keyring   _uid_ses.32: empty
298	00000892 I--QU-     1 perm 1f000000     0     0 user      metal:copper: 0
299	00000893 I--Q-N     1  35s 1f3f0000     0     0 user      metal:silver: 0
300	00000894 I--Q--     1  10h 003f0000     0     0 user      metal:gold: 0
301
302     The flags are::
303
304	I	Instantiated
305	R	Revoked
306	D	Dead
307	Q	Contributes to user's quota
308	U	Under construction by callback to userspace
309	N	Negative key
310
311
312  *  /proc/key-users
313
314     This file lists the tracking data for each user that has at least one key
315     on the system.  Such data includes quota information and statistics::
316
317	[root@andromeda root]# cat /proc/key-users
318	0:     46 45/45 1/100 13/10000
319	29:     2 2/2 2/100 40/10000
320	32:     2 2/2 2/100 40/10000
321	38:     2 2/2 2/100 40/10000
322
323     The format of each line is::
324
325	<UID>:			User ID to which this applies
326	<usage>			Structure refcount
327	<inst>/<keys>		Total number of keys and number instantiated
328	<keys>/<max>		Key count quota
329	<bytes>/<max>		Key size quota
330
331
332Four new sysctl files have been added also for the purpose of controlling the
333quota limits on keys:
334
335  *  /proc/sys/kernel/keys/root_maxkeys
336     /proc/sys/kernel/keys/root_maxbytes
337
338     These files hold the maximum number of keys that root may have and the
339     maximum total number of bytes of data that root may have stored in those
340     keys.
341
342  *  /proc/sys/kernel/keys/maxkeys
343     /proc/sys/kernel/keys/maxbytes
344
345     These files hold the maximum number of keys that each non-root user may
346     have and the maximum total number of bytes of data that each of those
347     users may have stored in their keys.
348
349Root may alter these by writing each new limit as a decimal number string to
350the appropriate file.
351
352
353Userspace System Call Interface
354===============================
355
356Userspace can manipulate keys directly through three new syscalls: add_key,
357request_key and keyctl. The latter provides a number of functions for
358manipulating keys.
359
360When referring to a key directly, userspace programs should use the key's
361serial number (a positive 32-bit integer). However, there are some special
362values available for referring to special keys and keyrings that relate to the
363process making the call::
364
365	CONSTANT			VALUE	KEY REFERENCED
366	==============================	======	===========================
367	KEY_SPEC_THREAD_KEYRING		-1	thread-specific keyring
368	KEY_SPEC_PROCESS_KEYRING	-2	process-specific keyring
369	KEY_SPEC_SESSION_KEYRING	-3	session-specific keyring
370	KEY_SPEC_USER_KEYRING		-4	UID-specific keyring
371	KEY_SPEC_USER_SESSION_KEYRING	-5	UID-session keyring
372	KEY_SPEC_GROUP_KEYRING		-6	GID-specific keyring
373	KEY_SPEC_REQKEY_AUTH_KEY	-7	assumed request_key()
374						  authorisation key
375
376
377The main syscalls are:
378
379  *  Create a new key of given type, description and payload and add it to the
380     nominated keyring::
381
382	key_serial_t add_key(const char *type, const char *desc,
383			     const void *payload, size_t plen,
384			     key_serial_t keyring);
385
386     If a key of the same type and description as that proposed already exists
387     in the keyring, this will try to update it with the given payload, or it
388     will return error EEXIST if that function is not supported by the key
389     type. The process must also have permission to write to the key to be able
390     to update it. The new key will have all user permissions granted and no
391     group or third party permissions.
392
393     Otherwise, this will attempt to create a new key of the specified type and
394     description, and to instantiate it with the supplied payload and attach it
395     to the keyring. In this case, an error will be generated if the process
396     does not have permission to write to the keyring.
397
398     If the key type supports it, if the description is NULL or an empty
399     string, the key type will try and generate a description from the content
400     of the payload.
401
402     The payload is optional, and the pointer can be NULL if not required by
403     the type. The payload is plen in size, and plen can be zero for an empty
404     payload.
405
406     A new keyring can be generated by setting type "keyring", the keyring name
407     as the description (or NULL) and setting the payload to NULL.
408
409     User defined keys can be created by specifying type "user". It is
410     recommended that a user defined key's description by prefixed with a type
411     ID and a colon, such as "krb5tgt:" for a Kerberos 5 ticket granting
412     ticket.
413
414     Any other type must have been registered with the kernel in advance by a
415     kernel service such as a filesystem.
416
417     The ID of the new or updated key is returned if successful.
418
419
420  *  Search the process's keyrings for a key, potentially calling out to
421     userspace to create it::
422
423	key_serial_t request_key(const char *type, const char *description,
424				 const char *callout_info,
425				 key_serial_t dest_keyring);
426
427     This function searches all the process's keyrings in the order thread,
428     process, session for a matching key. This works very much like
429     KEYCTL_SEARCH, including the optional attachment of the discovered key to
430     a keyring.
431
432     If a key cannot be found, and if callout_info is not NULL, then
433     /sbin/request-key will be invoked in an attempt to obtain a key. The
434     callout_info string will be passed as an argument to the program.
435
436     To link a key into the destination keyring the key must grant link
437     permission on the key to the caller and the keyring must grant write
438     permission.
439
440     See also Documentation/security/keys/request-key.rst.
441
442
443The keyctl syscall functions are:
444
445  *  Map a special key ID to a real key ID for this process::
446
447	key_serial_t keyctl(KEYCTL_GET_KEYRING_ID, key_serial_t id,
448			    int create);
449
450     The special key specified by "id" is looked up (with the key being created
451     if necessary) and the ID of the key or keyring thus found is returned if
452     it exists.
453
454     If the key does not yet exist, the key will be created if "create" is
455     non-zero; and the error ENOKEY will be returned if "create" is zero.
456
457
458  *  Replace the session keyring this process subscribes to with a new one::
459
460	key_serial_t keyctl(KEYCTL_JOIN_SESSION_KEYRING, const char *name);
461
462     If name is NULL, an anonymous keyring is created attached to the process
463     as its session keyring, displacing the old session keyring.
464
465     If name is not NULL, if a keyring of that name exists, the process
466     attempts to attach it as the session keyring, returning an error if that
467     is not permitted; otherwise a new keyring of that name is created and
468     attached as the session keyring.
469
470     To attach to a named keyring, the keyring must have search permission for
471     the process's ownership.
472
473     The ID of the new session keyring is returned if successful.
474
475
476  *  Update the specified key::
477
478	long keyctl(KEYCTL_UPDATE, key_serial_t key, const void *payload,
479		    size_t plen);
480
481     This will try to update the specified key with the given payload, or it
482     will return error EOPNOTSUPP if that function is not supported by the key
483     type. The process must also have permission to write to the key to be able
484     to update it.
485
486     The payload is of length plen, and may be absent or empty as for
487     add_key().
488
489
490  *  Revoke a key::
491
492	long keyctl(KEYCTL_REVOKE, key_serial_t key);
493
494     This makes a key unavailable for further operations. Further attempts to
495     use the key will be met with error EKEYREVOKED, and the key will no longer
496     be findable.
497
498
499  *  Change the ownership of a key::
500
501	long keyctl(KEYCTL_CHOWN, key_serial_t key, uid_t uid, gid_t gid);
502
503     This function permits a key's owner and group ID to be changed. Either one
504     of uid or gid can be set to -1 to suppress that change.
