1 ===================================== 1 =====================================
2 Filesystem-level encryption (fscrypt) 2 Filesystem-level encryption (fscrypt)
3 ===================================== 3 =====================================
4 4
5 Introduction 5 Introduction
6 ============ 6 ============
7 7
8 fscrypt is a library which filesystems can hoo 8 fscrypt is a library which filesystems can hook into to support
9 transparent encryption of files and directorie 9 transparent encryption of files and directories.
10 10
11 Note: "fscrypt" in this document refers to the 11 Note: "fscrypt" in this document refers to the kernel-level portion,
12 implemented in ``fs/crypto/``, as opposed to t 12 implemented in ``fs/crypto/``, as opposed to the userspace tool
13 `fscrypt <https://github.com/google/fscrypt>`_ 13 `fscrypt <https://github.com/google/fscrypt>`_. This document only
14 covers the kernel-level portion. For command- 14 covers the kernel-level portion. For command-line examples of how to
15 use encryption, see the documentation for the 15 use encryption, see the documentation for the userspace tool `fscrypt
16 <https://github.com/google/fscrypt>`_. Also, 16 <https://github.com/google/fscrypt>`_. Also, it is recommended to use
17 the fscrypt userspace tool, or other existing 17 the fscrypt userspace tool, or other existing userspace tools such as
18 `fscryptctl <https://github.com/google/fscrypt 18 `fscryptctl <https://github.com/google/fscryptctl>`_ or `Android's key
19 management system 19 management system
20 <https://source.android.com/security/encryptio 20 <https://source.android.com/security/encryption/file-based>`_, over
21 using the kernel's API directly. Using existi 21 using the kernel's API directly. Using existing tools reduces the
22 chance of introducing your own security bugs. 22 chance of introducing your own security bugs. (Nevertheless, for
23 completeness this documentation covers the ker 23 completeness this documentation covers the kernel's API anyway.)
24 24
25 Unlike dm-crypt, fscrypt operates at the files 25 Unlike dm-crypt, fscrypt operates at the filesystem level rather than
26 at the block device level. This allows it to 26 at the block device level. This allows it to encrypt different files
27 with different keys and to have unencrypted fi 27 with different keys and to have unencrypted files on the same
28 filesystem. This is useful for multi-user sys 28 filesystem. This is useful for multi-user systems where each user's
29 data-at-rest needs to be cryptographically iso 29 data-at-rest needs to be cryptographically isolated from the others.
30 However, except for filenames, fscrypt does no 30 However, except for filenames, fscrypt does not encrypt filesystem
31 metadata. 31 metadata.
32 32
33 Unlike eCryptfs, which is a stacked filesystem 33 Unlike eCryptfs, which is a stacked filesystem, fscrypt is integrated
34 directly into supported filesystems --- curren !! 34 directly into supported filesystems --- currently ext4, F2FS, and
35 and CephFS. This allows encrypted files to be !! 35 UBIFS. This allows encrypted files to be read and written without
36 without caching both the decrypted and encrypt !! 36 caching both the decrypted and encrypted pages in the pagecache,
37 pagecache, thereby nearly halving the memory u !! 37 thereby nearly halving the memory used and bringing it in line with
38 line with unencrypted files. Similarly, half !! 38 unencrypted files. Similarly, half as many dentries and inodes are
39 inodes are needed. eCryptfs also limits encry !! 39 needed. eCryptfs also limits encrypted filenames to 143 bytes,
40 bytes, causing application compatibility issue !! 40 causing application compatibility issues; fscrypt allows the full 255
41 full 255 bytes (NAME_MAX). Finally, unlike eC !! 41 bytes (NAME_MAX). Finally, unlike eCryptfs, the fscrypt API can be
42 can be used by unprivileged users, with no nee !! 42 used by unprivileged users, with no need to mount anything.
43 43
44 fscrypt does not support encrypting files in-p 44 fscrypt does not support encrypting files in-place. Instead, it
45 supports marking an empty directory as encrypt 45 supports marking an empty directory as encrypted. Then, after
46 userspace provides the key, all regular files, 46 userspace provides the key, all regular files, directories, and
47 symbolic links created in that directory tree 47 symbolic links created in that directory tree are transparently
48 encrypted. 48 encrypted.
