1 ============================ 2 KERNEL 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 - Key access filesystem 30 31 32============ 33KEY OVERVIEW 34============ 35 36In this context, keys represent units of cryptographic data, authentication 37tokens, keyrings, etc.. These are represented in the kernel by struct key. 38 39Each key has a number of attributes: 40 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. 48 49 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. 53 54 Userspace programs can use a key's serial numbers as a way to gain access 55 to it, subject to permission checking. 56 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. 60 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. 63 64 Should a type be removed from the system, all the keys of that type will 65 be invalidated. 66 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. 70 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. 74 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. 77 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. 82 83 Having a payload is not required; and the payload can, in fact, just be a 84 value stored in the struct key itself. 85 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. 89 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. 93 94 (*) Each key can be in one of a number of basic states: 95 96 (*) Uninstantiated. The key exists, but does not have any data attached. 97 Keys being requested from userspace will be in this state. 98 99 (*) Instantiated. This is the normal state. The key is fully formed, and 100 has data attached. 101 102 (*) 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. 106 107 (*) 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. 110 111 (*) 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). 113 114 (*) Dead. The key's type was unregistered, and so the key is now useless. 115 116 117==================== 118KEY SERVICE OVERVIEW 119==================== 120 121The key service provides a number of features besides keys: 122 123 (*) The key service defines two special key types: 124 125 (+) "keyring" 126 127 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. 130 131 (+) "user" 132 133 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. 136 137 (*) Each process subscribes to three keyrings: a thread-specific keyring, a 138 process-specific keyring, and a session-specific keyring. 139 140 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. 143 144 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. 148 149 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. 154 155 The ownership of the thread keyring changes when the real UID and GID of 156 the thread changes. 157 158 (*) 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. 161 162 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. 164 165 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. 167 168 (*) 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. 171 172 The user can view information on this and other statistics through procfs 173 files. The root user may also alter the quota limits through sysctl files 174 (see the section "New procfs files"). 175 176 Process-specific and thread-specific keyrings are not counted towards a 177 user's quota. 178 179 If a system call that modifies a key or keyring in some way would put the 180 user over quota, the operation is refused and error EDQUOT is returned. 181 182 (*) There's a system call interface by which userspace programs can create and 183 manipulate keys and keyrings. 184 185 (*) There's a kernel interface by which services can register types and search 186 for keys. 187 188 (*) There's a way for the a search done from the kernel to call back to 189 userspace to request a key that can't be found in a process's keyrings. 190 191 (*) An optional filesystem is available through which the key database can be 192 viewed and manipulated. 193 194 195====================== 196KEY ACCESS PERMISSIONS 197====================== 198 199Keys have an owner user ID, a group access ID, and a permissions mask. The mask 200has up to eight bits each for possessor, user, group and other access. Only 201six of each set of eight bits are defined. These permissions granted are: 202 203 (*) View 204 205 This permits a key or keyring's attributes to be viewed - including key 206 type and description. 207 208 (*) Read 209 210 This permits a key's payload to be viewed or a keyring's list of linked 211 keys. 212 213 (*) Write 214 215 This permits a key's payload to be instantiated or updated, or it allows a 216 link to be added to or removed from a keyring. 217 218 (*) Search 219 220 This permits keyrings to be searched and keys to be found. Searches can 221 only recurse into nested keyrings that have search permission set. 222 223 (*) Link 224 225 This permits a key or keyring to be linked to. To create a link from a 226 keyring to a key, a process must have Write permission on the keyring and 227 Link permission on the key. 228 229 (*) Set Attribute 230 231 This permits a key's UID, GID and permissions mask to be changed. 232 233For changing the ownership, group ID or permissions mask, being the owner of 234the key or having the sysadmin capability is sufficient. 235 236 237=============== 238SELINUX SUPPORT 239=============== 240 241The security class "key" has been added to SELinux so that mandatory access 242controls can be applied to keys created within various contexts. This support 243is preliminary, and is likely to change quite significantly in the near future. 