505
506     Only the superuser can change a key's owner to something other than the
507     key's current owner. Similarly, only the superuser can change a key's
508     group ID to something other than the calling process's group ID or one of
509     its group list members.
510
511
512  *  Change the permissions mask on a key::
513
514	long keyctl(KEYCTL_SETPERM, key_serial_t key, key_perm_t perm);
515
516     This function permits the owner of a key or the superuser to change the
517     permissions mask on a key.
518
519     Only bits the available bits are permitted; if any other bits are set,
520     error EINVAL will be returned.
521
522
523  *  Describe a key::
524
525	long keyctl(KEYCTL_DESCRIBE, key_serial_t key, char *buffer,
526		    size_t buflen);
527
528     This function returns a summary of the key's attributes (but not its
529     payload data) as a string in the buffer provided.
530
531     Unless there's an error, it always returns the amount of data it could
532     produce, even if that's too big for the buffer, but it won't copy more
533     than requested to userspace. If the buffer pointer is NULL then no copy
534     will take place.
535
536     A process must have view permission on the key for this function to be
537     successful.
538
539     If successful, a string is placed in the buffer in the following format::
540
541	<type>;<uid>;<gid>;<perm>;<description>
542
543     Where type and description are strings, uid and gid are decimal, and perm
544     is hexadecimal. A NUL character is included at the end of the string if
545     the buffer is sufficiently big.
546
547     This can be parsed with::
548
549	sscanf(buffer, "%[^;];%d;%d;%o;%s", type, &uid, &gid, &mode, desc);
550
551
552  *  Clear out a keyring::
553
554	long keyctl(KEYCTL_CLEAR, key_serial_t keyring);
555
556     This function clears the list of keys attached to a keyring. The calling
557     process must have write permission on the keyring, and it must be a
558     keyring (or else error ENOTDIR will result).
559
560     This function can also be used to clear special kernel keyrings if they
561     are appropriately marked if the user has CAP_SYS_ADMIN capability.  The
562     DNS resolver cache keyring is an example of this.
563
564
565  *  Link a key into a keyring::
566
567	long keyctl(KEYCTL_LINK, key_serial_t keyring, key_serial_t key);
568
569     This function creates a link from the keyring to the key. The process must
570     have write permission on the keyring and must have link permission on the
571     key.
572
573     Should the keyring not be a keyring, error ENOTDIR will result; and if the
574     keyring is full, error ENFILE will result.
575
576     The link procedure checks the nesting of the keyrings, returning ELOOP if
577     it appears too deep or EDEADLK if the link would introduce a cycle.
578
579     Any links within the keyring to keys that match the new key in terms of
580     type and description will be discarded from the keyring as the new one is
581     added.
582
583
584  *  Move a key from one keyring to another::
585
586	long keyctl(KEYCTL_MOVE,
587		    key_serial_t id,
588		    key_serial_t from_ring_id,
589		    key_serial_t to_ring_id,
590		    unsigned int flags);
591
592     Move the key specified by "id" from the keyring specified by
593     "from_ring_id" to the keyring specified by "to_ring_id".  If the two
594     keyrings are the same, nothing is done.
595
596     "flags" can have KEYCTL_MOVE_EXCL set in it to cause the operation to fail
597     with EEXIST if a matching key exists in the destination keyring, otherwise
598     such a key will be replaced.
599
600     A process must have link permission on the key for this function to be
601     successful and write permission on both keyrings.  Any errors that can
602     occur from KEYCTL_LINK also apply on the destination keyring here.
603
604
605  *  Unlink a key or keyring from another keyring::
606
607	long keyctl(KEYCTL_UNLINK, key_serial_t keyring, key_serial_t key);
608
609     This function looks through the keyring for the first link to the
610     specified key, and removes it if found. Subsequent links to that key are
611     ignored. The process must have write permission on the keyring.
612
613     If the keyring is not a keyring, error ENOTDIR will result; and if the key
614     is not present, error ENOENT will be the result.
615
616
617  *  Search a keyring tree for a key::
618
619	key_serial_t keyctl(KEYCTL_SEARCH, key_serial_t keyring,
620			    const char *type, const char *description,
621			    key_serial_t dest_keyring);
622
623     This searches the keyring tree headed by the specified keyring until a key
624     is found that matches the type and description criteria. Each keyring is
625     checked for keys before recursion into its children occurs.
626
627     The process must have search permission on the top level keyring, or else
628     error EACCES will result. Only keyrings that the process has search
629     permission on will be recursed into, and only keys and keyrings for which
630     a process has search permission can be matched. If the specified keyring
631     is not a keyring, ENOTDIR will result.
632
633     If the search succeeds, the function will attempt to link the found key
634     into the destination keyring if one is supplied (non-zero ID). All the
635     constraints applicable to KEYCTL_LINK apply in this case too.
636
637     Error ENOKEY, EKEYREVOKED or EKEYEXPIRED will be returned if the search
638     fails. On success, the resulting key ID will be returned.
639
640
641  *  Read the payload data from a key::
642
643	long keyctl(KEYCTL_READ, key_serial_t keyring, char *buffer,
644		    size_t buflen);
645
646     This function attempts to read the payload data from the specified key
647     into the buffer. The process must have read permission on the key to
648     succeed.
649
650     The returned data will be processed for presentation by the key type. For
651     instance, a keyring will return an array of key_serial_t entries
652     representing the IDs of all the keys to which it is subscribed. The user
653     defined key type will return its data as is. If a key type does not
654     implement this function, error EOPNOTSUPP will result.
655
656     If the specified buffer is too small, then the size of the buffer required
657     will be returned.  Note that in this case, the contents of the buffer may
658     have been overwritten in some undefined way.
659
660     Otherwise, on success, the function will return the amount of data copied
661     into the buffer.
662
663  *  Instantiate a partially constructed key::
664
665	long keyctl(KEYCTL_INSTANTIATE, key_serial_t key,
666		    const void *payload, size_t plen,
667		    key_serial_t keyring);
668	long keyctl(KEYCTL_INSTANTIATE_IOV, key_serial_t key,
669		    const struct iovec *payload_iov, unsigned ioc,
670		    key_serial_t keyring);
671
672     If the kernel calls back to userspace to complete the instantiation of a
673     key, userspace should use this call to supply data for the key before the
674     invoked process returns, or else the key will be marked negative
675     automatically.
676
677     The process must have write access on the key to be able to instantiate
678     it, and the key must be uninstantiated.
679
680     If a keyring is specified (non-zero), the key will also be linked into
681     that keyring, however all the constraints applying in KEYCTL_LINK apply in
682     this case too.
683
684     The payload and plen arguments describe the payload data as for add_key().
685
686     The payload_iov and ioc arguments describe the payload data in an iovec
687     array instead of a single buffer.
688
689
690  *  Negatively instantiate a partially constructed key::
691
692	long keyctl(KEYCTL_NEGATE, key_serial_t key,
693		    unsigned timeout, key_serial_t keyring);
694	long keyctl(KEYCTL_REJECT, key_serial_t key,
695		    unsigned timeout, unsigned error, key_serial_t keyring);
696
697     If the kernel calls back to userspace to complete the instantiation of a
698     key, userspace should use this call mark the key as negative before the
699     invoked process returns if it is unable to fulfill the request.
700
701     The process must have write access on the key to be able to instantiate
702     it, and the key must be uninstantiated.
