request-key.rst 7.9 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207
  1. ===================
  2. Key Request Service
  3. ===================
  4. The key request service is part of the key retention service (refer to
  5. Documentation/security/keys/core.rst). This document explains more fully how
  6. the requesting algorithm works.
  7. The process starts by either the kernel requesting a service by calling
  8. ``request_key*()``::
  9. struct key *request_key(const struct key_type *type,
  10. const char *description,
  11. const char *callout_info);
  12. or::
  13. struct key *request_key_tag(const struct key_type *type,
  14. const char *description,
  15. const struct key_tag *domain_tag,
  16. const char *callout_info);
  17. or::
  18. struct key *request_key_with_auxdata(const struct key_type *type,
  19. const char *description,
  20. const struct key_tag *domain_tag,
  21. const char *callout_info,
  22. size_t callout_len,
  23. void *aux);
  24. or::
  25. struct key *request_key_rcu(const struct key_type *type,
  26. const char *description,
  27. const struct key_tag *domain_tag);
  28. Or by userspace invoking the request_key system call::
  29. key_serial_t request_key(const char *type,
  30. const char *description,
  31. const char *callout_info,
  32. key_serial_t dest_keyring);
  33. The main difference between the access points is that the in-kernel interface
  34. does not need to link the key to a keyring to prevent it from being immediately
  35. destroyed. The kernel interface returns a pointer directly to the key, and
  36. it's up to the caller to destroy the key.
  37. The request_key_tag() call is like the in-kernel request_key(), except that it
  38. also takes a domain tag that allows keys to be separated by namespace and
  39. killed off as a group.
  40. The request_key_with_auxdata() calls is like the request_key_tag() call, except
  41. that they permit auxiliary data to be passed to the upcaller (the default is
  42. NULL). This is only useful for those key types that define their own upcall
  43. mechanism rather than using /sbin/request-key.
  44. The request_key_rcu() call is like the request_key_tag() call, except that it
  45. doesn't check for keys that are under construction and doesn't attempt to
  46. construct missing keys.
  47. The userspace interface links the key to a keyring associated with the process
  48. to prevent the key from going away, and returns the serial number of the key to
  49. the caller.
  50. The following example assumes that the key types involved don't define their
  51. own upcall mechanisms. If they do, then those should be substituted for the
  52. forking and execution of /sbin/request-key.
  53. The Process
  54. ===========
  55. A request proceeds in the following manner:
  56. 1) Process A calls request_key() [the userspace syscall calls the kernel
  57. interface].
  58. 2) request_key() searches the process's subscribed keyrings to see if there's
  59. a suitable key there. If there is, it returns the key. If there isn't,
  60. and callout_info is not set, an error is returned. Otherwise the process
  61. proceeds to the next step.
  62. 3) request_key() sees that A doesn't have the desired key yet, so it creates
  63. two things:
  64. a) An uninstantiated key U of requested type and description.
  65. b) An authorisation key V that refers to key U and notes that process A
  66. is the context in which key U should be instantiated and secured, and
  67. from which associated key requests may be satisfied.
  68. 4) request_key() then forks and executes /sbin/request-key with a new session
  69. keyring that contains a link to auth key V.
  70. 5) /sbin/request-key assumes the authority associated with key U.
  71. 6) /sbin/request-key execs an appropriate program to perform the actual
  72. instantiation.
  73. 7) The program may want to access another key from A's context (say a
  74. Kerberos TGT key). It just requests the appropriate key, and the keyring
  75. search notes that the session keyring has auth key V in its bottom level.
  76. This will permit it to then search the keyrings of process A with the
  77. UID, GID, groups and security info of process A as if it was process A,
  78. and come up with key W.
  79. 8) The program then does what it must to get the data with which to
  80. instantiate key U, using key W as a reference (perhaps it contacts a
  81. Kerberos server using the TGT) and then instantiates key U.
  82. 9) Upon instantiating key U, auth key V is automatically revoked so that it
  83. may not be used again.
  84. 10) The program then exits 0 and request_key() deletes key V and returns key
  85. U to the caller.
  86. This also extends further. If key W (step 7 above) didn't exist, key W would
  87. be created uninstantiated, another auth key (X) would be created (as per step
  88. 3) and another copy of /sbin/request-key spawned (as per step 4); but the
  89. context specified by auth key X will still be process A, as it was in auth key
  90. V.
  91. This is because process A's keyrings can't simply be attached to
  92. /sbin/request-key at the appropriate places because (a) execve will discard two
  93. of them, and (b) it requires the same UID/GID/Groups all the way through.
  94. Negative Instantiation And Rejection
  95. ====================================
  96. Rather than instantiating a key, it is possible for the possessor of an
  97. authorisation key to negatively instantiate a key that's under construction.
  98. This is a short duration placeholder that causes any attempt at re-requesting
  99. the key while it exists to fail with error ENOKEY if negated or the specified
  100. error if rejected.
  101. This is provided to prevent excessive repeated spawning of /sbin/request-key
  102. processes for a key that will never be obtainable.
  103. Should the /sbin/request-key process exit anything other than 0 or die on a
  104. signal, the key under construction will be automatically negatively
  105. instantiated for a short amount of time.
  106. The Search Algorithm
  107. ====================
  108. A search of any particular keyring proceeds in the following fashion:
  109. 1) When the key management code searches for a key (keyring_search_rcu) it
  110. firstly calls key_permission(SEARCH) on the keyring it's starting with,
  111. if this denies permission, it doesn't search further.
  112. 2) It considers all the non-keyring keys within that keyring and, if any key
  113. matches the criteria specified, calls key_permission(SEARCH) on it to see
  114. if the key is allowed to be found. If it is, that key is returned; if
  115. not, the search continues, and the error code is retained if of higher
  116. priority than the one currently set.
  117. 3) It then considers all the keyring-type keys in the keyring it's currently
  118. searching. It calls key_permission(SEARCH) on each keyring, and if this
  119. grants permission, it recurses, executing steps (2) and (3) on that
  120. keyring.
  121. The process stops immediately a valid key is found with permission granted to
  122. use it. Any error from a previous match attempt is discarded and the key is
  123. returned.
  124. When request_key() is invoked, if CONFIG_KEYS_REQUEST_CACHE=y, a per-task
  125. one-key cache is first checked for a match.
  126. When search_process_keyrings() is invoked, it performs the following searches
  127. until one succeeds:
  128. 1) If extant, the process's thread keyring is searched.
  129. 2) If extant, the process's process keyring is searched.
  130. 3) The process's session keyring is searched.
  131. 4) If the process has assumed the authority associated with a request_key()
  132. authorisation key then:
  133. a) If extant, the calling process's thread keyring is searched.
  134. b) If extant, the calling process's process keyring is searched.
  135. c) The calling process's session keyring is searched.
  136. The moment one succeeds, all pending errors are discarded and the found key is
  137. returned. If CONFIG_KEYS_REQUEST_CACHE=y, then that key is placed in the
  138. per-task cache, displacing the previous key. The cache is cleared on exit or
  139. just prior to resumption of userspace.
  140. Only if all these fail does the whole thing fail with the highest priority
  141. error. Note that several errors may have come from LSM.
  142. The error priority is::
  143. EKEYREVOKED > EKEYEXPIRED > ENOKEY
  144. EACCES/EPERM are only returned on a direct search of a specific keyring where
  145. the basal keyring does not grant Search permission.