view Documentation/keys.txt @ 897:329ea0ccb344

balloon: try harder to balloon up under memory pressure.

Currently if the balloon driver is unable to increase the guest's
reservation it assumes the failure was due to reaching its full
allocation, gives up on the ballooning operation and records the limit
it reached as the "hard limit". The driver will not try again until
the target is set again (even to the same value).

However it is possible that ballooning has in fact failed due to
memory pressure in the host and therefore it is desirable to keep
attempting to reach the target in case memory becomes available. The
most likely scenario is that some guests are ballooning down while
others are ballooning up and therefore there is temporary memory
pressure while things stabilise. You would not expect a well behaved
toolstack to ask a domain to balloon to more than its allocation nor
would you expect it to deliberately over-commit memory by setting
balloon targets which exceed the total host memory.

This patch drops the concept of a hard limit and causes the balloon
driver to retry increasing the reservation on a timer in the same
manner as when decreasing the reservation.

Also if we partially succeed in increasing the reservation
(i.e. receive less pages than we asked for) then we may as well keep
those pages rather than returning them to Xen.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
author Keir Fraser <keir.fraser@citrix.com>
date Fri Jun 05 14:01:20 2009 +0100 (2009-06-05)
parents 831230e53067
line source
1 ============================
3 ============================
5 This service allows cryptographic keys, authentication tokens, cross-domain
6 user mappings, and similar to be cached in the kernel for the use of
7 filesystems other kernel services.
9 Keyrings are permitted; these are a special type of key that can hold links to
10 other keys. Processes each have three standard keyring subscriptions that a
11 kernel service can search for relevant keys.
13 The key service can be configured on by enabling:
15 "Security options"/"Enable access key retention support" (CONFIG_KEYS)
17 This document has the following sections:
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 - Key access filesystem
32 ============
34 ============
36 In this context, keys represent units of cryptographic data, authentication
37 tokens, keyrings, etc.. These are represented in the kernel by struct key.
39 Each key has a number of attributes:
41 - A serial number.
42 - A type.
43 - A description (for matching a key in a search).
44 - Access control information.
45 - An expiry time.
46 - A payload.
47 - State.
50 (*) Each key is issued a serial number of type key_serial_t that is unique for
51 the lifetime of that key. All serial numbers are positive non-zero 32-bit
52 integers.
54 Userspace programs can use a key's serial numbers as a way to gain access
55 to it, subject to permission checking.
57 (*) Each key is of a defined "type". Types must be registered inside the
58 kernel by a kernel service (such as a filesystem) before keys of that type
59 can be added or used. Userspace programs cannot define new types directly.
61 Key types are represented in the kernel by struct key_type. This defines a
62 number of operations that can be performed on a key of that type.
64 Should a type be removed from the system, all the keys of that type will
65 be invalidated.
67 (*) Each key has a description. This should be a printable string. The key
68 type provides an operation to perform a match between the description on a
69 key and a criterion string.
71 (*) Each key has an owner user ID, a group ID and a permissions mask. These
72 are used to control what a process may do to a key from userspace, and
73 whether a kernel service will be able to find the key.
75 (*) Each key can be set to expire at a specific time by the key type's
76 instantiation function. Keys can also be immortal.
78 (*) Each key can have a payload. This is a quantity of data that represent the
79 actual "key". In the case of a keyring, this is a list of keys to which
80 the keyring links; in the case of a user-defined key, it's an arbitrary
81 blob of data.
83 Having a payload is not required; and the payload can, in fact, just be a
84 value stored in the struct key itself.
86 When a key is instantiated, the key type's instantiation function is
87 called with a blob of data, and that then creates the key's payload in
88 some way.
90 Similarly, when userspace wants to read back the contents of the key, if
91 permitted, another key type operation will be called to convert the key's
92 attached payload back into a blob of data.
94 (*) Each key can be in one of a number of basic states:
96 (*) Uninstantiated. The key exists, but does not have any data attached.
97 Keys being requested from userspace will be in this state.
99 (*) Instantiated. This is the normal state. The key is fully formed, and
100 has data attached.
102 (*) Negative. This is a relatively short-lived state. The key acts as a
103 note saying that a previous call out to userspace failed, and acts as
104 a throttle on key lookups. A negative key can be updated to a normal
105 state.
107 (*) Expired. Keys can have lifetimes set. If their lifetime is exceeded,
108 they traverse to this state. An expired key can be updated back to a
109 normal state.
111 (*) Revoked. A key is put in this state by userspace action. It can't be
112 found or operated upon (apart from by unlinking it).
114 (*) Dead. The key's type was unregistered, and so the key is now useless.