703
704     If a keyring is specified (non-zero), the key will also be linked into
705     that keyring, however all the constraints applying in KEYCTL_LINK apply in
706     this case too.
707
708     If the key is rejected, future searches for it will return the specified
709     error code until the rejected key expires.  Negating the key is the same
710     as rejecting the key with ENOKEY as the error code.
711
712
713  *  Set the default request-key destination keyring::
714
715	long keyctl(KEYCTL_SET_REQKEY_KEYRING, int reqkey_defl);
716
717     This sets the default keyring to which implicitly requested keys will be
718     attached for this thread. reqkey_defl should be one of these constants::
719
720	CONSTANT				VALUE	NEW DEFAULT KEYRING
721	======================================	======	=======================
722	KEY_REQKEY_DEFL_NO_CHANGE		-1	No change
723	KEY_REQKEY_DEFL_DEFAULT			0	Default[1]
724	KEY_REQKEY_DEFL_THREAD_KEYRING		1	Thread keyring
725	KEY_REQKEY_DEFL_PROCESS_KEYRING		2	Process keyring
726	KEY_REQKEY_DEFL_SESSION_KEYRING		3	Session keyring
727	KEY_REQKEY_DEFL_USER_KEYRING		4	User keyring
728	KEY_REQKEY_DEFL_USER_SESSION_KEYRING	5	User session keyring
729	KEY_REQKEY_DEFL_GROUP_KEYRING		6	Group keyring
730
731     The old default will be returned if successful and error EINVAL will be
732     returned if reqkey_defl is not one of the above values.
733
734     The default keyring can be overridden by the keyring indicated to the
735     request_key() system call.
736
737     Note that this setting is inherited across fork/exec.
738
739     [1] The default is: the thread keyring if there is one, otherwise
740     the process keyring if there is one, otherwise the session keyring if
741     there is one, otherwise the user default session keyring.
742
743
744  *  Set the timeout on a key::
745
746	long keyctl(KEYCTL_SET_TIMEOUT, key_serial_t key, unsigned timeout);
747
748     This sets or clears the timeout on a key. The timeout can be 0 to clear
749     the timeout or a number of seconds to set the expiry time that far into
750     the future.
751
752     The process must have attribute modification access on a key to set its
753     timeout. Timeouts may not be set with this function on negative, revoked
754     or expired keys.
755
756
757  *  Assume the authority granted to instantiate a key::
758
759	long keyctl(KEYCTL_ASSUME_AUTHORITY, key_serial_t key);
760
761     This assumes or divests the authority required to instantiate the
762     specified key. Authority can only be assumed if the thread has the
763     authorisation key associated with the specified key in its keyrings
764     somewhere.
765
766     Once authority is assumed, searches for keys will also search the
767     requester's keyrings using the requester's security label, UID, GID and
768     groups.
769
770     If the requested authority is unavailable, error EPERM will be returned,
771     likewise if the authority has been revoked because the target key is
772     already instantiated.
773
774     If the specified key is 0, then any assumed authority will be divested.
775
776     The assumed authoritative key is inherited across fork and exec.
777
778
779  *  Get the LSM security context attached to a key::
780
781	long keyctl(KEYCTL_GET_SECURITY, key_serial_t key, char *buffer,
782		    size_t buflen)
783
784     This function returns a string that represents the LSM security context
785     attached to a key in the buffer provided.
786
787     Unless there's an error, it always returns the amount of data it could
788     produce, even if that's too big for the buffer, but it won't copy more
789     than requested to userspace. If the buffer pointer is NULL then no copy
790     will take place.
791
792     A NUL character is included at the end of the string if the buffer is
793     sufficiently big.  This is included in the returned count.  If no LSM is
794     in force then an empty string will be returned.
795
796     A process must have view permission on the key for this function to be
797     successful.
798
799
800  *  Install the calling process's session keyring on its parent::
801
802	long keyctl(KEYCTL_SESSION_TO_PARENT);
803
804     This functions attempts to install the calling process's session keyring
805     on to the calling process's parent, replacing the parent's current session
806     keyring.
807
808     The calling process must have the same ownership as its parent, the
809     keyring must have the same ownership as the calling process, the calling
810     process must have LINK permission on the keyring and the active LSM module
811     mustn't deny permission, otherwise error EPERM will be returned.
812
813     Error ENOMEM will be returned if there was insufficient memory to complete
814     the operation, otherwise 0 will be returned to indicate success.
815
816     The keyring will be replaced next time the parent process leaves the
817     kernel and resumes executing userspace.
818
819
820  *  Invalidate a key::
821
822	long keyctl(KEYCTL_INVALIDATE, key_serial_t key);
823
824     This function marks a key as being invalidated and then wakes up the
825     garbage collector.  The garbage collector immediately removes invalidated
826     keys from all keyrings and deletes the key when its reference count
827     reaches zero.
828
829     Keys that are marked invalidated become invisible to normal key operations
830     immediately, though they are still visible in /proc/keys until deleted
831     (they're marked with an 'i' flag).
832
833     A process must have search permission on the key for this function to be
834     successful.
835
836  *  Compute a Diffie-Hellman shared secret or public key::
837
838	long keyctl(KEYCTL_DH_COMPUTE, struct keyctl_dh_params *params,
839		    char *buffer, size_t buflen, struct keyctl_kdf_params *kdf);
840
841     The params struct contains serial numbers for three keys::
842
843	 - The prime, p, known to both parties
844	 - The local private key
845	 - The base integer, which is either a shared generator or the
846	   remote public key
847
848     The value computed is::
849
850	result = base ^ private (mod prime)
851
852     If the base is the shared generator, the result is the local
853     public key.  If the base is the remote public key, the result is
854     the shared secret.
855
856     If the parameter kdf is NULL, the following applies:
857
858	 - The buffer length must be at least the length of the prime, or zero.
859
860	 - If the buffer length is nonzero, the length of the result is
861	   returned when it is successfully calculated and copied in to the
862	   buffer. When the buffer length is zero, the minimum required
863	   buffer length is returned.
864
865     The kdf parameter allows the caller to apply a key derivation function
866     (KDF) on the Diffie-Hellman computation where only the result
867     of the KDF is returned to the caller. The KDF is characterized with
868     struct keyctl_kdf_params as follows:
869
870	 - ``char *hashname`` specifies the NUL terminated string identifying
871	   the hash used from the kernel crypto API and applied for the KDF
872	   operation. The KDF implemenation complies with SP800-56A as well
873	   as with SP800-108 (the counter KDF).
874
875	 - ``char *otherinfo`` specifies the OtherInfo data as documented in
876	   SP800-56A section 5.8.1.2. The length of the buffer is given with
877	   otherinfolen. The format of OtherInfo is defined by the caller.
878	   The otherinfo pointer may be NULL if no OtherInfo shall be used.
879
880     This function will return error EOPNOTSUPP if the key type is not
881     supported, error ENOKEY if the key could not be found, or error
882     EACCES if the key is not readable by the caller. In addition, the
883     function will return EMSGSIZE when the parameter kdf is non-NULL
884     and either the buffer length or the OtherInfo length exceeds the
885     allowed length.
886
887
888  *  Restrict keyring linkage::
889
890	long keyctl(KEYCTL_RESTRICT_KEYRING, key_serial_t keyring,
891		    const char *type, const char *restriction);
892
893     An existing keyring can restrict linkage of additional keys by evaluating
894     the contents of the key according to a restriction scheme.