244Currently, all of the basic permissions explained above are provided in SELinux 245as well; SELinux is simply invoked after all basic permission checks have been 246performed. 247 248The value of the file /proc/self/attr/keycreate influences the labeling of 249newly-created keys. If the contents of that file correspond to an SELinux 250security context, then the key will be assigned that context. Otherwise, the 251key will be assigned the current context of the task that invoked the key 252creation request. Tasks must be granted explicit permission to assign a 253particular context to newly-created keys, using the "create" permission in the 254key security class. 255 256The default keyrings associated with users will be labeled with the default 257context of the user if and only if the login programs have been instrumented to 258properly initialize keycreate during the login process. Otherwise, they will 259be labeled with the context of the login program itself. 260 261Note, however, that the default keyrings associated with the root user are 262labeled with the default kernel context, since they are created early in the 263boot process, before root has a chance to log in. 264 265The keyrings associated with new threads are each labeled with the context of 266their associated thread, and both session and process keyrings are handled 267similarly. 268 269 270================ 271NEW PROCFS FILES 272================ 273 274Two files have been added to procfs by which an administrator can find out 275about the status of the key service: 276 277 (*) /proc/keys 278 279 This lists the keys that are currently viewable by the task reading the 280 file, giving information about their type, description and permissions. 281 It is not possible to view the payload of the key this way, though some 282 information about it may be given. 283 284 The only keys included in the list are those that grant View permission to 285 the reading process whether or not it possesses them. Note that LSM 286 security checks are still performed, and may further filter out keys that 287 the current process is not authorised to view. 288 289 The contents of the file look like this: 290 291 SERIAL FLAGS USAGE EXPY PERM UID GID TYPE DESCRIPTION: SUMMARY 292 00000001 I----- 39 perm 1f3f0000 0 0 keyring _uid_ses.0: 1/4 293 00000002 I----- 2 perm 1f3f0000 0 0 keyring _uid.0: empty 294 00000007 I----- 1 perm 1f3f0000 0 0 keyring _pid.1: empty 295 0000018d I----- 1 perm 1f3f0000 0 0 keyring _pid.412: empty 296 000004d2 I--Q-- 1 perm 1f3f0000 32 -1 keyring _uid.32: 1/4 297 000004d3 I--Q-- 3 perm 1f3f0000 32 -1 keyring _uid_ses.32: empty 298 00000892 I--QU- 1 perm 1f000000 0 0 user metal:copper: 0 299 00000893 I--Q-N 1 35s 1f3f0000 0 0 user metal:silver: 0 300 00000894 I--Q-- 1 10h 003f0000 0 0 user metal:gold: 0 301 302 The flags are: 303 304 I Instantiated 305 R Revoked 306 D Dead 307 Q Contributes to user's quota 308 U Under construction by callback to userspace 309 N Negative key 310 311 This file must be enabled at kernel configuration time as it allows anyone 312 to list the keys database. 313 314 (*) /proc/key-users 315 316 This file lists the tracking data for each user that has at least one key 317 on the system. Such data includes quota information and statistics: 318 319 [root@andromeda root]# cat /proc/key-users 320 0: 46 45/45 1/100 13/10000 321 29: 2 2/2 2/100 40/10000 322 32: 2 2/2 2/100 40/10000 323 38: 2 2/2 2/100 40/10000 324 325 The format of each line is 326 <UID>: User ID to which this applies 327 <usage> Structure refcount 328 <inst>/<keys> Total number of keys and number instantiated 329 <keys>/<max> Key count quota 330 <bytes>/<max> Key size quota 331 332 333Four new sysctl files have been added also for the purpose of controlling the 334quota limits on keys: 335 336 (*) /proc/sys/kernel/keys/root_maxkeys 337 /proc/sys/kernel/keys/root_maxbytes 338 339 These files hold the maximum number of keys that root may have and the 340 maximum total number of bytes of data that root may have stored in those 341 keys. 342 343 (*) /proc/sys/kernel/keys/maxkeys 344 /proc/sys/kernel/keys/maxbytes 345 346 These files hold the maximum number of keys that each non-root user may 347 have and the maximum total number of bytes of data that each of those 348 users may have stored in their keys. 349 350Root may alter these by writing each new limit as a decimal number string to 351the appropriate file. 352 353 354=============================== 355USERSPACE SYSTEM CALL INTERFACE 356=============================== 357 358Userspace can manipulate keys directly through three new syscalls: add_key, 359request_key and keyctl. The latter provides a number of functions for 360manipulating keys. 361 362When referring to a key directly, userspace programs should use the key's 363serial number (a positive 32-bit integer). However, there are some special 364values available for referring to special keys and keyrings that relate to the 365process making the call: 366 367 CONSTANT VALUE KEY REFERENCED 368 ============================== ====== =========================== 369 KEY_SPEC_THREAD_KEYRING -1 thread-specific keyring 370 KEY_SPEC_PROCESS_KEYRING -2 process-specific keyring 371 KEY_SPEC_SESSION_KEYRING -3 session-specific keyring 372 KEY_SPEC_USER_KEYRING -4 UID-specific keyring 373 KEY_SPEC_USER_SESSION_KEYRING -5 UID-session keyring 374 KEY_SPEC_GROUP_KEYRING -6 GID-specific keyring 375 KEY_SPEC_REQKEY_AUTH_KEY -7 assumed request_key() 376 authorisation key 377 378 379The main syscalls are: 380 381 (*) Create a new key of given type, description and payload and add it to the 382 nominated keyring: 383 384 key_serial_t add_key(const char *type, const char *desc, 385 const void *payload, size_t plen, 386 key_serial_t keyring); 387 388 If a key of the same type and description as that proposed already exists 389 in the keyring, this will try to update it with the given payload, or it 390 will return error EEXIST if that function is not supported by the key 391 type. The process must also have permission to write to the key to be able 392 to update it. The new key will have all user permissions granted and no 393 group or third party permissions. 