1 =================== 1 ===================
2 Key Request Service 2 Key Request Service
3 =================== 3 ===================
4 4
5 The key request service is part of the key ret 5 The key request service is part of the key retention service (refer to
6 Documentation/security/keys/core.rst). This d 6 Documentation/security/keys/core.rst). This document explains more fully how
7 the requesting algorithm works. 7 the requesting algorithm works.
8 8
9 The process starts by either the kernel reques 9 The process starts by either the kernel requesting a service by calling
10 ``request_key*()``:: 10 ``request_key*()``::
11 11
12 struct key *request_key(const struct k 12 struct key *request_key(const struct key_type *type,
13 const char *de 13 const char *description,
14 const char *ca 14 const char *callout_info);
15 15
16 or:: 16 or::
17 17
18 struct key *request_key_tag(const stru 18 struct key *request_key_tag(const struct key_type *type,
19 const char 19 const char *description,
20 const stru 20 const struct key_tag *domain_tag,
21 const char 21 const char *callout_info);
22 22
23 or:: 23 or::
24 24
25 struct key *request_key_with_auxdata(c 25 struct key *request_key_with_auxdata(const struct key_type *type,
26 c 26 const char *description,
27 c 27 const struct key_tag *domain_tag,
28 c 28 const char *callout_info,
29 s 29 size_t callout_len,
30 v 30 void *aux);
31 31
32 or:: 32 or::
33 33
34 struct key *request_key_rcu(const stru 34 struct key *request_key_rcu(const struct key_type *type,
35 const char 35 const char *description,
36 const stru 36 const struct key_tag *domain_tag);
37 37
38 Or by userspace invoking the request_key syste 38 Or by userspace invoking the request_key system call::
39 39
40 key_serial_t request_key(const char *t 40 key_serial_t request_key(const char *type,
41 const char *d 41 const char *description,
42 const char *c 42 const char *callout_info,
43 key_serial_t 43 key_serial_t dest_keyring);
44 44
45 The main difference between the access points 45 The main difference between the access points is that the in-kernel interface
46 does not need to link the key to a keyring to 46 does not need to link the key to a keyring to prevent it from being immediately
47 destroyed. The kernel interface returns a poi 47 destroyed. The kernel interface returns a pointer directly to the key, and
48 it's up to the caller to destroy the key. 48 it's up to the caller to destroy the key.
49 49
50 The request_key_tag() call is like the in-kern 50 The request_key_tag() call is like the in-kernel request_key(), except that it
51 also takes a domain tag that allows keys to be 51 also takes a domain tag that allows keys to be separated by namespace and
52 killed off as a group. 52 killed off as a group.
53 53
54 The request_key_with_auxdata() calls is like t 54 The request_key_with_auxdata() calls is like the request_key_tag() call, except
55 that they permit auxiliary data to be passed t 55 that they permit auxiliary data to be passed to the upcaller (the default is
56 NULL). This is only useful for those key type 56 NULL). This is only useful for those key types that define their own upcall
57 mechanism rather than using /sbin/request-key. 57 mechanism rather than using /sbin/request-key.
58 58
59 The request_key_rcu() call is like the request 59 The request_key_rcu() call is like the request_key_tag() call, except that it
60 doesn't check for keys that are under construc 60 doesn't check for keys that are under construction and doesn't attempt to
61 construct missing keys. 61 construct missing keys.
62 62
63 The userspace interface links the key to a key 63 The userspace interface links the key to a keyring associated with the process
64 to prevent the key from going away, and return 64 to prevent the key from going away, and returns the serial number of the key to
65 the caller. 65 the caller.
66 66
67 67
68 The following example assumes that the key typ 68 The following example assumes that the key types involved don't define their
69 own upcall mechanisms. If they do, then those 69 own upcall mechanisms. If they do, then those should be substituted for the
70 forking and execution of /sbin/request-key. 70 forking and execution of /sbin/request-key.
71 71
72 72
73 The Process 73 The Process
74 =========== 74 ===========
75 75
76 A request proceeds in the following manner: 76 A request proceeds in the following manner:
77 77
78 1) Process A calls request_key() [the usersp 78 1) Process A calls request_key() [the userspace syscall calls the kernel
79 interface]. 79 interface].