895
896     "keyring" is the key ID for an existing keyring to apply a restriction
897     to. It may be empty or may already have keys linked. Existing linked keys
898     will remain in the keyring even if the new restriction would reject them.
899
900     "type" is a registered key type.
901
902     "restriction" is a string describing how key linkage is to be restricted.
903     The format varies depending on the key type, and the string is passed to
904     the lookup_restriction() function for the requested type.  It may specify
905     a method and relevant data for the restriction such as signature
906     verification or constraints on key payload. If the requested key type is
907     later unregistered, no keys may be added to the keyring after the key type
908     is removed.
909
910     To apply a keyring restriction the process must have Set Attribute
911     permission and the keyring must not be previously restricted.
912
913     One application of restricted keyrings is to verify X.509 certificate
914     chains or individual certificate signatures using the asymmetric key type.
915     See Documentation/crypto/asymmetric-keys.rst for specific restrictions
916     applicable to the asymmetric key type.
917
918
919  *  Query an asymmetric key::
920
921	long keyctl(KEYCTL_PKEY_QUERY,
922		    key_serial_t key_id, unsigned long reserved,
923		    const char *params,
924		    struct keyctl_pkey_query *info);
925
926     Get information about an asymmetric key.  Specific algorithms and
927     encodings may be queried by using the ``params`` argument.  This is a
928     string containing a space- or tab-separated string of key-value pairs.
929     Currently supported keys include ``enc`` and ``hash``.  The information
930     is returned in the keyctl_pkey_query struct::
931
932	__u32	supported_ops;
933	__u32	key_size;
934	__u16	max_data_size;
935	__u16	max_sig_size;
936	__u16	max_enc_size;
937	__u16	max_dec_size;
938	__u32	__spare[10];
939
940     ``supported_ops`` contains a bit mask of flags indicating which ops are
941     supported.  This is constructed from a bitwise-OR of::
942
943	KEYCTL_SUPPORTS_{ENCRYPT,DECRYPT,SIGN,VERIFY}
944
945     ``key_size`` indicated the size of the key in bits.
946
947     ``max_*_size`` indicate the maximum sizes in bytes of a blob of data to be
948     signed, a signature blob, a blob to be encrypted and a blob to be
949     decrypted.
950
951     ``__spare[]`` must be set to 0.  This is intended for future use to hand
952     over one or more passphrases needed unlock a key.
953
954     If successful, 0 is returned.  If the key is not an asymmetric key,
955     EOPNOTSUPP is returned.
956
957
958  *  Encrypt, decrypt, sign or verify a blob using an asymmetric key::
959
960	long keyctl(KEYCTL_PKEY_ENCRYPT,
961		    const struct keyctl_pkey_params *params,
962		    const char *info,
963		    const void *in,
964		    void *out);
965
966	long keyctl(KEYCTL_PKEY_DECRYPT,
967		    const struct keyctl_pkey_params *params,
968		    const char *info,
969		    const void *in,
970		    void *out);
971
972	long keyctl(KEYCTL_PKEY_SIGN,
973		    const struct keyctl_pkey_params *params,
974		    const char *info,
975		    const void *in,
976		    void *out);
977
978	long keyctl(KEYCTL_PKEY_VERIFY,
979		    const struct keyctl_pkey_params *params,
980		    const char *info,
981		    const void *in,
982		    const void *in2);
983
984     Use an asymmetric key to perform a public-key cryptographic operation a
985     blob of data.  For encryption and verification, the asymmetric key may
986     only need the public parts to be available, but for decryption and signing
987     the private parts are required also.
988
989     The parameter block pointed to by params contains a number of integer
990     values::
991
992	__s32		key_id;
993	__u32		in_len;
994	__u32		out_len;
995	__u32		in2_len;
996
997     ``key_id`` is the ID of the asymmetric key to be used.  ``in_len`` and
998     ``in2_len`` indicate the amount of data in the in and in2 buffers and
999     ``out_len`` indicates the size of the out buffer as appropriate for the
1000     above operations.
1001
1002     For a given operation, the in and out buffers are used as follows::
1003
1004	Operation ID		in,in_len	out,out_len	in2,in2_len
1005	=======================	===============	===============	===============
1006	KEYCTL_PKEY_ENCRYPT	Raw data	Encrypted data	-
1007	KEYCTL_PKEY_DECRYPT	Encrypted data	Raw data	-
1008	KEYCTL_PKEY_SIGN	Raw data	Signature	-
1009	KEYCTL_PKEY_VERIFY	Raw data	-		Signature
1010
1011     ``info`` is a string of key=value pairs that supply supplementary
1012     information.  These include:
1013
1014	``enc=<encoding>`` The encoding of the encrypted/signature blob.  This
1015			can be "pkcs1" for RSASSA-PKCS1-v1.5 or
1016			RSAES-PKCS1-v1.5; "pss" for "RSASSA-PSS"; "oaep" for
1017			"RSAES-OAEP".  If omitted or is "raw", the raw output
1018			of the encryption function is specified.
1019
1020	``hash=<algo>``	If the data buffer contains the output of a hash
1021			function and the encoding includes some indication of
1022			which hash function was used, the hash function can be
1023			specified with this, eg. "hash=sha256".
1024
1025     The ``__spare[]`` space in the parameter block must be set to 0.  This is
1026     intended, amongst other things, to allow the passing of passphrases
1027     required to unlock a key.
1028
1029     If successful, encrypt, decrypt and sign all return the amount of data
1030     written into the output buffer.  Verification returns 0 on success.
1031
1032
1033  *  Watch a key or keyring for changes::
1034
1035	long keyctl(KEYCTL_WATCH_KEY, key_serial_t key, int queue_fd,
1036		    const struct watch_notification_filter *filter);
1037
1038     This will set or remove a watch for changes on the specified key or
1039     keyring.
1040
1041     "key" is the ID of the key to be watched.
1042
1043     "queue_fd" is a file descriptor referring to an open "/dev/watch_queue"
1044     which manages the buffer into which notifications will be delivered.
1045
1046     "filter" is either NULL to remove a watch or a filter specification to
1047     indicate what events are required from the key.
1048
1049     See Documentation/watch_queue.rst for more information.
1050
1051     Note that only one watch may be emplaced for any particular { key,
1052     queue_fd } combination.
1053
1054     Notification records look like::
1055
1056	struct key_notification {
1057		struct watch_notification watch;
1058		__u32	key_id;
1059		__u32	aux;
1060	};
1061
1062     In this, watch::type will be "WATCH_TYPE_KEY_NOTIFY" and subtype will be
1063     one of::
1064
1065	NOTIFY_KEY_INSTANTIATED
1066	NOTIFY_KEY_UPDATED
1067	NOTIFY_KEY_LINKED
1068	NOTIFY_KEY_UNLINKED
1069	NOTIFY_KEY_CLEARED
1070	NOTIFY_KEY_REVOKED
1071	NOTIFY_KEY_INVALIDATED
1072	NOTIFY_KEY_SETATTR
1073
1074     Where these indicate a key being instantiated/rejected, updated, a link
1075     being made in a keyring, a link being removed from a keyring, a keyring
1076     being cleared, a key being revoked, a key being invalidated or a key
1077     having one of its attributes changed (user, group, perm, timeout,
1078     restriction).