394 395 Otherwise, this will attempt to create a new key of the specified type and 396 description, and to instantiate it with the supplied payload and attach it 397 to the keyring. In this case, an error will be generated if the process 398 does not have permission to write to the keyring. 399 400 The payload is optional, and the pointer can be NULL if not required by 401 the type. The payload is plen in size, and plen can be zero for an empty 402 payload. 403 404 A new keyring can be generated by setting type "keyring", the keyring name 405 as the description (or NULL) and setting the payload to NULL. 406 407 User defined keys can be created by specifying type "user". It is 408 recommended that a user defined key's description by prefixed with a type 409 ID and a colon, such as "krb5tgt:" for a Kerberos 5 ticket granting 410 ticket. 411 412 Any other type must have been registered with the kernel in advance by a 413 kernel service such as a filesystem. 414 415 The ID of the new or updated key is returned if successful. 416 417 418 (*) Search the process's keyrings for a key, potentially calling out to 419 userspace to create it. 420 421 key_serial_t request_key(const char *type, const char *description, 422 const char *callout_info, 423 key_serial_t dest_keyring); 424 425 This function searches all the process's keyrings in the order thread, 426 process, session for a matching key. This works very much like 427 KEYCTL_SEARCH, including the optional attachment of the discovered key to 428 a keyring. 429 430 If a key cannot be found, and if callout_info is not NULL, then 431 /sbin/request-key will be invoked in an attempt to obtain a key. The 432 callout_info string will be passed as an argument to the program. 433 434 See also Documentation/keys-request-key.txt. 435 436 437The keyctl syscall functions are: 438 439 (*) Map a special key ID to a real key ID for this process: 440 441 key_serial_t keyctl(KEYCTL_GET_KEYRING_ID, key_serial_t id, 442 int create); 443 444 The special key specified by "id" is looked up (with the key being created 445 if necessary) and the ID of the key or keyring thus found is returned if 446 it exists. 447 448 If the key does not yet exist, the key will be created if "create" is 449 non-zero; and the error ENOKEY will be returned if "create" is zero. 450 451 452 (*) Replace the session keyring this process subscribes to with a new one: 453 454 key_serial_t keyctl(KEYCTL_JOIN_SESSION_KEYRING, const char *name); 455 456 If name is NULL, an anonymous keyring is created attached to the process 457 as its session keyring, displacing the old session keyring. 458 459 If name is not NULL, if a keyring of that name exists, the process 460 attempts to attach it as the session keyring, returning an error if that 461 is not permitted; otherwise a new keyring of that name is created and 462 attached as the session keyring. 463 464 To attach to a named keyring, the keyring must have search permission for 465 the process's ownership. 466 467 The ID of the new session keyring is returned if successful. 468 469 470 (*) Update the specified key: 471 472 long keyctl(KEYCTL_UPDATE, key_serial_t key, const void *payload, 473 size_t plen); 474 475 This will try to update the specified key with the given payload, or it 476 will return error EOPNOTSUPP if that function is not supported by the key 477 type. The process must also have permission to write to the key to be able 478 to update it. 479 480 The payload is of length plen, and may be absent or empty as for 481 add_key(). 482 483 484 (*) Revoke a key: 485 486 long keyctl(KEYCTL_REVOKE, key_serial_t key); 487 488 This makes a key unavailable for further operations. Further attempts to 489 use the key will be met with error EKEYREVOKED, and the key will no longer 490 be findable. 491 492 493 (*) Change the ownership of a key: 494 495 long keyctl(KEYCTL_CHOWN, key_serial_t key, uid_t uid, gid_t gid); 496 497 This function permits a key's owner and group ID to be changed. Either one 498 of uid or gid can be set to -1 to suppress that change. 499 500 Only the superuser can change a key's owner to something other than the 501 key's current owner. Similarly, only the superuser can change a key's 502 group ID to something other than the calling process's group ID or one of 503 its group list members. 504 505 506 (*) Change the permissions mask on a key: 507 508 long keyctl(KEYCTL_SETPERM, key_serial_t key, key_perm_t perm); 509 510 This function permits the owner of a key or the superuser to change the 511 permissions mask on a key. 512 513 Only bits the available bits are permitted; if any other bits are set, 514 error EINVAL will be returned. 515 516 517 (*) Describe a key: 518 519 long keyctl(KEYCTL_DESCRIBE, key_serial_t key, char *buffer, 520 size_t buflen); 521 522 This function returns a summary of the key's attributes (but not its 523 payload data) as a string in the buffer provided. 