49 49
50 Threat model 50 Threat model
51 ============ 51 ============
52 52
53 Offline attacks 53 Offline attacks
54 --------------- 54 ---------------
55 55
56 Provided that userspace chooses a strong encry 56 Provided that userspace chooses a strong encryption key, fscrypt
57 protects the confidentiality of file contents 57 protects the confidentiality of file contents and filenames in the
58 event of a single point-in-time permanent offl 58 event of a single point-in-time permanent offline compromise of the
59 block device content. fscrypt does not protec 59 block device content. fscrypt does not protect the confidentiality of
60 non-filename metadata, e.g. file sizes, file p 60 non-filename metadata, e.g. file sizes, file permissions, file
61 timestamps, and extended attributes. Also, th 61 timestamps, and extended attributes. Also, the existence and location
62 of holes (unallocated blocks which logically c 62 of holes (unallocated blocks which logically contain all zeroes) in
63 files is not protected. 63 files is not protected.
64 64
65 fscrypt is not guaranteed to protect confident 65 fscrypt is not guaranteed to protect confidentiality or authenticity
66 if an attacker is able to manipulate the files 66 if an attacker is able to manipulate the filesystem offline prior to
67 an authorized user later accessing the filesys 67 an authorized user later accessing the filesystem.
68 68
69 Online attacks 69 Online attacks
70 -------------- 70 --------------
71 71
72 fscrypt (and storage encryption in general) ca 72 fscrypt (and storage encryption in general) can only provide limited
73 protection, if any at all, against online atta 73 protection, if any at all, against online attacks. In detail:
74 74
75 Side-channel attacks 75 Side-channel attacks
76 ~~~~~~~~~~~~~~~~~~~~ 76 ~~~~~~~~~~~~~~~~~~~~
77 77
78 fscrypt is only resistant to side-channel atta 78 fscrypt is only resistant to side-channel attacks, such as timing or
79 electromagnetic attacks, to the extent that th 79 electromagnetic attacks, to the extent that the underlying Linux
80 Cryptographic API algorithms or inline encrypt !! 80 Cryptographic API algorithms are. If a vulnerable algorithm is used,
81 vulnerable algorithm is used, such as a table- !! 81 such as a table-based implementation of AES, it may be possible for an
82 AES, it may be possible for an attacker to mou !! 82 attacker to mount a side channel attack against the online system.
83 against the online system. Side channel attac !! 83 Side channel attacks may also be mounted against applications
84 against applications consuming decrypted data. !! 84 consuming decrypted data.
85 85
86 Unauthorized file access 86 Unauthorized file access
87 ~~~~~~~~~~~~~~~~~~~~~~~~ 87 ~~~~~~~~~~~~~~~~~~~~~~~~
88 88
89 After an encryption key has been added, fscryp 89 After an encryption key has been added, fscrypt does not hide the
90 plaintext file contents or filenames from othe 90 plaintext file contents or filenames from other users on the same
91 system. Instead, existing access control mech 91 system. Instead, existing access control mechanisms such as file mode
92 bits, POSIX ACLs, LSMs, or namespaces should b 92 bits, POSIX ACLs, LSMs, or namespaces should be used for this purpose.
93 93
94 (For the reasoning behind this, understand tha 94 (For the reasoning behind this, understand that while the key is
95 added, the confidentiality of the data, from t 95 added, the confidentiality of the data, from the perspective of the
96 system itself, is *not* protected by the mathe 96 system itself, is *not* protected by the mathematical properties of
97 encryption but rather only by the correctness 97 encryption but rather only by the correctness of the kernel.
98 Therefore, any encryption-specific access cont 98 Therefore, any encryption-specific access control checks would merely
99 be enforced by kernel *code* and therefore wou 99 be enforced by kernel *code* and therefore would be largely redundant
100 with the wide variety of access control mechan 100 with the wide variety of access control mechanisms already available.)