1079
1080     If a watched key is deleted, a basic watch_notification will be issued
1081     with "type" set to WATCH_TYPE_META and "subtype" set to
1082     watch_meta_removal_notification.  The watchpoint ID will be set in the
1083     "info" field.
1084
1085     This needs to be configured by enabling:
1086
1087	"Provide key/keyring change notifications" (KEY_NOTIFICATIONS)
1088
1089
1090Kernel Services
1091===============
1092
1093The kernel services for key management are fairly simple to deal with. They can
1094be broken down into two areas: keys and key types.
1095
1096Dealing with keys is fairly straightforward. Firstly, the kernel service
1097registers its type, then it searches for a key of that type. It should retain
1098the key as long as it has need of it, and then it should release it. For a
1099filesystem or device file, a search would probably be performed during the open
1100call, and the key released upon close. How to deal with conflicting keys due to
1101two different users opening the same file is left to the filesystem author to
1102solve.
1103
1104To access the key manager, the following header must be #included::
1105
1106	<linux/key.h>
1107
1108Specific key types should have a header file under include/keys/ that should be
1109used to access that type.  For keys of type "user", for example, that would be::
1110
1111	<keys/user-type.h>
1112
1113Note that there are two different types of pointers to keys that may be
1114encountered:
1115
1116  *  struct key *
1117
1118     This simply points to the key structure itself. Key structures will be at
1119     least four-byte aligned.
1120
1121  *  key_ref_t
1122
1123     This is equivalent to a ``struct key *``, but the least significant bit is set
1124     if the caller "possesses" the key. By "possession" it is meant that the
1125     calling processes has a searchable link to the key from one of its
1126     keyrings. There are three functions for dealing with these::
1127
1128	key_ref_t make_key_ref(const struct key *key, bool possession);
1129
1130	struct key *key_ref_to_ptr(const key_ref_t key_ref);
1131
1132	bool is_key_possessed(const key_ref_t key_ref);
1133
1134     The first function constructs a key reference from a key pointer and
1135     possession information (which must be true or false).
1136
1137     The second function retrieves the key pointer from a reference and the
1138     third retrieves the possession flag.
1139
1140When accessing a key's payload contents, certain precautions must be taken to
1141prevent access vs modification races. See the section "Notes on accessing
1142payload contents" for more information.
1143
1144 *  To search for a key, call::
1145
1146	struct key *request_key(const struct key_type *type,
1147				const char *description,
1148				const char *callout_info);
1149
1150    This is used to request a key or keyring with a description that matches
1151    the description specified according to the key type's match_preparse()
1152    method. This permits approximate matching to occur. If callout_string is
1153    not NULL, then /sbin/request-key will be invoked in an attempt to obtain
1154    the key from userspace. In that case, callout_string will be passed as an
1155    argument to the program.
1156
1157    Should the function fail error ENOKEY, EKEYEXPIRED or EKEYREVOKED will be
1158    returned.
1159
1160    If successful, the key will have been attached to the default keyring for
1161    implicitly obtained request-key keys, as set by KEYCTL_SET_REQKEY_KEYRING.
1162
1163    See also Documentation/security/keys/request-key.rst.
1164
1165
1166 *  To search for a key in a specific domain, call::
1167
1168	struct key *request_key_tag(const struct key_type *type,
1169				    const char *description,
1170				    struct key_tag *domain_tag,
1171				    const char *callout_info);
1172
1173    This is identical to request_key(), except that a domain tag may be
1174    specifies that causes search algorithm to only match keys matching that
1175    tag.  The domain_tag may be NULL, specifying a global domain that is
1176    separate from any nominated domain.
1177
1178
1179 *  To search for a key, passing auxiliary data to the upcaller, call::
1180
1181	struct key *request_key_with_auxdata(const struct key_type *type,
1182					     const char *description,
1183					     struct key_tag *domain_tag,
1184					     const void *callout_info,
1185					     size_t callout_len,
1186					     void *aux);
1187
1188    This is identical to request_key_tag(), except that the auxiliary data is
1189    passed to the key_type->request_key() op if it exists, and the
1190    callout_info is a blob of length callout_len, if given (the length may be
1191    0).
1192
1193
1194 *  To search for a key under RCU conditions, call::
1195
1196	struct key *request_key_rcu(const struct key_type *type,
1197				    const char *description,
1198				    struct key_tag *domain_tag);
1199
1200    which is similar to request_key_tag() except that it does not check for
1201    keys that are under construction and it will not call out to userspace to
1202    construct a key if it can't find a match.
1203
1204
1205 *  When it is no longer required, the key should be released using::
1206
1207	void key_put(struct key *key);
1208
1209    Or::
1210
1211	void key_ref_put(key_ref_t key_ref);
1212
1213    These can be called from interrupt context. If CONFIG_KEYS is not set then
1214    the argument will not be parsed.
1215
1216
1217 *  Extra references can be made to a key by calling one of the following
1218    functions::
1219
1220	struct key *__key_get(struct key *key);
1221	struct key *key_get(struct key *key);
1222
1223    Keys so references will need to be disposed of by calling key_put() when
1224    they've been finished with.  The key pointer passed in will be returned.
1225
1226    In the case of key_get(), if the pointer is NULL or CONFIG_KEYS is not set
1227    then the key will not be dereferenced and no increment will take place.
1228
1229
1230 *  A key's serial number can be obtained by calling::
1231
1232	key_serial_t key_serial(struct key *key);
1233
1234    If key is NULL or if CONFIG_KEYS is not set then 0 will be returned (in the
1235    latter case without parsing the argument).
1236
1237
1238 *  If a keyring was found in the search, this can be further searched by::
1239
1240	key_ref_t keyring_search(key_ref_t keyring_ref,
1241				 const struct key_type *type,
1242				 const char *description,
1243				 bool recurse)
1244
1245    This searches the specified keyring only (recurse == false) or keyring tree
1246    (recurse == true) specified for a matching key. Error ENOKEY is returned
1247    upon failure (use IS_ERR/PTR_ERR to determine). If successful, the returned
1248    key will need to be released.
1249
1250    The possession attribute from the keyring reference is used to control
1251    access through the permissions mask and is propagated to the returned key
1252    reference pointer if successful.
1253
1254
1255 *  A keyring can be created by::
1256
1257	struct key *keyring_alloc(const char *description, uid_t uid, gid_t gid,
1258				  const struct cred *cred,
1259				  key_perm_t perm,
1260				  struct key_restriction *restrict_link,
1261				  unsigned long flags,
1262				  struct key *dest);
1263
1264    This creates a keyring with the given attributes and returns it.  If dest
1265    is not NULL, the new keyring will be linked into the keyring to which it
1266    points.  No permission checks are made upon the destination keyring.
1267
1268    Error EDQUOT can be returned if the keyring would overload the quota (pass
1269    KEY_ALLOC_NOT_IN_QUOTA in flags if the keyring shouldn't be accounted
1270    towards the user's quota).  Error ENOMEM can also be returned.
1271
1272    If restrict_link is not NULL, it should point to a structure that contains
1273    the function that will be called each time an attempt is made to link a
1274    key into the new keyring.  The structure may also contain a key pointer
1275    and an associated key type.  The function is called to check whether a key
1276    may be added into the keyring or not.  The key type is used by the garbage
1277    collector to clean up function or data pointers in this structure if the
1278    given key type is unregistered.  Callers of key_create_or_update() within
1279    the kernel can pass KEY_ALLOC_BYPASS_RESTRICTION to suppress the check.