524 525 Unless there's an error, it always returns the amount of data it could 526 produce, even if that's too big for the buffer, but it won't copy more 527 than requested to userspace. If the buffer pointer is NULL then no copy 528 will take place. 529 530 A process must have view permission on the key for this function to be 531 successful. 532 533 If successful, a string is placed in the buffer in the following format: 534 535 <type>;<uid>;<gid>;<perm>;<description> 536 537 Where type and description are strings, uid and gid are decimal, and perm 538 is hexadecimal. A NUL character is included at the end of the string if 539 the buffer is sufficiently big. 540 541 This can be parsed with 542 543 sscanf(buffer, "%[^;];%d;%d;%o;%s", type, &uid, &gid, &mode, desc); 544 545 546 (*) Clear out a keyring: 547 548 long keyctl(KEYCTL_CLEAR, key_serial_t keyring); 549 550 This function clears the list of keys attached to a keyring. The calling 551 process must have write permission on the keyring, and it must be a 552 keyring (or else error ENOTDIR will result). 553 554 555 (*) Link a key into a keyring: 556 557 long keyctl(KEYCTL_LINK, key_serial_t keyring, key_serial_t key); 558 559 This function creates a link from the keyring to the key. The process must 560 have write permission on the keyring and must have link permission on the 561 key. 562 563 Should the keyring not be a keyring, error ENOTDIR will result; and if the 564 keyring is full, error ENFILE will result. 565 566 The link procedure checks the nesting of the keyrings, returning ELOOP if 567 it appears too deep or EDEADLK if the link would introduce a cycle. 568 569 Any links within the keyring to keys that match the new key in terms of 570 type and description will be discarded from the keyring as the new one is 571 added. 572 573 574 (*) Unlink a key or keyring from another keyring: 575 576 long keyctl(KEYCTL_UNLINK, key_serial_t keyring, key_serial_t key); 577 578 This function looks through the keyring for the first link to the 579 specified key, and removes it if found. Subsequent links to that key are 580 ignored. The process must have write permission on the keyring. 581 582 If the keyring is not a keyring, error ENOTDIR will result; and if the key 583 is not present, error ENOENT will be the result. 584 585 586 (*) Search a keyring tree for a key: 587 588 key_serial_t keyctl(KEYCTL_SEARCH, key_serial_t keyring, 589 const char *type, const char *description, 590 key_serial_t dest_keyring); 591 592 This searches the keyring tree headed by the specified keyring until a key 593 is found that matches the type and description criteria. Each keyring is 594 checked for keys before recursion into its children occurs. 595 596 The process must have search permission on the top level keyring, or else 597 error EACCES will result. Only keyrings that the process has search 598 permission on will be recursed into, and only keys and keyrings for which 599 a process has search permission can be matched. If the specified keyring 600 is not a keyring, ENOTDIR will result. 601 602 If the search succeeds, the function will attempt to link the found key 603 into the destination keyring if one is supplied (non-zero ID). All the 604 constraints applicable to KEYCTL_LINK apply in this case too. 605 606 Error ENOKEY, EKEYREVOKED or EKEYEXPIRED will be returned if the search 607 fails. On success, the resulting key ID will be returned. 608 609 610 (*) Read the payload data from a key: 611 612 long keyctl(KEYCTL_READ, key_serial_t keyring, char *buffer, 613 size_t buflen); 614 615 This function attempts to read the payload data from the specified key 616 into the buffer. The process must have read permission on the key to 617 succeed. 618 619 The returned data will be processed for presentation by the key type. For 620 instance, a keyring will return an array of key_serial_t entries 621 representing the IDs of all the keys to which it is subscribed. The user 622 defined key type will return its data as is. If a key type does not 623 implement this function, error EOPNOTSUPP will result. 624 625 As much of the data as can be fitted into the buffer will be copied to 626 userspace if the buffer pointer is not NULL. 627 628 On a successful return, the function will always return the amount of data 629 available rather than the amount copied. 630 631 632 (*) Instantiate a partially constructed key. 633 634 long keyctl(KEYCTL_INSTANTIATE, key_serial_t key, 635 const void *payload, size_t plen, 636 key_serial_t keyring); 637 638 If the kernel calls back to userspace to complete the instantiation of a 639 key, userspace should use this call to supply data for the key before the 640 invoked process returns, or else the key will be marked negative 641 automatically. 642 643 The process must have write access on the key to be able to instantiate 644 it, and the key must be uninstantiated. 645 646 If a keyring is specified (non-zero), the key will also be linked into 647 that keyring, however all the constraints applying in KEYCTL_LINK apply in 648 this case too. 649 650 The payload and plen arguments describe the payload data as for add_key(). 651 652 653 (*) Negatively instantiate a partially constructed key. 654 655 long keyctl(KEYCTL_NEGATE, key_serial_t key, 656 unsigned timeout, key_serial_t keyring); 657 658 If the kernel calls back to userspace to complete the instantiation of a 659 key, userspace should use this call mark the key as negative before the 660 invoked process returns if it is unable to fulfil the request. 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. 668 669 670 (*) Set the default request-key destination keyring. 