117 ====================
119 ====================
121 The key service provides a number of features besides keys:
123 (*) The key service defines two special key types:
125 (+) "keyring"
127 Keyrings are special keys that contain a list of other keys. Keyring
128 lists can be modified using various system calls. Keyrings should not
129 be given a payload when created.
131 (+) "user"
133 A key of this type has a description and a payload that are arbitrary
134 blobs of data. These can be created, updated and read by userspace,
135 and aren't intended for use by kernel services.
137 (*) Each process subscribes to three keyrings: a thread-specific keyring, a
138 process-specific keyring, and a session-specific keyring.
140 The thread-specific keyring is discarded from the child when any sort of
141 clone, fork, vfork or execve occurs. A new keyring is created only when
142 required.
144 The process-specific keyring is replaced with an empty one in the child on
145 clone, fork, vfork unless CLONE_THREAD is supplied, in which case it is
146 shared. execve also discards the process's process keyring and creates a
147 new one.
149 The session-specific keyring is persistent across clone, fork, vfork and
150 execve, even when the latter executes a set-UID or set-GID binary. A
151 process can, however, replace its current session keyring with a new one
152 by using PR_JOIN_SESSION_KEYRING. It is permitted to request an anonymous
153 new one, or to attempt to create or join one of a specific name.
155 The ownership of the thread keyring changes when the real UID and GID of
156 the thread changes.
158 (*) Each user ID resident in the system holds two special keyrings: a user
159 specific keyring and a default user session keyring. The default session
160 keyring is initialised with a link to the user-specific keyring.
162 When a process changes its real UID, if it used to have no session key, it
163 will be subscribed to the default session key for the new UID.
165 If a process attempts to access its session key when it doesn't have one,
166 it will be subscribed to the default for its current UID.
168 (*) Each user has two quotas against which the keys they own are tracked. One
169 limits the total number of keys and keyrings, the other limits the total
170 amount of description and payload space that can be consumed.
172 The user can view information on this and other statistics through procfs
173 files.
175 Process-specific and thread-specific keyrings are not counted towards a
176 user's quota.
178 If a system call that modifies a key or keyring in some way would put the
179 user over quota, the operation is refused and error EDQUOT is returned.
181 (*) There's a system call interface by which userspace programs can create and
182 manipulate keys and keyrings.
184 (*) There's a kernel interface by which services can register types and search
185 for keys.
187 (*) There's a way for the a search done from the kernel to call back to
188 userspace to request a key that can't be found in a process's keyrings.
190 (*) An optional filesystem is available through which the key database can be
191 viewed and manipulated.
194 ======================
196 ======================
198 Keys have an owner user ID, a group access ID, and a permissions mask. The mask
199 has up to eight bits each for possessor, user, group and other access. Only
200 six of each set of eight bits are defined. These permissions granted are:
202 (*) View
204 This permits a key or keyring's attributes to be viewed - including key
205 type and description.
207 (*) Read
209 This permits a key's payload to be viewed or a keyring's list of linked
210 keys.
212 (*) Write
214 This permits a key's payload to be instantiated or updated, or it allows a
215 link to be added to or removed from a keyring.
217 (*) Search
219 This permits keyrings to be searched and keys to be found. Searches can
220 only recurse into nested keyrings that have search permission set.
222 (*) Link
224 This permits a key or keyring to be linked to. To create a link from a
225 keyring to a key, a process must have Write permission on the keyring and
226 Link permission on the key.
228 (*) Set Attribute
230 This permits a key's UID, GID and permissions mask to be changed.
232 For changing the ownership, group ID or permissions mask, being the owner of
233 the key or having the sysadmin capability is sufficient.
236 ===============
238 ===============
240 The security class "key" has been added to SELinux so that mandatory access
241 controls can be applied to keys created within various contexts. This support
242 is preliminary, and is likely to change quite significantly in the near future.
243 Currently, all of the basic permissions explained above are provided in SELinux
244 as well; SELinux is simply invoked after all basic permission checks have been
245 performed.
247 The value of the file /proc/self/attr/keycreate influences the labeling of
248 newly-created keys. If the contents of that file correspond to an SELinux
249 security context, then the key will be assigned that context. Otherwise, the
250 key will be assigned the current context of the task that invoked the key
251 creation request. Tasks must be granted explicit permission to assign a
252 particular context to newly-created keys, using the "create" permission in the
253 key security class.
255 The default keyrings associated with users will be labeled with the default
256 context of the user if and only if the login programs have been instrumented to
257 properly initialize keycreate during the login process. Otherwise, they will
258 be labeled with the context of the login program itself.
260 Note, however, that the default keyrings associated with the root user are
261 labeled with the default kernel context, since they are created early in the
262 boot process, before root has a chance to log in.