1280    An example of using this is to manage rings of cryptographic keys that are
1281    set up when the kernel boots where userspace is also permitted to add keys
1282    - provided they can be verified by a key the kernel already has.
1283
1284    When called, the restriction function will be passed the keyring being
1285    added to, the key type, the payload of the key being added, and data to be
1286    used in the restriction check.  Note that when a new key is being created,
1287    this is called between payload preparsing and actual key creation.  The
1288    function should return 0 to allow the link or an error to reject it.
1289
1290    A convenience function, restrict_link_reject, exists to always return
1291    -EPERM to in this case.
1292
1293
1294 *  To check the validity of a key, this function can be called::
1295
1296	int validate_key(struct key *key);
1297
1298    This checks that the key in question hasn't expired or and hasn't been
1299    revoked. Should the key be invalid, error EKEYEXPIRED or EKEYREVOKED will
1300    be returned. If the key is NULL or if CONFIG_KEYS is not set then 0 will be
1301    returned (in the latter case without parsing the argument).
1302
1303
1304 *  To register a key type, the following function should be called::
1305
1306	int register_key_type(struct key_type *type);
1307
1308    This will return error EEXIST if a type of the same name is already
1309    present.
1310
1311
1312 *  To unregister a key type, call::
1313
1314	void unregister_key_type(struct key_type *type);
1315
1316
1317Under some circumstances, it may be desirable to deal with a bundle of keys.
1318The facility provides access to the keyring type for managing such a bundle::
1319
1320	struct key_type key_type_keyring;
1321
1322This can be used with a function such as request_key() to find a specific
1323keyring in a process's keyrings.  A keyring thus found can then be searched
1324with keyring_search().  Note that it is not possible to use request_key() to
1325search a specific keyring, so using keyrings in this way is of limited utility.
1326
1327
1328Notes On Accessing Payload Contents
1329===================================
1330
1331The simplest payload is just data stored in key->payload directly.  In this
1332case, there's no need to indulge in RCU or locking when accessing the payload.
1333
1334More complex payload contents must be allocated and pointers to them set in the
1335key->payload.data[] array.  One of the following ways must be selected to
1336access the data:
1337
1338  1) Unmodifiable key type.
1339
1340     If the key type does not have a modify method, then the key's payload can
1341     be accessed without any form of locking, provided that it's known to be
1342     instantiated (uninstantiated keys cannot be "found").
1343
1344  2) The key's semaphore.
1345
1346     The semaphore could be used to govern access to the payload and to control
1347     the payload pointer. It must be write-locked for modifications and would
1348     have to be read-locked for general access. The disadvantage of doing this
1349     is that the accessor may be required to sleep.
1350
1351  3) RCU.
1352
1353     RCU must be used when the semaphore isn't already held; if the semaphore
1354     is held then the contents can't change under you unexpectedly as the
1355     semaphore must still be used to serialise modifications to the key. The
1356     key management code takes care of this for the key type.
1357
1358     However, this means using::
1359
1360	rcu_read_lock() ... rcu_dereference() ... rcu_read_unlock()
1361
1362     to read the pointer, and::
1363
1364	rcu_dereference() ... rcu_assign_pointer() ... call_rcu()
1365
1366     to set the pointer and dispose of the old contents after a grace period.
1367     Note that only the key type should ever modify a key's payload.
1368
1369     Furthermore, an RCU controlled payload must hold a struct rcu_head for the
1370     use of call_rcu() and, if the payload is of variable size, the length of
1371     the payload. key->datalen cannot be relied upon to be consistent with the
1372     payload just dereferenced if the key's semaphore is not held.
1373
1374     Note that key->payload.data[0] has a shadow that is marked for __rcu
1375     usage.  This is called key->payload.rcu_data0.  The following accessors
1376     wrap the RCU calls to this element:
1377
1378     a) Set or change the first payload pointer::
1379
1380		rcu_assign_keypointer(struct key *key, void *data);
1381
1382     b) Read the first payload pointer with the key semaphore held::
1383
1384		[const] void *dereference_key_locked([const] struct key *key);
1385
1386	 Note that the return value will inherit its constness from the key
1387	 parameter.  Static analysis will give an error if it things the lock
1388	 isn't held.
1389
1390     c) Read the first payload pointer with the RCU read lock held::
1391
1392		const void *dereference_key_rcu(const struct key *key);
1393
1394
1395Defining a Key Type
1396===================
1397
1398A kernel service may want to define its own key type. For instance, an AFS
1399filesystem might want to define a Kerberos 5 ticket key type. To do this, it
1400author fills in a key_type struct and registers it with the system.
1401
1402Source files that implement key types should include the following header file::
1403
1404	<linux/key-type.h>
1405
1406The structure has a number of fields, some of which are mandatory:
1407
1408  *  ``const char *name``
1409
1410     The name of the key type. This is used to translate a key type name
1411     supplied by userspace into a pointer to the structure.
1412
1413
1414  *  ``size_t def_datalen``
1415
1416     This is optional - it supplies the default payload data length as
1417     contributed to the quota. If the key type's payload is always or almost
1418     always the same size, then this is a more efficient way to do things.
1419
1420     The data length (and quota) on a particular key can always be changed
1421     during instantiation or update by calling::
1422
1423	int key_payload_reserve(struct key *key, size_t datalen);
1424
1425     With the revised data length. Error EDQUOT will be returned if this is not
1426     viable.
1427
1428
1429  *  ``int (*vet_description)(const char *description);``
1430
1431     This optional method is called to vet a key description.  If the key type
1432     doesn't approve of the key description, it may return an error, otherwise
1433     it should return 0.
1434
1435
1436  *  ``int (*preparse)(struct key_preparsed_payload *prep);``
1437
1438     This optional method permits the key type to attempt to parse payload
1439     before a key is created (add key) or the key semaphore is taken (update or
1440     instantiate key).  The structure pointed to by prep looks like::
1441
1442	struct key_preparsed_payload {
1443		char		*description;
1444		union key_payload payload;
1445		const void	*data;
1446		size_t		datalen;
1447		size_t		quotalen;
1448		time_t		expiry;
1449	};
1450
1451     Before calling the method, the caller will fill in data and datalen with
1452     the payload blob parameters; quotalen will be filled in with the default
1453     quota size from the key type; expiry will be set to TIME_T_MAX and the
1454     rest will be cleared.
1455
1456     If a description can be proposed from the payload contents, that should be
1457     attached as a string to the description field.  This will be used for the
1458     key description if the caller of add_key() passes NULL or "".
1459
1460     The method can attach anything it likes to payload.  This is merely passed
1461     along to the instantiate() or update() operations.  If set, the expiry
1462     time will be applied to the key if it is instantiated from this data.
1463
1464     The method should return 0 if successful or a negative error code
1465     otherwise.
1466
1467
1468  *  ``void (*free_preparse)(struct key_preparsed_payload *prep);``
1469
1470     This method is only required if the preparse() method is provided,
1471     otherwise it is unused.  It cleans up anything attached to the description
1472     and payload fields of the key_preparsed_payload struct as filled in by the
1473     preparse() method.  It will always be called after preparse() returns
1474     successfully, even if instantiate() or update() succeed.
1475
1476
1477  *  ``int (*instantiate)(struct key *key, struct key_preparsed_payload *prep);``
1478
1479     This method is called to attach a payload to a key during construction.
1480     The payload attached need not bear any relation to the data passed to this
1481     function.