671 672 long keyctl(KEYCTL_SET_REQKEY_KEYRING, int reqkey_defl); 673 674 This sets the default keyring to which implicitly requested keys will be 675 attached for this thread. reqkey_defl should be one of these constants: 676 677 CONSTANT VALUE NEW DEFAULT KEYRING 678 ====================================== ====== ======================= 679 KEY_REQKEY_DEFL_NO_CHANGE -1 No change 680 KEY_REQKEY_DEFL_DEFAULT 0 Default[1] 681 KEY_REQKEY_DEFL_THREAD_KEYRING 1 Thread keyring 682 KEY_REQKEY_DEFL_PROCESS_KEYRING 2 Process keyring 683 KEY_REQKEY_DEFL_SESSION_KEYRING 3 Session keyring 684 KEY_REQKEY_DEFL_USER_KEYRING 4 User keyring 685 KEY_REQKEY_DEFL_USER_SESSION_KEYRING 5 User session keyring 686 KEY_REQKEY_DEFL_GROUP_KEYRING 6 Group keyring 687 688 The old default will be returned if successful and error EINVAL will be 689 returned if reqkey_defl is not one of the above values. 690 691 The default keyring can be overridden by the keyring indicated to the 692 request_key() system call. 693 694 Note that this setting is inherited across fork/exec. 695 696 [1] The default is: the thread keyring if there is one, otherwise 697 the process keyring if there is one, otherwise the session keyring if 698 there is one, otherwise the user default session keyring. 699 700 701 (*) Set the timeout on a key. 702 703 long keyctl(KEYCTL_SET_TIMEOUT, key_serial_t key, unsigned timeout); 704 705 This sets or clears the timeout on a key. The timeout can be 0 to clear 706 the timeout or a number of seconds to set the expiry time that far into 707 the future. 708 709 The process must have attribute modification access on a key to set its 710 timeout. Timeouts may not be set with this function on negative, revoked 711 or expired keys. 712 713 714 (*) Assume the authority granted to instantiate a key 715 716 long keyctl(KEYCTL_ASSUME_AUTHORITY, key_serial_t key); 717 718 This assumes or divests the authority required to instantiate the 719 specified key. Authority can only be assumed if the thread has the 720 authorisation key associated with the specified key in its keyrings 721 somewhere. 722 723 Once authority is assumed, searches for keys will also search the 724 requester's keyrings using the requester's security label, UID, GID and 725 groups. 726 727 If the requested authority is unavailable, error EPERM will be returned, 728 likewise if the authority has been revoked because the target key is 729 already instantiated. 730 731 If the specified key is 0, then any assumed authority will be divested. 732 733 The assumed authoritative key is inherited across fork and exec. 734 735 736 (*) Get the LSM security context attached to a key. 737 738 long keyctl(KEYCTL_GET_SECURITY, key_serial_t key, char *buffer, 739 size_t buflen) 740 741 This function returns a string that represents the LSM security context 742 attached to a key in the buffer provided. 743 744 Unless there's an error, it always returns the amount of data it could 745 produce, even if that's too big for the buffer, but it won't copy more 746 than requested to userspace. If the buffer pointer is NULL then no copy 747 will take place. 748 749 A NUL character is included at the end of the string if the buffer is 750 sufficiently big. This is included in the returned count. If no LSM is 751 in force then an empty string will be returned. 752 753 A process must have view permission on the key for this function to be 754 successful. 755 756 757=============== 758KERNEL SERVICES 759=============== 760 761The kernel services for key management are fairly simple to deal with. They can 762be broken down into two areas: keys and key types. 763 764Dealing with keys is fairly straightforward. Firstly, the kernel service 765registers its type, then it searches for a key of that type. It should retain 766the key as long as it has need of it, and then it should release it. For a 767filesystem or device file, a search would probably be performed during the open 768call, and the key released upon close. How to deal with conflicting keys due to 769two different users opening the same file is left to the filesystem author to 770solve. 771 772To access the key manager, the following header must be #included: 773 774 <linux/key.h> 775 776Specific key types should have a header file under include/keys/ that should be 777used to access that type. For keys of type "user", for example, that would be: 778 779 <keys/user-type.h> 780 781Note that there are two different types of pointers to keys that may be 782encountered: 783 784 (*) struct key * 785 786 This simply points to the key structure itself. Key structures will be at 787 least four-byte aligned. 788 789 (*) key_ref_t 790 791 This is equivalent to a struct key *, but the least significant bit is set 792 if the caller "possesses" the key. By "possession" it is meant that the 793 calling processes has a searchable link to the key from one of its 794 keyrings. There are three functions for dealing with these: 795 796 key_ref_t make_key_ref(const struct key *key, 797 unsigned long possession); 798 799 struct key *key_ref_to_ptr(const key_ref_t key_ref); 800 801 unsigned long is_key_possessed(const key_ref_t key_ref); 802 803 The first function constructs a key reference from a key pointer and 804 possession information (which must be 0 or 1 and not any other value). 805 806 The second function retrieves the key pointer from a reference and the 807 third retrieves the possession flag. 808 809When accessing a key's payload contents, certain precautions must be taken to 810prevent access vs modification races. See the section "Notes on accessing 811payload contents" for more information. 