264 The keyrings associated with new threads are each labeled with the context of
265 their associated thread, and both session and process keyrings are handled
266 similarly.
269 ================
271 ================
273 Two files have been added to procfs by which an administrator can find out
274 about the status of the key service:
276 (*) /proc/keys
278 This lists the keys that are currently viewable by the task reading the
279 file, giving information about their type, description and permissions.
280 It is not possible to view the payload of the key this way, though some
281 information about it may be given.
283 The only keys included in the list are those that grant View permission to
284 the reading process whether or not it possesses them. Note that LSM
285 security checks are still performed, and may further filter out keys that
286 the current process is not authorised to view.
288 The contents of the file look like this:
291 00000001 I----- 39 perm 1f3f0000 0 0 keyring _uid_ses.0: 1/4
292 00000002 I----- 2 perm 1f3f0000 0 0 keyring _uid.0: empty
293 00000007 I----- 1 perm 1f3f0000 0 0 keyring _pid.1: empty
294 0000018d I----- 1 perm 1f3f0000 0 0 keyring _pid.412: empty
295 000004d2 I--Q-- 1 perm 1f3f0000 32 -1 keyring _uid.32: 1/4
296 000004d3 I--Q-- 3 perm 1f3f0000 32 -1 keyring _uid_ses.32: empty
297 00000892 I--QU- 1 perm 1f000000 0 0 user metal:copper: 0
298 00000893 I--Q-N 1 35s 1f3f0000 0 0 user metal:silver: 0
299 00000894 I--Q-- 1 10h 003f0000 0 0 user metal:gold: 0
301 The flags are:
303 I Instantiated
304 R Revoked
305 D Dead
306 Q Contributes to user's quota
307 U Under contruction by callback to userspace
308 N Negative key
310 This file must be enabled at kernel configuration time as it allows anyone
311 to list the keys database.
313 (*) /proc/key-users
315 This file lists the tracking data for each user that has at least one key
316 on the system. Such data includes quota information and statistics:
318 [root@andromeda root]# cat /proc/key-users
319 0: 46 45/45 1/100 13/10000
320 29: 2 2/2 2/100 40/10000
321 32: 2 2/2 2/100 40/10000
322 38: 2 2/2 2/100 40/10000
324 The format of each line is
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
332 ===============================
334 ===============================
336 Userspace can manipulate keys directly through three new syscalls: add_key,
337 request_key and keyctl. The latter provides a number of functions for
338 manipulating keys.
340 When referring to a key directly, userspace programs should use the key's
341 serial number (a positive 32-bit integer). However, there are some special
342 values available for referring to special keys and keyrings that relate to the
343 process making the call:
346 ============================== ====== ===========================
347 KEY_SPEC_THREAD_KEYRING -1 thread-specific keyring
348 KEY_SPEC_PROCESS_KEYRING -2 process-specific keyring
349 KEY_SPEC_SESSION_KEYRING -3 session-specific keyring
350 KEY_SPEC_USER_KEYRING -4 UID-specific keyring
351 KEY_SPEC_USER_SESSION_KEYRING -5 UID-session keyring
352 KEY_SPEC_GROUP_KEYRING -6 GID-specific keyring
353 KEY_SPEC_REQKEY_AUTH_KEY -7 assumed request_key()
354 authorisation key
357 The main syscalls are:
359 (*) Create a new key of given type, description and payload and add it to the
360 nominated keyring:
362 key_serial_t add_key(const char *type, const char *desc,
363 const void *payload, size_t plen,
364 key_serial_t keyring);
366 If a key of the same type and description as that proposed already exists
367 in the keyring, this will try to update it with the given payload, or it
368 will return error EEXIST if that function is not supported by the key
369 type. The process must also have permission to write to the key to be able
370 to update it. The new key will have all user permissions granted and no
371 group or third party permissions.
373 Otherwise, this will attempt to create a new key of the specified type and
374 description, and to instantiate it with the supplied payload and attach it
375 to the keyring. In this case, an error will be generated if the process
376 does not have permission to write to the keyring.
378 The payload is optional, and the pointer can be NULL if not required by
379 the type. The payload is plen in size, and plen can be zero for an empty
380 payload.
382 A new keyring can be generated by setting type "keyring", the keyring name
383 as the description (or NULL) and setting the payload to NULL.
385 User defined keys can be created by specifying type "user". It is
386 recommended that a user defined key's description by prefixed with a type
387 ID and a colon, such as "krb5tgt:" for a Kerberos 5 ticket granting
388 ticket.
390 Any other type must have been registered with the kernel in advance by a
391 kernel service such as a filesystem.