1482
1483     The prep->data and prep->datalen fields will define the original payload
1484     blob.  If preparse() was supplied then other fields may be filled in also.
1485
1486     If the amount of data attached to the key differs from the size in
1487     keytype->def_datalen, then key_payload_reserve() should be called.
1488
1489     This method does not have to lock the key in order to attach a payload.
1490     The fact that KEY_FLAG_INSTANTIATED is not set in key->flags prevents
1491     anything else from gaining access to the key.
1492
1493     It is safe to sleep in this method.
1494
1495     generic_key_instantiate() is provided to simply copy the data from
1496     prep->payload.data[] to key->payload.data[], with RCU-safe assignment on
1497     the first element.  It will then clear prep->payload.data[] so that the
1498     free_preparse method doesn't release the data.
1499
1500
1501  *  ``int (*update)(struct key *key, const void *data, size_t datalen);``
1502
1503     If this type of key can be updated, then this method should be provided.
1504     It is called to update a key's payload from the blob of data provided.
1505
1506     The prep->data and prep->datalen fields will define the original payload
1507     blob.  If preparse() was supplied then other fields may be filled in also.
1508
1509     key_payload_reserve() should be called if the data length might change
1510     before any changes are actually made. Note that if this succeeds, the type
1511     is committed to changing the key because it's already been altered, so all
1512     memory allocation must be done first.
1513
1514     The key will have its semaphore write-locked before this method is called,
1515     but this only deters other writers; any changes to the key's payload must
1516     be made under RCU conditions, and call_rcu() must be used to dispose of
1517     the old payload.
1518
1519     key_payload_reserve() should be called before the changes are made, but
1520     after all allocations and other potentially failing function calls are
1521     made.
1522
1523     It is safe to sleep in this method.
1524
1525
1526  *  ``int (*match_preparse)(struct key_match_data *match_data);``
1527
1528     This method is optional.  It is called when a key search is about to be
1529     performed.  It is given the following structure::
1530
1531	struct key_match_data {
1532		bool (*cmp)(const struct key *key,
1533			    const struct key_match_data *match_data);
1534		const void	*raw_data;
1535		void		*preparsed;
1536		unsigned	lookup_type;
1537	};
1538
1539     On entry, raw_data will be pointing to the criteria to be used in matching
1540     a key by the caller and should not be modified.  ``(*cmp)()`` will be pointing
1541     to the default matcher function (which does an exact description match
1542     against raw_data) and lookup_type will be set to indicate a direct lookup.
1543
1544     The following lookup_type values are available:
1545
1546       *  KEYRING_SEARCH_LOOKUP_DIRECT - A direct lookup hashes the type and
1547      	  description to narrow down the search to a small number of keys.
1548
1549       *  KEYRING_SEARCH_LOOKUP_ITERATE - An iterative lookup walks all the
1550      	  keys in the keyring until one is matched.  This must be used for any
1551      	  search that's not doing a simple direct match on the key description.
1552
1553     The method may set cmp to point to a function of its choice that does some
1554     other form of match, may set lookup_type to KEYRING_SEARCH_LOOKUP_ITERATE
1555     and may attach something to the preparsed pointer for use by ``(*cmp)()``.
1556     ``(*cmp)()`` should return true if a key matches and false otherwise.
1557
1558     If preparsed is set, it may be necessary to use the match_free() method to
1559     clean it up.
1560
1561     The method should return 0 if successful or a negative error code
1562     otherwise.
1563
1564     It is permitted to sleep in this method, but ``(*cmp)()`` may not sleep as
1565     locks will be held over it.
1566
1567     If match_preparse() is not provided, keys of this type will be matched
1568     exactly by their description.
1569
1570
1571  *  ``void (*match_free)(struct key_match_data *match_data);``
1572
1573     This method is optional.  If given, it called to clean up
1574     match_data->preparsed after a successful call to match_preparse().
1575
1576
1577  *  ``void (*revoke)(struct key *key);``
1578
1579     This method is optional.  It is called to discard part of the payload
1580     data upon a key being revoked.  The caller will have the key semaphore
1581     write-locked.
1582
1583     It is safe to sleep in this method, though care should be taken to avoid
1584     a deadlock against the key semaphore.
1585
1586
1587  *  ``void (*destroy)(struct key *key);``
1588
1589     This method is optional. It is called to discard the payload data on a key
1590     when it is being destroyed.
1591
1592     This method does not need to lock the key to access the payload; it can
1593     consider the key as being inaccessible at this time. Note that the key's
1594     type may have been changed before this function is called.
1595
1596     It is not safe to sleep in this method; the caller may hold spinlocks.
1597
1598
1599  *  ``void (*describe)(const struct key *key, struct seq_file *p);``
1600
1601     This method is optional. It is called during /proc/keys reading to
1602     summarise a key's description and payload in text form.
1603
1604     This method will be called with the RCU read lock held. rcu_dereference()
1605     should be used to read the payload pointer if the payload is to be
1606     accessed. key->datalen cannot be trusted to stay consistent with the
1607     contents of the payload.
1608
1609     The description will not change, though the key's state may.
1610
1611     It is not safe to sleep in this method; the RCU read lock is held by the
1612     caller.
1613
1614
1615  *  ``long (*read)(const struct key *key, char __user *buffer, size_t buflen);``
1616
1617     This method is optional. It is called by KEYCTL_READ to translate the
1618     key's payload into something a blob of data for userspace to deal with.
1619     Ideally, the blob should be in the same format as that passed in to the
1620     instantiate and update methods.
1621
1622     If successful, the blob size that could be produced should be returned
1623     rather than the size copied.
1624
1625     This method will be called with the key's semaphore read-locked. This will
1626     prevent the key's payload changing. It is not necessary to use RCU locking
1627     when accessing the key's payload. It is safe to sleep in this method, such
1628     as might happen when the userspace buffer is accessed.
1629
1630
1631  *  ``int (*request_key)(struct key_construction *cons, const char *op, void *aux);``
1632
1633     This method is optional.  If provided, request_key() and friends will
1634     invoke this function rather than upcalling to /sbin/request-key to operate
1635     upon a key of this type.
1636
1637     The aux parameter is as passed to request_key_async_with_auxdata() and
1638     similar or is NULL otherwise.  Also passed are the construction record for
1639     the key to be operated upon and the operation type (currently only
1640     "create").
1641
1642     This method is permitted to return before the upcall is complete, but the
1643     following function must be called under all circumstances to complete the
1644     instantiation process, whether or not it succeeds, whether or not there's
1645     an error::
1646
1647	void complete_request_key(struct key_construction *cons, int error);
1648
1649     The error parameter should be 0 on success, -ve on error.  The
1650     construction record is destroyed by this action and the authorisation key
1651     will be revoked.  If an error is indicated, the key under construction
1652     will be negatively instantiated if it wasn't already instantiated.
1653
1654     If this method returns an error, that error will be returned to the
1655     caller of request_key*().  complete_request_key() must be called prior to
1656     returning.