812 813(*) To search for a key, call: 814 815 struct key *request_key(const struct key_type *type, 816 const char *description, 817 const char *callout_info); 818 819 This is used to request a key or keyring with a description that matches 820 the description specified according to the key type's match function. This 821 permits approximate matching to occur. If callout_string is not NULL, then 822 /sbin/request-key will be invoked in an attempt to obtain the key from 823 userspace. In that case, callout_string will be passed as an argument to 824 the program. 825 826 Should the function fail error ENOKEY, EKEYEXPIRED or EKEYREVOKED will be 827 returned. 828 829 If successful, the key will have been attached to the default keyring for 830 implicitly obtained request-key keys, as set by KEYCTL_SET_REQKEY_KEYRING. 831 832 See also Documentation/keys-request-key.txt. 833 834 835(*) To search for a key, passing auxiliary data to the upcaller, call: 836 837 struct key *request_key_with_auxdata(const struct key_type *type, 838 const char *description, 839 const void *callout_info, 840 size_t callout_len, 841 void *aux); 842 843 This is identical to request_key(), except that the auxiliary data is 844 passed to the key_type->request_key() op if it exists, and the callout_info 845 is a blob of length callout_len, if given (the length may be 0). 846 847 848(*) A key can be requested asynchronously by calling one of: 849 850 struct key *request_key_async(const struct key_type *type, 851 const char *description, 852 const void *callout_info, 853 size_t callout_len); 854 855 or: 856 857 struct key *request_key_async_with_auxdata(const struct key_type *type, 858 const char *description, 859 const char *callout_info, 860 size_t callout_len, 861 void *aux); 862 863 which are asynchronous equivalents of request_key() and 864 request_key_with_auxdata() respectively. 865 866 These two functions return with the key potentially still under 867 construction. To wait for construction completion, the following should be 868 called: 869 870 int wait_for_key_construction(struct key *key, bool intr); 871 872 The function will wait for the key to finish being constructed and then 873 invokes key_validate() to return an appropriate value to indicate the state 874 of the key (0 indicates the key is usable). 875 876 If intr is true, then the wait can be interrupted by a signal, in which 877 case error ERESTARTSYS will be returned. 878 879 880(*) When it is no longer required, the key should be released using: 881 882 void key_put(struct key *key); 883 884 Or: 885 886 void key_ref_put(key_ref_t key_ref); 887 888 These can be called from interrupt context. If CONFIG_KEYS is not set then 889 the argument will not be parsed. 890 891 892(*) Extra references can be made to a key by calling the following function: 893 894 struct key *key_get(struct key *key); 895 896 These need to be disposed of by calling key_put() when they've been 897 finished with. The key pointer passed in will be returned. If the pointer 898 is NULL or CONFIG_KEYS is not set then the key will not be dereferenced and 899 no increment will take place. 900 901 902(*) A key's serial number can be obtained by calling: 903 904 key_serial_t key_serial(struct key *key); 905 906 If key is NULL or if CONFIG_KEYS is not set then 0 will be returned (in the 907 latter case without parsing the argument). 908 909 910(*) If a keyring was found in the search, this can be further searched by: 911 912 key_ref_t keyring_search(key_ref_t keyring_ref, 913 const struct key_type *type, 914 const char *description) 915 916 This searches the keyring tree specified for a matching key. Error ENOKEY 917 is returned upon failure (use IS_ERR/PTR_ERR to determine). If successful, 918 the returned key will need to be released. 919 920 The possession attribute from the keyring reference is used to control 921 access through the permissions mask and is propagated to the returned key 922 reference pointer if successful. 923 924 925(*) To check the validity of a key, this function can be called: 926 927 int validate_key(struct key *key); 928 929 This checks that the key in question hasn't expired or and hasn't been 930 revoked. Should the key be invalid, error EKEYEXPIRED or EKEYREVOKED will 931 be returned. If the key is NULL or if CONFIG_KEYS is not set then 0 will be 932 returned (in the latter case without parsing the argument). 933 934 935(*) To register a key type, the following function should be called: 936 937 int register_key_type(struct key_type *type); 938 939 This will return error EEXIST if a type of the same name is already 940 present. 941 942 943(*) To unregister a key type, call: 944 945 void unregister_key_type(struct key_type *type); 946 947 948Under some circumstances, it may be desirable to deal with a bundle of keys. 949The facility provides access to the keyring type for managing such a bundle: 950 951 struct key_type key_type_keyring; 952 953This can be used with a function such as request_key() to find a specific 954keyring in a process's keyrings. A keyring thus found can then be searched 955with keyring_search(). Note that it is not possible to use request_key() to 956search a specific keyring, so using keyrings in this way is of limited utility. 957 958 959=================================== 960NOTES ON ACCESSING PAYLOAD CONTENTS 961=================================== 962 963The simplest payload is just a number in key->payload.value. In this case, 964there's no need to indulge in RCU or locking when accessing the payload. 