393 The ID of the new or updated key is returned if successful.
396 (*) Search the process's keyrings for a key, potentially calling out to
397 userspace to create it.
399 key_serial_t request_key(const char *type, const char *description,
400 const char *callout_info,
401 key_serial_t dest_keyring);
403 This function searches all the process's keyrings in the order thread,
404 process, session for a matching key. This works very much like
405 KEYCTL_SEARCH, including the optional attachment of the discovered key to
406 a keyring.
408 If a key cannot be found, and if callout_info is not NULL, then
409 /sbin/request-key will be invoked in an attempt to obtain a key. The
410 callout_info string will be passed as an argument to the program.
412 See also Documentation/keys-request-key.txt.
415 The keyctl syscall functions are:
417 (*) Map a special key ID to a real key ID for this process:
419 key_serial_t keyctl(KEYCTL_GET_KEYRING_ID, key_serial_t id,
420 int create);
422 The special key specified by "id" is looked up (with the key being created
423 if necessary) and the ID of the key or keyring thus found is returned if
424 it exists.
426 If the key does not yet exist, the key will be created if "create" is
427 non-zero; and the error ENOKEY will be returned if "create" is zero.
430 (*) Replace the session keyring this process subscribes to with a new one:
432 key_serial_t keyctl(KEYCTL_JOIN_SESSION_KEYRING, const char *name);
434 If name is NULL, an anonymous keyring is created attached to the process
435 as its session keyring, displacing the old session keyring.
437 If name is not NULL, if a keyring of that name exists, the process
438 attempts to attach it as the session keyring, returning an error if that
439 is not permitted; otherwise a new keyring of that name is created and
440 attached as the session keyring.
442 To attach to a named keyring, the keyring must have search permission for
443 the process's ownership.
445 The ID of the new session keyring is returned if successful.
448 (*) Update the specified key:
450 long keyctl(KEYCTL_UPDATE, key_serial_t key, const void *payload,
451 size_t plen);
453 This will try to update the specified key with the given payload, or it
454 will return error EOPNOTSUPP if that function is not supported by the key
455 type. The process must also have permission to write to the key to be able
456 to update it.
458 The payload is of length plen, and may be absent or empty as for
459 add_key().
462 (*) Revoke a key:
464 long keyctl(KEYCTL_REVOKE, key_serial_t key);
466 This makes a key unavailable for further operations. Further attempts to
467 use the key will be met with error EKEYREVOKED, and the key will no longer
468 be findable.
471 (*) Change the ownership of a key:
473 long keyctl(KEYCTL_CHOWN, key_serial_t key, uid_t uid, gid_t gid);
475 This function permits a key's owner and group ID to be changed. Either one
476 of uid or gid can be set to -1 to suppress that change.
478 Only the superuser can change a key's owner to something other than the
479 key's current owner. Similarly, only the superuser can change a key's
480 group ID to something other than the calling process's group ID or one of
481 its group list members.
484 (*) Change the permissions mask on a key:
486 long keyctl(KEYCTL_SETPERM, key_serial_t key, key_perm_t perm);
488 This function permits the owner of a key or the superuser to change the
489 permissions mask on a key.
491 Only bits the available bits are permitted; if any other bits are set,
492 error EINVAL will be returned.
495 (*) Describe a key:
497 long keyctl(KEYCTL_DESCRIBE, key_serial_t key, char *buffer,
498 size_t buflen);
500 This function returns a summary of the key's attributes (but not its
501 payload data) as a string in the buffer provided.
503 Unless there's an error, it always returns the amount of data it could
504 produce, even if that's too big for the buffer, but it won't copy more
505 than requested to userspace. If the buffer pointer is NULL then no copy
506 will take place.
508 A process must have view permission on the key for this function to be
509 successful.
511 If successful, a string is placed in the buffer in the following format:
513 <type>;<uid>;<gid>;<perm>;<description>
515 Where type and description are strings, uid and gid are decimal, and perm
516 is hexadecimal. A NUL character is included at the end of the string if
517 the buffer is sufficiently big.
519 This can be parsed with
521 sscanf(buffer, "%[^;];%d;%d;%o;%s", type, &uid, &gid, &mode, desc);
524 (*) Clear out a keyring:
526 long keyctl(KEYCTL_CLEAR, key_serial_t keyring);
528 This function clears the list of keys attached to a keyring. The calling
529 process must have write permission on the keyring, and it must be a
530 keyring (or else error ENOTDIR will result).
533 (*) Link a key into a keyring:
535 long keyctl(KEYCTL_LINK, key_serial_t keyring, key_serial_t key);
537 This function creates a link from the keyring to the key. The process must
538 have write permission on the keyring and must have link permission on the
539 key.