101 101
102 Kernel memory compromise 102 Kernel memory compromise
103 ~~~~~~~~~~~~~~~~~~~~~~~~ 103 ~~~~~~~~~~~~~~~~~~~~~~~~
104 104
105 An attacker who compromises the system enough 105 An attacker who compromises the system enough to read from arbitrary
106 memory, e.g. by mounting a physical attack or 106 memory, e.g. by mounting a physical attack or by exploiting a kernel
107 security vulnerability, can compromise all enc 107 security vulnerability, can compromise all encryption keys that are
108 currently in use. 108 currently in use.
109 109
110 However, fscrypt allows encryption keys to be 110 However, fscrypt allows encryption keys to be removed from the kernel,
111 which may protect them from later compromise. 111 which may protect them from later compromise.
112 112
113 In more detail, the FS_IOC_REMOVE_ENCRYPTION_K 113 In more detail, the FS_IOC_REMOVE_ENCRYPTION_KEY ioctl (or the
114 FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS ioctl) 114 FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS ioctl) can wipe a master
115 encryption key from kernel memory. If it does 115 encryption key from kernel memory. If it does so, it will also try to
116 evict all cached inodes which had been "unlock 116 evict all cached inodes which had been "unlocked" using the key,
117 thereby wiping their per-file keys and making 117 thereby wiping their per-file keys and making them once again appear
118 "locked", i.e. in ciphertext or encrypted form 118 "locked", i.e. in ciphertext or encrypted form.
119 119
120 However, these ioctls have some limitations: 120 However, these ioctls have some limitations:
121 121
122 - Per-file keys for in-use files will *not* be 122 - Per-file keys for in-use files will *not* be removed or wiped.
123 Therefore, for maximum effect, userspace sho 123 Therefore, for maximum effect, userspace should close the relevant
124 encrypted files and directories before remov 124 encrypted files and directories before removing a master key, as
125 well as kill any processes whose working dir 125 well as kill any processes whose working directory is in an affected
126 encrypted directory. 126 encrypted directory.
127 127
128 - The kernel cannot magically wipe copies of t 128 - The kernel cannot magically wipe copies of the master key(s) that
129 userspace might have as well. Therefore, us 129 userspace might have as well. Therefore, userspace must wipe all
130 copies of the master key(s) it makes as well 130 copies of the master key(s) it makes as well; normally this should
131 be done immediately after FS_IOC_ADD_ENCRYPT 131 be done immediately after FS_IOC_ADD_ENCRYPTION_KEY, without waiting
132 for FS_IOC_REMOVE_ENCRYPTION_KEY. Naturally 132 for FS_IOC_REMOVE_ENCRYPTION_KEY. Naturally, the same also applies
133 to all higher levels in the key hierarchy. 133 to all higher levels in the key hierarchy. Userspace should also
134 follow other security precautions such as ml 134 follow other security precautions such as mlock()ing memory
135 containing keys to prevent it from being swa 135 containing keys to prevent it from being swapped out.
136 136
137 - In general, decrypted contents and filenames 137 - In general, decrypted contents and filenames in the kernel VFS
138 caches are freed but not wiped. Therefore, 138 caches are freed but not wiped. Therefore, portions thereof may be
139 recoverable from freed memory, even after th 139 recoverable from freed memory, even after the corresponding key(s)
140 were wiped. To partially solve this, you ca 140 were wiped. To partially solve this, you can set
141 CONFIG_PAGE_POISONING=y in your kernel confi 141 CONFIG_PAGE_POISONING=y in your kernel config and add page_poison=1
142 to your kernel command line. However, this 142 to your kernel command line. However, this has a performance cost.
143 143
144 - Secret keys might still exist in CPU registe 144 - Secret keys might still exist in CPU registers, in crypto
145 accelerator hardware (if used by the crypto 145 accelerator hardware (if used by the crypto API to implement any of
146 the algorithms), or in other places not expl 146 the algorithms), or in other places not explicitly considered here.