1657
1658     The key under construction and the authorisation key can be found in the
1659     key_construction struct pointed to by cons:
1660
1661      *  ``struct key *key;``
1662
1663     	 The key under construction.
1664
1665      *  ``struct key *authkey;``
1666
1667     	 The authorisation key.
1668
1669
1670  *  ``struct key_restriction *(*lookup_restriction)(const char *params);``
1671
1672     This optional method is used to enable userspace configuration of keyring
1673     restrictions. The restriction parameter string (not including the key type
1674     name) is passed in, and this method returns a pointer to a key_restriction
1675     structure containing the relevant functions and data to evaluate each
1676     attempted key link operation. If there is no match, -EINVAL is returned.
1677
1678
1679  *  ``asym_eds_op`` and ``asym_verify_signature``::
1680
1681       int (*asym_eds_op)(struct kernel_pkey_params *params,
1682			  const void *in, void *out);
1683       int (*asym_verify_signature)(struct kernel_pkey_params *params,
1684				    const void *in, const void *in2);
1685
1686     These methods are optional.  If provided the first allows a key to be
1687     used to encrypt, decrypt or sign a blob of data, and the second allows a
1688     key to verify a signature.
1689
1690     In all cases, the following information is provided in the params block::
1691
1692	struct kernel_pkey_params {
1693		struct key	*key;
1694		const char	*encoding;
1695		const char	*hash_algo;
1696		char		*info;
1697		__u32		in_len;
1698		union {
1699			__u32	out_len;
1700			__u32	in2_len;
1701		};
1702		enum kernel_pkey_operation op : 8;
1703	};
1704
1705     This includes the key to be used; a string indicating the encoding to use
1706     (for instance, "pkcs1" may be used with an RSA key to indicate
1707     RSASSA-PKCS1-v1.5 or RSAES-PKCS1-v1.5 encoding or "raw" if no encoding);
1708     the name of the hash algorithm used to generate the data for a signature
1709     (if appropriate); the sizes of the input and output (or second input)
1710     buffers; and the ID of the operation to be performed.
1711
1712     For a given operation ID, the input and output buffers are used as
1713     follows::
1714
1715	Operation ID		in,in_len	out,out_len	in2,in2_len
1716	=======================	===============	===============	===============
1717	kernel_pkey_encrypt	Raw data	Encrypted data	-
1718	kernel_pkey_decrypt	Encrypted data	Raw data	-
1719	kernel_pkey_sign	Raw data	Signature	-
1720	kernel_pkey_verify	Raw data	-		Signature
1721
1722     asym_eds_op() deals with encryption, decryption and signature creation as
1723     specified by params->op.  Note that params->op is also set for
1724     asym_verify_signature().
1725
1726     Encrypting and signature creation both take raw data in the input buffer
1727     and return the encrypted result in the output buffer.  Padding may have
1728     been added if an encoding was set.  In the case of signature creation,
1729     depending on the encoding, the padding created may need to indicate the
1730     digest algorithm - the name of which should be supplied in hash_algo.
1731
1732     Decryption takes encrypted data in the input buffer and returns the raw
1733     data in the output buffer.  Padding will get checked and stripped off if
1734     an encoding was set.
1735
1736     Verification takes raw data in the input buffer and the signature in the
1737     second input buffer and checks that the one matches the other.  Padding
1738     will be validated.  Depending on the encoding, the digest algorithm used
1739     to generate the raw data may need to be indicated in hash_algo.
1740
1741     If successful, asym_eds_op() should return the number of bytes written
1742     into the output buffer.  asym_verify_signature() should return 0.
1743
1744     A variety of errors may be returned, including EOPNOTSUPP if the operation
1745     is not supported; EKEYREJECTED if verification fails; ENOPKG if the
1746     required crypto isn't available.
1747
1748
1749  *  ``asym_query``::
1750
1751       int (*asym_query)(const struct kernel_pkey_params *params,
1752			 struct kernel_pkey_query *info);
1753
1754     This method is optional.  If provided it allows information about the
1755     public or asymmetric key held in the key to be determined.
1756
1757     The parameter block is as for asym_eds_op() and co. but in_len and out_len
1758     are unused.  The encoding and hash_algo fields should be used to reduce
1759     the returned buffer/data sizes as appropriate.
1760
1761     If successful, the following information is filled in::
1762
1763	struct kernel_pkey_query {
1764		__u32		supported_ops;
1765		__u32		key_size;
1766		__u16		max_data_size;
1767		__u16		max_sig_size;
1768		__u16		max_enc_size;
1769		__u16		max_dec_size;
1770	};
1771
1772     The supported_ops field will contain a bitmask indicating what operations
1773     are supported by the key, including encryption of a blob, decryption of a
1774     blob, signing a blob and verifying the signature on a blob.  The following
1775     constants are defined for this::
1776
1777	KEYCTL_SUPPORTS_{ENCRYPT,DECRYPT,SIGN,VERIFY}
1778
1779     The key_size field is the size of the key in bits.  max_data_size and
1780     max_sig_size are the maximum raw data and signature sizes for creation and
1781     verification of a signature; max_enc_size and max_dec_size are the maximum
1782     raw data and signature sizes for encryption and decryption.  The
1783     max_*_size fields are measured in bytes.
1784
1785     If successful, 0 will be returned.  If the key doesn't support this,
1786     EOPNOTSUPP will be returned.
1787
1788
1789Request-Key Callback Service
1790============================
1791
1792To create a new key, the kernel will attempt to execute the following command
1793line::
1794
1795	/sbin/request-key create <key> <uid> <gid> \
1796		<threadring> <processring> <sessionring> <callout_info>
1797
1798<key> is the key being constructed, and the three keyrings are the process
1799keyrings from the process that caused the search to be issued. These are
1800included for two reasons:
1801
1802   1  There may be an authentication token in one of the keyrings that is
1803      required to obtain the key, eg: a Kerberos Ticket-Granting Ticket.
1804
1805   2  The new key should probably be cached in one of these rings.
1806
1807This program should set it UID and GID to those specified before attempting to
1808access any more keys. It may then look around for a user specific process to
1809hand the request off to (perhaps a path held in placed in another key by, for
1810example, the KDE desktop manager).
1811
1812The program (or whatever it calls) should finish construction of the key by
1813calling KEYCTL_INSTANTIATE or KEYCTL_INSTANTIATE_IOV, which also permits it to
1814cache the key in one of the keyrings (probably the session ring) before
1815returning.  Alternatively, the key can be marked as negative with KEYCTL_NEGATE
1816or KEYCTL_REJECT; this also permits the key to be cached in one of the
1817keyrings.
1818
1819If it returns with the key remaining in the unconstructed state, the key will
1820be marked as being negative, it will be added to the session keyring, and an
1821error will be returned to the key requestor.
1822
1823Supplementary information may be provided from whoever or whatever invoked this
1824service. This will be passed as the <callout_info> parameter. If no such
1825information was made available, then "-" will be passed as this parameter
1826instead.
1827
1828
1829Similarly, the kernel may attempt to update an expired or a soon to expire key
1830by executing::
1831
1832	/sbin/request-key update <key> <uid> <gid> \
1833		<threadring> <processring> <sessionring>
1834
1835In this case, the program isn't required to actually attach the key to a ring;
1836the rings are provided for reference.
1837
1838
1839Garbage Collection
1840==================
1841
1842Dead keys (for which the type has been removed) will be automatically unlinked
1843from those keyrings that point to them and deleted as soon as possible by a
1844background garbage collector.
1845
1846Similarly, revoked and expired keys will be garbage collected, but only after a
1847certain amount of time has passed.  This time is set as a number of seconds in::
1848
1849	/proc/sys/kernel/keys/gc_delay
1850