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