147 147
148 Limitations of v1 policies 148 Limitations of v1 policies
149 ~~~~~~~~~~~~~~~~~~~~~~~~~~ 149 ~~~~~~~~~~~~~~~~~~~~~~~~~~
150 150
151 v1 encryption policies have some weaknesses wi 151 v1 encryption policies have some weaknesses with respect to online
152 attacks: 152 attacks:
153 153
154 - There is no verification that the provided m 154 - There is no verification that the provided master key is correct.
155 Therefore, a malicious user can temporarily 155 Therefore, a malicious user can temporarily associate the wrong key
156 with another user's encrypted files to which 156 with another user's encrypted files to which they have read-only
157 access. Because of filesystem caching, the 157 access. Because of filesystem caching, the wrong key will then be
158 used by the other user's accesses to those f 158 used by the other user's accesses to those files, even if the other
159 user has the correct key in their own keyrin 159 user has the correct key in their own keyring. This violates the
160 meaning of "read-only access". 160 meaning of "read-only access".
161 161
162 - A compromise of a per-file key also compromi 162 - A compromise of a per-file key also compromises the master key from
163 which it was derived. 163 which it was derived.
164 164
165 - Non-root users cannot securely remove encryp 165 - Non-root users cannot securely remove encryption keys.
166 166
167 All the above problems are fixed with v2 encry 167 All the above problems are fixed with v2 encryption policies. For
168 this reason among others, it is recommended to 168 this reason among others, it is recommended to use v2 encryption
169 policies on all new encrypted directories. 169 policies on all new encrypted directories.
170 170
171 Key hierarchy 171 Key hierarchy
172 ============= 172 =============
173 173
174 Master Keys 174 Master Keys
175 ----------- 175 -----------
176 176
177 Each encrypted directory tree is protected by 177 Each encrypted directory tree is protected by a *master key*. Master
178 keys can be up to 64 bytes long, and must be a 178 keys can be up to 64 bytes long, and must be at least as long as the
179 greater of the security strength of the conten !! 179 greater of the key length needed by the contents and filenames
180 encryption modes being used. For example, if !! 180 encryption modes being used. For example, if AES-256-XTS is used for
181 used, the master key must be at least 256 bits !! 181 contents encryption, the master key must be 64 bytes (512 bits). Note
182 stricter requirement applies if the key is use !! 182 that the XTS mode is defined to require a key twice as long as that
183 policy and AES-256-XTS is used; such keys must !! 183 required by the underlying block cipher.
184 184
185 To "unlock" an encrypted directory tree, users 185 To "unlock" an encrypted directory tree, userspace must provide the
186 appropriate master key. There can be any numb 186 appropriate master key. There can be any number of master keys, each
187 of which protects any number of directory tree 187 of which protects any number of directory trees on any number of
188 filesystems. 188 filesystems.
189 189
190 Master keys must be real cryptographic keys, i 190 Master keys must be real cryptographic keys, i.e. indistinguishable
191 from random bytestrings of the same length. T 191 from random bytestrings of the same length. This implies that users
192 **must not** directly use a password as a mast 192 **must not** directly use a password as a master key, zero-pad a
193 shorter key, or repeat a shorter key. Securit 193 shorter key, or repeat a shorter key. Security cannot be guaranteed
194 if userspace makes any such error, as the cryp 194 if userspace makes any such error, as the cryptographic proofs and
195 analysis would no longer apply. 195 analysis would no longer apply.
196 196
197 Instead, users should generate master keys eit 197 Instead, users should generate master keys either using a
198 cryptographically secure random number generat 198 cryptographically secure random number generator, or by using a KDF
199 (Key Derivation Function). The kernel does no 199 (Key Derivation Function). The kernel does not do any key stretching;
200 therefore, if userspace derives the key from a 200 therefore, if userspace derives the key from a low-entropy secret such
201 as a passphrase, it is critical that a KDF des 201 as a passphrase, it is critical that a KDF designed for this purpose
202 be used, such as scrypt, PBKDF2, or Argon2. 202 be used, such as scrypt, PBKDF2, or Argon2.