965 966More complex payload contents must be allocated and a pointer to them set in 967key->payload.data. One of the following ways must be selected to access the 968data: 969 970 (1) Unmodifiable key type. 971 972 If the key type does not have a modify method, then the key's payload can 973 be accessed without any form of locking, provided that it's known to be 974 instantiated (uninstantiated keys cannot be "found"). 975 976 (2) The key's semaphore. 977 978 The semaphore could be used to govern access to the payload and to control 979 the payload pointer. It must be write-locked for modifications and would 980 have to be read-locked for general access. The disadvantage of doing this 981 is that the accessor may be required to sleep. 982 983 (3) RCU. 984 985 RCU must be used when the semaphore isn't already held; if the semaphore 986 is held then the contents can't change under you unexpectedly as the 987 semaphore must still be used to serialise modifications to the key. The 988 key management code takes care of this for the key type. 989 990 However, this means using: 991 992 rcu_read_lock() ... rcu_dereference() ... rcu_read_unlock() 993 994 to read the pointer, and: 995 996 rcu_dereference() ... rcu_assign_pointer() ... call_rcu() 997 998 to set the pointer and dispose of the old contents after a grace period. 999 Note that only the key type should ever modify a key's payload. 1000 1001 Furthermore, an RCU controlled payload must hold a struct rcu_head for the 1002 use of call_rcu() and, if the payload is of variable size, the length of 1003 the payload. key->datalen cannot be relied upon to be consistent with the 1004 payload just dereferenced if the key's semaphore is not held. 1005 1006 1007=================== 1008DEFINING A KEY TYPE 1009=================== 1010 1011A kernel service may want to define its own key type. For instance, an AFS 1012filesystem might want to define a Kerberos 5 ticket key type. To do this, it 1013author fills in a key_type struct and registers it with the system. 1014 1015Source files that implement key types should include the following header file: 1016 1017 <linux/key-type.h> 1018 1019The structure has a number of fields, some of which are mandatory: 1020 1021 (*) const char *name 1022 1023 The name of the key type. This is used to translate a key type name 1024 supplied by userspace into a pointer to the structure. 1025 1026 1027 (*) size_t def_datalen 1028 1029 This is optional - it supplies the default payload data length as 1030 contributed to the quota. If the key type's payload is always or almost 1031 always the same size, then this is a more efficient way to do things. 1032 1033 The data length (and quota) on a particular key can always be changed 1034 during instantiation or update by calling: 1035 1036 int key_payload_reserve(struct key *key, size_t datalen); 1037 1038 With the revised data length. Error EDQUOT will be returned if this is not 1039 viable. 1040 1041 1042 (*) int (*instantiate)(struct key *key, const void *data, size_t datalen); 1043 1044 This method is called to attach a payload to a key during construction. 1045 The payload attached need not bear any relation to the data passed to this 1046 function. 1047 1048 If the amount of data attached to the key differs from the size in 1049 keytype->def_datalen, then key_payload_reserve() should be called. 1050 1051 This method does not have to lock the key in order to attach a payload. 1052 The fact that KEY_FLAG_INSTANTIATED is not set in key->flags prevents 1053 anything else from gaining access to the key. 1054 1055 It is safe to sleep in this method. 1056 1057 1058 (*) int (*update)(struct key *key, const void *data, size_t datalen); 1059 1060 If this type of key can be updated, then this method should be provided. 1061 It is called to update a key's payload from the blob of data provided. 1062 1063 key_payload_reserve() should be called if the data length might change 1064 before any changes are actually made. Note that if this succeeds, the type 1065 is committed to changing the key because it's already been altered, so all 1066 memory allocation must be done first. 1067 1068 The key will have its semaphore write-locked before this method is called, 1069 but this only deters other writers; any changes to the key's payload must 1070 be made under RCU conditions, and call_rcu() must be used to dispose of 1071 the old payload. 1072 1073 key_payload_reserve() should be called before the changes are made, but 1074 after all allocations and other potentially failing function calls are 1075 made. 1076 1077 It is safe to sleep in this method. 1078 1079 1080 (*) int (*match)(const struct key *key, const void *desc); 1081 1082 This method is called to match a key against a description. It should 1083 return non-zero if the two match, zero if they don't. 1084 1085 This method should not need to lock the key in any way. The type and 1086 description can be considered invariant, and the payload should not be 1087 accessed (the key may not yet be instantiated). 1088 1089 It is not safe to sleep in this method; the caller may hold spinlocks. 1090 1091 1092 (*) void (*revoke)(struct key *key); 1093 1094 This method is optional. It is called to discard part of the payload 1095 data upon a key being revoked. The caller will have the key semaphore 1096 write-locked. 1097 1098 It is safe to sleep in this method, though care should be taken to avoid 1099 a deadlock against the key semaphore. 1100 1101 1102 (*) void (*destroy)(struct key *key); 1103 1104 This method is optional. It is called to discard the payload data on a key 1105 when it is being destroyed. 