541 Should the keyring not be a keyring, error ENOTDIR will result; and if the
542 keyring is full, error ENFILE will result.
544 The link procedure checks the nesting of the keyrings, returning ELOOP if
545 it appears too deep or EDEADLK if the link would introduce a cycle.
547 Any links within the keyring to keys that match the new key in terms of
548 type and description will be discarded from the keyring as the new one is
549 added.
552 (*) Unlink a key or keyring from another keyring:
554 long keyctl(KEYCTL_UNLINK, key_serial_t keyring, key_serial_t key);
556 This function looks through the keyring for the first link to the
557 specified key, and removes it if found. Subsequent links to that key are
558 ignored. The process must have write permission on the keyring.
560 If the keyring is not a keyring, error ENOTDIR will result; and if the key
561 is not present, error ENOENT will be the result.
564 (*) Search a keyring tree for a key:
566 key_serial_t keyctl(KEYCTL_SEARCH, key_serial_t keyring,
567 const char *type, const char *description,
568 key_serial_t dest_keyring);
570 This searches the keyring tree headed by the specified keyring until a key
571 is found that matches the type and description criteria. Each keyring is
572 checked for keys before recursion into its children occurs.
574 The process must have search permission on the top level keyring, or else
575 error EACCES will result. Only keyrings that the process has search
576 permission on will be recursed into, and only keys and keyrings for which
577 a process has search permission can be matched. If the specified keyring
578 is not a keyring, ENOTDIR will result.
580 If the search succeeds, the function will attempt to link the found key
581 into the destination keyring if one is supplied (non-zero ID). All the
582 constraints applicable to KEYCTL_LINK apply in this case too.
584 Error ENOKEY, EKEYREVOKED or EKEYEXPIRED will be returned if the search
585 fails. On success, the resulting key ID will be returned.
588 (*) Read the payload data from a key:
590 long keyctl(KEYCTL_READ, key_serial_t keyring, char *buffer,
591 size_t buflen);
593 This function attempts to read the payload data from the specified key
594 into the buffer. The process must have read permission on the key to
595 succeed.
597 The returned data will be processed for presentation by the key type. For
598 instance, a keyring will return an array of key_serial_t entries
599 representing the IDs of all the keys to which it is subscribed. The user
600 defined key type will return its data as is. If a key type does not
601 implement this function, error EOPNOTSUPP will result.
603 As much of the data as can be fitted into the buffer will be copied to
604 userspace if the buffer pointer is not NULL.
606 On a successful return, the function will always return the amount of data
607 available rather than the amount copied.
610 (*) Instantiate a partially constructed key.
612 long keyctl(KEYCTL_INSTANTIATE, key_serial_t key,
613 const void *payload, size_t plen,
614 key_serial_t keyring);
616 If the kernel calls back to userspace to complete the instantiation of a
617 key, userspace should use this call to supply data for the key before the
618 invoked process returns, or else the key will be marked negative
619 automatically.
621 The process must have write access on the key to be able to instantiate
622 it, and the key must be uninstantiated.
624 If a keyring is specified (non-zero), the key will also be linked into
625 that keyring, however all the constraints applying in KEYCTL_LINK apply in
626 this case too.
628 The payload and plen arguments describe the payload data as for add_key().
631 (*) Negatively instantiate a partially constructed key.
633 long keyctl(KEYCTL_NEGATE, key_serial_t key,
634 unsigned timeout, key_serial_t keyring);
636 If the kernel calls back to userspace to complete the instantiation of a
637 key, userspace should use this call mark the key as negative before the
638 invoked process returns if it is unable to fulfil the request.
640 The process must have write access on the key to be able to instantiate
641 it, and the key must be uninstantiated.
643 If a keyring is specified (non-zero), the key will also be linked into
644 that keyring, however all the constraints applying in KEYCTL_LINK apply in
645 this case too.
648 (*) Set the default request-key destination keyring.
650 long keyctl(KEYCTL_SET_REQKEY_KEYRING, int reqkey_defl);
652 This sets the default keyring to which implicitly requested keys will be
653 attached for this thread. reqkey_defl should be one of these constants:
656 ====================================== ====== =======================
666 The old default will be returned if successful and error EINVAL will be
667 returned if reqkey_defl is not one of the above values.
669 The default keyring can be overridden by the keyring indicated to the
670 request_key() system call.
672 Note that this setting is inherited across fork/exec.
674 [1] The default default is: the thread keyring if there is one, otherwise
675 the process keyring if there is one, otherwise the session keyring if
676 there is one, otherwise the user default session keyring.
679 (*) Set the timeout on a key.