203 203
204 Key derivation function 204 Key derivation function
205 ----------------------- 205 -----------------------
206 206
207 With one exception, fscrypt never uses the mas 207 With one exception, fscrypt never uses the master key(s) for
208 encryption directly. Instead, they are only u 208 encryption directly. Instead, they are only used as input to a KDF
209 (Key Derivation Function) to derive the actual 209 (Key Derivation Function) to derive the actual keys.
210 210
211 The KDF used for a particular master key diffe 211 The KDF used for a particular master key differs depending on whether
212 the key is used for v1 encryption policies or 212 the key is used for v1 encryption policies or for v2 encryption
213 policies. Users **must not** use the same key 213 policies. Users **must not** use the same key for both v1 and v2
214 encryption policies. (No real-world attack is 214 encryption policies. (No real-world attack is currently known on this
215 specific case of key reuse, but its security c 215 specific case of key reuse, but its security cannot be guaranteed
216 since the cryptographic proofs and analysis wo 216 since the cryptographic proofs and analysis would no longer apply.)
217 217
218 For v1 encryption policies, the KDF only suppo 218 For v1 encryption policies, the KDF only supports deriving per-file
219 encryption keys. It works by encrypting the m 219 encryption keys. It works by encrypting the master key with
220 AES-128-ECB, using the file's 16-byte nonce as 220 AES-128-ECB, using the file's 16-byte nonce as the AES key. The
221 resulting ciphertext is used as the derived ke 221 resulting ciphertext is used as the derived key. If the ciphertext is
222 longer than needed, then it is truncated to th 222 longer than needed, then it is truncated to the needed length.
223 223
224 For v2 encryption policies, the KDF is HKDF-SH 224 For v2 encryption policies, the KDF is HKDF-SHA512. The master key is
225 passed as the "input keying material", no salt 225 passed as the "input keying material", no salt is used, and a distinct
226 "application-specific information string" is u 226 "application-specific information string" is used for each distinct
227 key to be derived. For example, when a per-fi 227 key to be derived. For example, when a per-file encryption key is
228 derived, the application-specific information 228 derived, the application-specific information string is the file's
229 nonce prefixed with "fscrypt\\0" and a context 229 nonce prefixed with "fscrypt\\0" and a context byte. Different
230 context bytes are used for other types of deri 230 context bytes are used for other types of derived keys.
231 231
232 HKDF-SHA512 is preferred to the original AES-1 232 HKDF-SHA512 is preferred to the original AES-128-ECB based KDF because
233 HKDF is more flexible, is nonreversible, and e 233 HKDF is more flexible, is nonreversible, and evenly distributes
234 entropy from the master key. HKDF is also sta 234 entropy from the master key. HKDF is also standardized and widely
235 used by other software, whereas the AES-128-EC 235 used by other software, whereas the AES-128-ECB based KDF is ad-hoc.
236 236
237 Per-file encryption keys 237 Per-file encryption keys
238 ------------------------ 238 ------------------------
239 239
240 Since each master key can protect many files, 240 Since each master key can protect many files, it is necessary to
241 "tweak" the encryption of each file so that th 241 "tweak" the encryption of each file so that the same plaintext in two
242 files doesn't map to the same ciphertext, or v 242 files doesn't map to the same ciphertext, or vice versa. In most
243 cases, fscrypt does this by deriving per-file 243 cases, fscrypt does this by deriving per-file keys. When a new
244 encrypted inode (regular file, directory, or s 244 encrypted inode (regular file, directory, or symlink) is created,
245 fscrypt randomly generates a 16-byte nonce and 245 fscrypt randomly generates a 16-byte nonce and stores it in the
246 inode's encryption xattr. Then, it uses a KDF 246 inode's encryption xattr. Then, it uses a KDF (as described in `Key
247 derivation function`_) to derive the file's ke 247 derivation function`_) to derive the file's key from the master key
248 and nonce. 248 and nonce.