80 80
81 2) request_key() searches the process's subs 81 2) request_key() searches the process's subscribed keyrings to see if there's
82 a suitable key there. If there is, it re 82 a suitable key there. If there is, it returns the key. If there isn't,
83 and callout_info is not set, an error is 83 and callout_info is not set, an error is returned. Otherwise the process
84 proceeds to the next step. 84 proceeds to the next step.
85 85
86 3) request_key() sees that A doesn't have th 86 3) request_key() sees that A doesn't have the desired key yet, so it creates
87 two things: 87 two things:
88 88
89 a) An uninstantiated key U of requested 89 a) An uninstantiated key U of requested type and description.
90 90
91 b) An authorisation key V that refers to 91 b) An authorisation key V that refers to key U and notes that process A
92 is the context in which key U should 92 is the context in which key U should be instantiated and secured, and
93 from which associated key requests ma 93 from which associated key requests may be satisfied.
94 94
95 4) request_key() then forks and executes /sb 95 4) request_key() then forks and executes /sbin/request-key with a new session
96 keyring that contains a link to auth key 96 keyring that contains a link to auth key V.
97 97
98 5) /sbin/request-key assumes the authority a 98 5) /sbin/request-key assumes the authority associated with key U.
99 99
100 6) /sbin/request-key execs an appropriate pr 100 6) /sbin/request-key execs an appropriate program to perform the actual
101 instantiation. 101 instantiation.
102 102
103 7) The program may want to access another ke 103 7) The program may want to access another key from A's context (say a
104 Kerberos TGT key). It just requests the 104 Kerberos TGT key). It just requests the appropriate key, and the keyring
105 search notes that the session keyring has 105 search notes that the session keyring has auth key V in its bottom level.
106 106
107 This will permit it to then search the ke 107 This will permit it to then search the keyrings of process A with the
108 UID, GID, groups and security info of pro 108 UID, GID, groups and security info of process A as if it was process A,
109 and come up with key W. 109 and come up with key W.
110 110
111 8) The program then does what it must to get 111 8) The program then does what it must to get the data with which to
112 instantiate key U, using key W as a refer 112 instantiate key U, using key W as a reference (perhaps it contacts a
113 Kerberos server using the TGT) and then i 113 Kerberos server using the TGT) and then instantiates key U.
114 114
115 9) Upon instantiating key U, auth key V is a 115 9) Upon instantiating key U, auth key V is automatically revoked so that it
116 may not be used again. 116 may not be used again.
117 117
118 10) The program then exits 0 and request_key 118 10) The program then exits 0 and request_key() deletes key V and returns key
119 U to the caller. 119 U to the caller.
120 120
121 This also extends further. If key W (step 7 a 121 This also extends further. If key W (step 7 above) didn't exist, key W would
122 be created uninstantiated, another auth key (X 122 be created uninstantiated, another auth key (X) would be created (as per step
123 3) and another copy of /sbin/request-key spawn 123 3) and another copy of /sbin/request-key spawned (as per step 4); but the
124 context specified by auth key X will still be 124 context specified by auth key X will still be process A, as it was in auth key
125 V. 125 V.
126 126
127 This is because process A's keyrings can't sim 127 This is because process A's keyrings can't simply be attached to
128 /sbin/request-key at the appropriate places be 128 /sbin/request-key at the appropriate places because (a) execve will discard two
129 of them, and (b) it requires the same UID/GID/ 129 of them, and (b) it requires the same UID/GID/Groups all the way through.
130 130
131 131
132 Negative Instantiation And Rejection 132 Negative Instantiation And Rejection
133 ==================================== 133 ====================================
134 134
135 Rather than instantiating a key, it is possibl 135 Rather than instantiating a key, it is possible for the possessor of an
136 authorisation key to negatively instantiate a 136 authorisation key to negatively instantiate a key that's under construction.
137 This is a short duration placeholder that caus 137 This is a short duration placeholder that causes any attempt at re-requesting
138 the key while it exists to fail with error ENO 138 the key while it exists to fail with error ENOKEY if negated or the specified
139 error if rejected. 139 error if rejected.
140 140
141 This is provided to prevent excessive repeated 141 This is provided to prevent excessive repeated spawning of /sbin/request-key
142 processes for a key that will never be obtaina 142 processes for a key that will never be obtainable.