1106 1107 This method does not need to lock the key to access the payload; it can 1108 consider the key as being inaccessible at this time. Note that the key's 1109 type may have been changed before this function is called. 1110 1111 It is not safe to sleep in this method; the caller may hold spinlocks. 1112 1113 1114 (*) void (*describe)(const struct key *key, struct seq_file *p); 1115 1116 This method is optional. It is called during /proc/keys reading to 1117 summarise a key's description and payload in text form. 1118 1119 This method will be called with the RCU read lock held. rcu_dereference() 1120 should be used to read the payload pointer if the payload is to be 1121 accessed. key->datalen cannot be trusted to stay consistent with the 1122 contents of the payload. 1123 1124 The description will not change, though the key's state may. 1125 1126 It is not safe to sleep in this method; the RCU read lock is held by the 1127 caller. 1128 1129 1130 (*) long (*read)(const struct key *key, char __user *buffer, size_t buflen); 1131 1132 This method is optional. It is called by KEYCTL_READ to translate the 1133 key's payload into something a blob of data for userspace to deal with. 1134 Ideally, the blob should be in the same format as that passed in to the 1135 instantiate and update methods. 1136 1137 If successful, the blob size that could be produced should be returned 1138 rather than the size copied. 1139 1140 This method will be called with the key's semaphore read-locked. This will 1141 prevent the key's payload changing. It is not necessary to use RCU locking 1142 when accessing the key's payload. It is safe to sleep in this method, such 1143 as might happen when the userspace buffer is accessed. 1144 1145 1146 (*) int (*request_key)(struct key_construction *cons, const char *op, 1147 void *aux); 1148 1149 This method is optional. If provided, request_key() and friends will 1150 invoke this function rather than upcalling to /sbin/request-key to operate 1151 upon a key of this type. 1152 1153 The aux parameter is as passed to request_key_async_with_auxdata() and 1154 similar or is NULL otherwise. Also passed are the construction record for 1155 the key to be operated upon and the operation type (currently only 1156 "create"). 1157 1158 This method is permitted to return before the upcall is complete, but the 1159 following function must be called under all circumstances to complete the 1160 instantiation process, whether or not it succeeds, whether or not there's 1161 an error: 1162 1163 void complete_request_key(struct key_construction *cons, int error); 1164 1165 The error parameter should be 0 on success, -ve on error. The 1166 construction record is destroyed by this action and the authorisation key 1167 will be revoked. If an error is indicated, the key under construction 1168 will be negatively instantiated if it wasn't already instantiated. 1169 1170 If this method returns an error, that error will be returned to the 1171 caller of request_key*(). complete_request_key() must be called prior to 1172 returning. 1173 1174 The key under construction and the authorisation key can be found in the 1175 key_construction struct pointed to by cons: 1176 1177 (*) struct key *key; 1178 1179 The key under construction. 1180 1181 (*) struct key *authkey; 1182 1183 The authorisation key. 1184 1185 1186============================ 1187REQUEST-KEY CALLBACK SERVICE 1188============================ 1189 1190To create a new key, the kernel will attempt to execute the following command 1191line: 1192 1193 /sbin/request-key create <key> <uid> <gid> \ 1194 <threadring> <processring> <sessionring> <callout_info> 1195 1196<key> is the key being constructed, and the three keyrings are the process 1197keyrings from the process that caused the search to be issued. These are 1198included for two reasons: 1199 1200 (1) There may be an authentication token in one of the keyrings that is 1201 required to obtain the key, eg: a Kerberos Ticket-Granting Ticket. 1202 1203 (2) The new key should probably be cached in one of these rings. 1204 1205This program should set it UID and GID to those specified before attempting to 1206access any more keys. It may then look around for a user specific process to 1207hand the request off to (perhaps a path held in placed in another key by, for 1208example, the KDE desktop manager). 1209 1210The program (or whatever it calls) should finish construction of the key by 1211calling KEYCTL_INSTANTIATE, which also permits it to cache the key in one of 1212the keyrings (probably the session ring) before returning. Alternatively, the 1213key can be marked as negative with KEYCTL_NEGATE; this also permits the key to 1214be cached in one of the keyrings. 1215 1216If it returns with the key remaining in the unconstructed state, the key will 1217be marked as being negative, it will be added to the session keyring, and an 1218error will be returned to the key requestor. 1219 1220Supplementary information may be provided from whoever or whatever invoked this 1221service. This will be passed as the <callout_info> parameter. If no such 1222information was made available, then "-" will be passed as this parameter 1223instead. 1224 1225 1226Similarly, the kernel may attempt to update an expired or a soon to expire key 1227by executing: 1228 1229 /sbin/request-key update <key> <uid> <gid> \ 1230 <threadring> <processring> <sessionring> 1231 1232In this case, the program isn't required to actually attach the key to a ring; 1233the rings are provided for reference. 1234