249 249
250 Key derivation was chosen over key wrapping be 250 Key derivation was chosen over key wrapping because wrapped keys would
251 require larger xattrs which would be less like 251 require larger xattrs which would be less likely to fit in-line in the
252 filesystem's inode table, and there didn't app 252 filesystem's inode table, and there didn't appear to be any
253 significant advantages to key wrapping. In pa 253 significant advantages to key wrapping. In particular, currently
254 there is no requirement to support unlocking a 254 there is no requirement to support unlocking a file with multiple
255 alternative master keys or to support rotating 255 alternative master keys or to support rotating master keys. Instead,
256 the master keys may be wrapped in userspace, e 256 the master keys may be wrapped in userspace, e.g. as is done by the
257 `fscrypt <https://github.com/google/fscrypt>`_ 257 `fscrypt <https://github.com/google/fscrypt>`_ tool.
258 258
259 DIRECT_KEY policies 259 DIRECT_KEY policies
260 ------------------- 260 -------------------
261 261
262 The Adiantum encryption mode (see `Encryption 262 The Adiantum encryption mode (see `Encryption modes and usage`_) is
263 suitable for both contents and filenames encry 263 suitable for both contents and filenames encryption, and it accepts
264 long IVs --- long enough to hold both an 8-byt !! 264 long IVs --- long enough to hold both an 8-byte logical block number
265 16-byte per-file nonce. Also, the overhead of !! 265 and a 16-byte per-file nonce. Also, the overhead of each Adiantum key
266 greater than that of an AES-256-XTS key. !! 266 is greater than that of an AES-256-XTS key.
267 267
268 Therefore, to improve performance and save mem 268 Therefore, to improve performance and save memory, for Adiantum a
269 "direct key" configuration is supported. When 269 "direct key" configuration is supported. When the user has enabled
270 this by setting FSCRYPT_POLICY_FLAG_DIRECT_KEY 270 this by setting FSCRYPT_POLICY_FLAG_DIRECT_KEY in the fscrypt policy,
271 per-file encryption keys are not used. Instea 271 per-file encryption keys are not used. Instead, whenever any data
272 (contents or filenames) is encrypted, the file 272 (contents or filenames) is encrypted, the file's 16-byte nonce is
273 included in the IV. Moreover: 273 included in the IV. Moreover:
274 274
275 - For v1 encryption policies, the encryption i 275 - For v1 encryption policies, the encryption is done directly with the
276 master key. Because of this, users **must n 276 master key. Because of this, users **must not** use the same master
277 key for any other purpose, even for other v1 277 key for any other purpose, even for other v1 policies.
278 278
279 - For v2 encryption policies, the encryption i 279 - For v2 encryption policies, the encryption is done with a per-mode
280 key derived using the KDF. Users may use th 280 key derived using the KDF. Users may use the same master key for
281 other v2 encryption policies. 281 other v2 encryption policies.
282 282
283 IV_INO_LBLK_64 policies 283 IV_INO_LBLK_64 policies
284 ----------------------- 284 -----------------------
285 285
286 When FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64 is set 286 When FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64 is set in the fscrypt policy,
287 the encryption keys are derived from the maste 287 the encryption keys are derived from the master key, encryption mode
288 number, and filesystem UUID. This normally re 288 number, and filesystem UUID. This normally results in all files
289 protected by the same master key sharing a sin 289 protected by the same master key sharing a single contents encryption
290 key and a single filenames encryption key. To 290 key and a single filenames encryption key. To still encrypt different
291 files' data differently, inode numbers are inc 291 files' data differently, inode numbers are included in the IVs.