681 long keyctl(KEYCTL_SET_TIMEOUT, key_serial_t key, unsigned timeout);
683 This sets or clears the timeout on a key. The timeout can be 0 to clear
684 the timeout or a number of seconds to set the expiry time that far into
685 the future.
687 The process must have attribute modification access on a key to set its
688 timeout. Timeouts may not be set with this function on negative, revoked
689 or expired keys.
692 (*) Assume the authority granted to instantiate a key
694 long keyctl(KEYCTL_ASSUME_AUTHORITY, key_serial_t key);
696 This assumes or divests the authority required to instantiate the
697 specified key. Authority can only be assumed if the thread has the
698 authorisation key associated with the specified key in its keyrings
699 somewhere.
701 Once authority is assumed, searches for keys will also search the
702 requester's keyrings using the requester's security label, UID, GID and
703 groups.
705 If the requested authority is unavailable, error EPERM will be returned,
706 likewise if the authority has been revoked because the target key is
707 already instantiated.
709 If the specified key is 0, then any assumed authority will be divested.
711 The assumed authorititive key is inherited across fork and exec.
714 ===============
716 ===============
718 The kernel services for key managment are fairly simple to deal with. They can
719 be broken down into two areas: keys and key types.
721 Dealing with keys is fairly straightforward. Firstly, the kernel service
722 registers its type, then it searches for a key of that type. It should retain
723 the key as long as it has need of it, and then it should release it. For a
724 filesystem or device file, a search would probably be performed during the open
725 call, and the key released upon close. How to deal with conflicting keys due to
726 two different users opening the same file is left to the filesystem author to
727 solve.
729 Note that there are two different types of pointers to keys that may be
730 encountered:
732 (*) struct key *
734 This simply points to the key structure itself. Key structures will be at
735 least four-byte aligned.
737 (*) key_ref_t
739 This is equivalent to a struct key *, but the least significant bit is set
740 if the caller "possesses" the key. By "possession" it is meant that the
741 calling processes has a searchable link to the key from one of its
742 keyrings. There are three functions for dealing with these:
744 key_ref_t make_key_ref(const struct key *key,
745 unsigned long possession);
747 struct key *key_ref_to_ptr(const key_ref_t key_ref);
749 unsigned long is_key_possessed(const key_ref_t key_ref);
751 The first function constructs a key reference from a key pointer and
752 possession information (which must be 0 or 1 and not any other value).
754 The second function retrieves the key pointer from a reference and the
755 third retrieves the possession flag.
757 When accessing a key's payload contents, certain precautions must be taken to
758 prevent access vs modification races. See the section "Notes on accessing
759 payload contents" for more information.
761 (*) To search for a key, call:
763 struct key *request_key(const struct key_type *type,
764 const char *description,
765 const char *callout_string);
767 This is used to request a key or keyring with a description that matches
768 the description specified according to the key type's match function. This
769 permits approximate matching to occur. If callout_string is not NULL, then
770 /sbin/request-key will be invoked in an attempt to obtain the key from
771 userspace. In that case, callout_string will be passed as an argument to
772 the program.
774 Should the function fail error ENOKEY, EKEYEXPIRED or EKEYREVOKED will be
775 returned.
777 If successful, the key will have been attached to the default keyring for
778 implicitly obtained request-key keys, as set by KEYCTL_SET_REQKEY_KEYRING.
780 See also Documentation/keys-request-key.txt.
783 (*) To search for a key, passing auxiliary data to the upcaller, call:
785 struct key *request_key_with_auxdata(const struct key_type *type,
786 const char *description,
787 const char *callout_string,
788 void *aux);
790 This is identical to request_key(), except that the auxiliary data is
791 passed to the key_type->request_key() op if it exists.
794 (*) When it is no longer required, the key should be released using:
796 void key_put(struct key *key);
798 Or:
800 void key_ref_put(key_ref_t key_ref);
802 These can be called from interrupt context. If CONFIG_KEYS is not set then
803 the argument will not be parsed.
806 (*) Extra references can be made to a key by calling the following function:
808 struct key *key_get(struct key *key);
810 These need to be disposed of by calling key_put() when they've been
811 finished with. The key pointer passed in will be returned. If the pointer
812 is NULL or CONFIG_KEYS is not set then the key will not be dereferenced and
813 no increment will take place.
816 (*) A key's serial number can be obtained by calling:
818 key_serial_t key_serial(struct key *key);
820 If key is NULL or if CONFIG_KEYS is not set then 0 will be returned (in the
821 latter case without parsing the argument).
824 (*) If a keyring was found in the search, this can be further searched by:
826 key_ref_t keyring_search(key_ref_t keyring_ref,
827 const struct key_type *type,
828 const char *description)
830 This searches the keyring tree specified for a matching key. Error ENOKEY
831 is returned upon failure (use IS_ERR/PTR_ERR to determine). If successful,
832 the returned key will need to be released.