143 143
144 Should the /sbin/request-key process exit anyt 144 Should the /sbin/request-key process exit anything other than 0 or die on a
145 signal, the key under construction will be aut 145 signal, the key under construction will be automatically negatively
146 instantiated for a short amount of time. 146 instantiated for a short amount of time.
147 147
148 148
149 The Search Algorithm 149 The Search Algorithm
150 ==================== 150 ====================
151 151
152 A search of any particular keyring proceeds in 152 A search of any particular keyring proceeds in the following fashion:
153 153
154 1) When the key management code searches for 154 1) When the key management code searches for a key (keyring_search_rcu) it
155 firstly calls key_permission(SEARCH) on t 155 firstly calls key_permission(SEARCH) on the keyring it's starting with,
156 if this denies permission, it doesn't sea 156 if this denies permission, it doesn't search further.
157 157
158 2) It considers all the non-keyring keys wit 158 2) It considers all the non-keyring keys within that keyring and, if any key
159 matches the criteria specified, calls key 159 matches the criteria specified, calls key_permission(SEARCH) on it to see
160 if the key is allowed to be found. If it 160 if the key is allowed to be found. If it is, that key is returned; if
161 not, the search continues, and the error 161 not, the search continues, and the error code is retained if of higher
162 priority than the one currently set. 162 priority than the one currently set.
163 163
164 3) It then considers all the keyring-type ke 164 3) It then considers all the keyring-type keys in the keyring it's currently
165 searching. It calls key_permission(SEARC 165 searching. It calls key_permission(SEARCH) on each keyring, and if this
166 grants permission, it recurses, executing 166 grants permission, it recurses, executing steps (2) and (3) on that
167 keyring. 167 keyring.
168 168
169 The process stops immediately a valid key is f 169 The process stops immediately a valid key is found with permission granted to
170 use it. Any error from a previous match attem 170 use it. Any error from a previous match attempt is discarded and the key is
171 returned. 171 returned.
172 172
173 When request_key() is invoked, if CONFIG_KEYS_ 173 When request_key() is invoked, if CONFIG_KEYS_REQUEST_CACHE=y, a per-task
174 one-key cache is first checked for a match. 174 one-key cache is first checked for a match.
175 175
176 When search_process_keyrings() is invoked, it 176 When search_process_keyrings() is invoked, it performs the following searches
177 until one succeeds: 177 until one succeeds:
178 178
179 1) If extant, the process's thread keyring i 179 1) If extant, the process's thread keyring is searched.
180 180
181 2) If extant, the process's process keyring 181 2) If extant, the process's process keyring is searched.
182 182
183 3) The process's session keyring is searched 183 3) The process's session keyring is searched.
184 184
185 4) If the process has assumed the authority 185 4) If the process has assumed the authority associated with a request_key()
186 authorisation key then: 186 authorisation key then:
187 187
188 a) If extant, the calling process's thre 188 a) If extant, the calling process's thread keyring is searched.
189 189
190 b) If extant, the calling process's proc 190 b) If extant, the calling process's process keyring is searched.
191 191
192 c) The calling process's session keyring 192 c) The calling process's session keyring is searched.
193 193
194 The moment one succeeds, all pending errors ar 194 The moment one succeeds, all pending errors are discarded and the found key is
195 returned. If CONFIG_KEYS_REQUEST_CACHE=y, the 195 returned. If CONFIG_KEYS_REQUEST_CACHE=y, then that key is placed in the
196 per-task cache, displacing the previous key. 196 per-task cache, displacing the previous key. The cache is cleared on exit or
197 just prior to resumption of userspace. 197 just prior to resumption of userspace.
198 198
199 Only if all these fail does the whole thing fa 199 Only if all these fail does the whole thing fail with the highest priority
200 error. Note that several errors may have come 200 error. Note that several errors may have come from LSM.
201 201
202 The error priority is:: 202 The error priority is::
203 203
204 EKEYREVOKED > EKEYEXPIRED > ENOKEY 204 EKEYREVOKED > EKEYEXPIRED > ENOKEY
205 205
206 EACCES/EPERM are only returned on a direct sea 206 EACCES/EPERM are only returned on a direct search of a specific keyring where
207 the basal keyring does not grant Search permis 207 the basal keyring does not grant Search permission.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.