834 The possession attribute from the keyring reference is used to control
835 access through the permissions mask and is propagated to the returned key
836 reference pointer if successful.
839 (*) To check the validity of a key, this function can be called:
841 int validate_key(struct key *key);
843 This checks that the key in question hasn't expired or and hasn't been
844 revoked. Should the key be invalid, error EKEYEXPIRED or EKEYREVOKED will
845 be returned. If the key is NULL or if CONFIG_KEYS is not set then 0 will be
846 returned (in the latter case without parsing the argument).
849 (*) To register a key type, the following function should be called:
851 int register_key_type(struct key_type *type);
853 This will return error EEXIST if a type of the same name is already
854 present.
857 (*) To unregister a key type, call:
859 void unregister_key_type(struct key_type *type);
862 ===================================
864 ===================================
866 The simplest payload is just a number in key->payload.value. In this case,
867 there's no need to indulge in RCU or locking when accessing the payload.
869 More complex payload contents must be allocated and a pointer to them set in
870 key->payload.data. One of the following ways must be selected to access the
871 data:
873 (1) Unmodifiable key type.
875 If the key type does not have a modify method, then the key's payload can
876 be accessed without any form of locking, provided that it's known to be
877 instantiated (uninstantiated keys cannot be "found").
879 (2) The key's semaphore.
881 The semaphore could be used to govern access to the payload and to control
882 the payload pointer. It must be write-locked for modifications and would
883 have to be read-locked for general access. The disadvantage of doing this
884 is that the accessor may be required to sleep.
886 (3) RCU.
888 RCU must be used when the semaphore isn't already held; if the semaphore
889 is held then the contents can't change under you unexpectedly as the
890 semaphore must still be used to serialise modifications to the key. The
891 key management code takes care of this for the key type.
893 However, this means using:
895 rcu_read_lock() ... rcu_dereference() ... rcu_read_unlock()
897 to read the pointer, and:
899 rcu_dereference() ... rcu_assign_pointer() ... call_rcu()
901 to set the pointer and dispose of the old contents after a grace period.
902 Note that only the key type should ever modify a key's payload.
904 Furthermore, an RCU controlled payload must hold a struct rcu_head for the
905 use of call_rcu() and, if the payload is of variable size, the length of
906 the payload. key->datalen cannot be relied upon to be consistent with the
907 payload just dereferenced if the key's semaphore is not held.
910 ===================
912 ===================
914 A kernel service may want to define its own key type. For instance, an AFS
915 filesystem might want to define a Kerberos 5 ticket key type. To do this, it
916 author fills in a struct key_type and registers it with the system.
918 The structure has a number of fields, some of which are mandatory:
920 (*) const char *name
922 The name of the key type. This is used to translate a key type name
923 supplied by userspace into a pointer to the structure.
926 (*) size_t def_datalen
928 This is optional - it supplies the default payload data length as
929 contributed to the quota. If the key type's payload is always or almost
930 always the same size, then this is a more efficient way to do things.
932 The data length (and quota) on a particular key can always be changed
933 during instantiation or update by calling:
935 int key_payload_reserve(struct key *key, size_t datalen);
937 With the revised data length. Error EDQUOT will be returned if this is not
938 viable.
941 (*) int (*instantiate)(struct key *key, const void *data, size_t datalen);
943 This method is called to attach a payload to a key during construction.
944 The payload attached need not bear any relation to the data passed to this
945 function.
947 If the amount of data attached to the key differs from the size in
948 keytype->def_datalen, then key_payload_reserve() should be called.
950 This method does not have to lock the key in order to attach a payload.
951 The fact that KEY_FLAG_INSTANTIATED is not set in key->flags prevents
952 anything else from gaining access to the key.
954 It is safe to sleep in this method.
957 (*) int (*update)(struct key *key, const void *data, size_t datalen);
959 If this type of key can be updated, then this method should be provided.
960 It is called to update a key's payload from the blob of data provided.
962 key_payload_reserve() should be called if the data length might change
963 before any changes are actually made. Note that if this succeeds, the type
964 is committed to changing the key because it's already been altered, so all
965 memory allocation must be done first.
967 The key will have its semaphore write-locked before this method is called,
968 but this only deters other writers; any changes to the key's payload must
969 be made under RCU conditions, and call_rcu() must be used to dispose of
970 the old payload.
972 key_payload_reserve() should be called before the changes are made, but
973 after all allocations and other potentially failing function calls are
974 made.
976 It is safe to sleep in this method.
979 (*) int (*match)(const struct key *key, const void *desc);
981 This method is called to match a key against a description. It should
982 return non-zero if the two match, zero if they don't.
984 This method should not need to lock the key in any way. The type and
985 description can be considered invariant, and the payload should not be
986 accessed (the key may not yet be instantiated).
988 It is not safe to sleep in this method; the caller may hold spinlocks.
991 (*) void (*revoke)(struct key *key);
993 This method is optional. It is called to discard part of the payload
994 data upon a key being revoked. The caller will have the key semaphore
995 write-locked.
997 It is safe to sleep in this method, though care should be taken to avoid
998 a deadlock against the key semaphore.
1001 (*) void (*destroy)(struct key *key);
1003 This method is optional. It is called to discard the payload data on a key
1004 when it is being destroyed.
1006 This method does not need to lock the key to access the payload; it can
1007 consider the key as being inaccessible at this time. Note that the key's
1008 type may have been changed before this function is called.
1010 It is not safe to sleep in this method; the caller may hold spinlocks.
1013 (*) void (*describe)(const struct key *key, struct seq_file *p);
1015 This method is optional. It is called during /proc/keys reading to
1016 summarise a key's description and payload in text form.
1018 This method will be called with the RCU read lock held. rcu_dereference()
1019 should be used to read the payload pointer if the payload is to be
1020 accessed. key->datalen cannot be trusted to stay consistent with the
1021 contents of the payload.
1023 The description will not change, though the key's state may.
1025 It is not safe to sleep in this method; the RCU read lock is held by the
1026 caller.
1029 (*) long (*read)(const struct key *key, char __user *buffer, size_t buflen);
1031 This method is optional. It is called by KEYCTL_READ to translate the
1032 key's payload into something a blob of data for userspace to deal with.
1033 Ideally, the blob should be in the same format as that passed in to the
1034 instantiate and update methods.
1036 If successful, the blob size that could be produced should be returned
1037 rather than the size copied.
1039 This method will be called with the key's semaphore read-locked. This will
1040 prevent the key's payload changing. It is not necessary to use RCU locking
1041 when accessing the key's payload. It is safe to sleep in this method, such
1042 as might happen when the userspace buffer is accessed.
1045 (*) int (*request_key)(struct key *key, struct key *authkey, const char *op,
1046 void *aux);
1048 This method is optional. If provided, request_key() and
1049 request_key_with_auxdata() will invoke this function rather than
1050 upcalling to /sbin/request-key to operate upon a key of this type.
1052 The aux parameter is as passed to request_key_with_auxdata() or is NULL
1053 otherwise. Also passed are the key to be operated upon, the
1054 authorisation key for this operation and the operation type (currently
1055 only "create").
1057 This function should return only when the upcall is complete. Upon return
1058 the authorisation key will be revoked, and the target key will be
1059 negatively instantiated if it is still uninstantiated. The error will be
1060 returned to the caller of request_key*().
1063 ============================
1065 ============================
1067 To create a new key, the kernel will attempt to execute the following command
1068 line:
1070 /sbin/request-key create <key> <uid> <gid> \
1071 <threadring> <processring> <sessionring> <callout_info>
1073 <key> is the key being constructed, and the three keyrings are the process
1074 keyrings from the process that caused the search to be issued. These are
1075 included for two reasons:
1077 (1) There may be an authentication token in one of the keyrings that is
1078 required to obtain the key, eg: a Kerberos Ticket-Granting Ticket.
1080 (2) The new key should probably be cached in one of these rings.
1082 This program should set it UID and GID to those specified before attempting to
1083 access any more keys. It may then look around for a user specific process to
1084 hand the request off to (perhaps a path held in placed in another key by, for
1085 example, the KDE desktop manager).
1087 The program (or whatever it calls) should finish construction of the key by
1088 calling KEYCTL_INSTANTIATE, which also permits it to cache the key in one of
1089 the keyrings (probably the session ring) before returning. Alternatively, the
1090 key can be marked as negative with KEYCTL_NEGATE; this also permits the key to
1091 be cached in one of the keyrings.
1093 If it returns with the key remaining in the unconstructed state, the key will
1094 be marked as being negative, it will be added to the session keyring, and an
1095 error will be returned to the key requestor.
1097 Supplementary information may be provided from whoever or whatever invoked this
1098 service. This will be passed as the <callout_info> parameter. If no such
1099 information was made available, then "-" will be passed as this parameter
1100 instead.
1103 Similarly, the kernel may attempt to update an expired or a soon to expire key
1104 by executing:
1106 /sbin/request-key update <key> <uid> <gid> \
1107 <threadring> <processring> <sessionring>
1109 In this case, the program isn't required to actually attach the key to a ring;
1110 the rings are provided for reference.