292 Consequently, shrinking the filesystem may not 292 Consequently, shrinking the filesystem may not be allowed.
293 293
294 This format is optimized for use with inline e 294 This format is optimized for use with inline encryption hardware
295 compliant with the UFS standard, which support 295 compliant with the UFS standard, which supports only 64 IV bits per
296 I/O request and may have only a small number o 296 I/O request and may have only a small number of keyslots.
297 297
298 IV_INO_LBLK_32 policies 298 IV_INO_LBLK_32 policies
299 ----------------------- 299 -----------------------
300 300
301 IV_INO_LBLK_32 policies work like IV_INO_LBLK_ 301 IV_INO_LBLK_32 policies work like IV_INO_LBLK_64, except that for
302 IV_INO_LBLK_32, the inode number is hashed wit 302 IV_INO_LBLK_32, the inode number is hashed with SipHash-2-4 (where the
303 SipHash key is derived from the master key) an !! 303 SipHash key is derived from the master key) and added to the file
304 unit index mod 2^32 to produce a 32-bit IV. !! 304 logical block number mod 2^32 to produce a 32-bit IV.
305 305
306 This format is optimized for use with inline e 306 This format is optimized for use with inline encryption hardware
307 compliant with the eMMC v5.2 standard, which s 307 compliant with the eMMC v5.2 standard, which supports only 32 IV bits
308 per I/O request and may have only a small numb 308 per I/O request and may have only a small number of keyslots. This
309 format results in some level of IV reuse, so i 309 format results in some level of IV reuse, so it should only be used
310 when necessary due to hardware limitations. 310 when necessary due to hardware limitations.
311 311
312 Key identifiers 312 Key identifiers
313 --------------- 313 ---------------
314 314
315 For master keys used for v2 encryption policie 315 For master keys used for v2 encryption policies, a unique 16-byte "key
316 identifier" is also derived using the KDF. Th 316 identifier" is also derived using the KDF. This value is stored in
317 the clear, since it is needed to reliably iden 317 the clear, since it is needed to reliably identify the key itself.
318 318
319 Dirhash keys 319 Dirhash keys
320 ------------ 320 ------------
321 321
322 For directories that are indexed using a secre 322 For directories that are indexed using a secret-keyed dirhash over the
323 plaintext filenames, the KDF is also used to d 323 plaintext filenames, the KDF is also used to derive a 128-bit
324 SipHash-2-4 key per directory in order to hash 324 SipHash-2-4 key per directory in order to hash filenames. This works
325 just like deriving a per-file encryption key, 325 just like deriving a per-file encryption key, except that a different
326 KDF context is used. Currently, only casefold 326 KDF context is used. Currently, only casefolded ("case-insensitive")
327 encrypted directories use this style of hashin 327 encrypted directories use this style of hashing.
328 328
329 Encryption modes and usage 329 Encryption modes and usage
330 ========================== 330 ==========================
331 331
332 fscrypt allows one encryption mode to be speci 332 fscrypt allows one encryption mode to be specified for file contents
333 and one encryption mode to be specified for fi 333 and one encryption mode to be specified for filenames. Different
334 directory trees are permitted to use different 334 directory trees are permitted to use different encryption modes.
335 <<
336 Supported modes <<
337 --------------- <<
338 <<
339 Currently, the following pairs of encryption m 335 Currently, the following pairs of encryption modes are supported:
340 336
341 - AES-256-XTS for contents and AES-256-CBC-CTS !! 337 - AES-256-XTS for contents and AES-256-CTS-CBC for filenames
342 - AES-256-XTS for contents and AES-256-HCTR2 f !! 338 - AES-128-CBC for contents and AES-128-CTS-CBC for filenames
343 - Adiantum for both contents and filenames 339 - Adiantum for both contents and filenames
344 - AES-128-CBC-ESSIV for contents and AES-128-C <<
345 - SM4-XTS for contents and SM4-CBC-CTS for fil <<
346 340
347 Note: in the API, "CBC" means CBC-ESSIV, and " !! 341 If unsure, you should use the (AES-256-XTS, AES-256-CTS-CBC) pair.
348 So, for example, FSCRYPT_MODE_AES_256_CTS mean <<
349 342
350 Authenticated encryption modes are not current !! 343 AES-128-CBC was added only for low-powered embedded devices with
351 the difficulty of dealing with ciphertext expa !! 344 crypto accelerators such as CAAM or CESA that do not support XTS. To
352 contents encryption uses a block cipher in `XT !! 345 use AES-128-CBC, CONFIG_CRYPTO_ESSIV and CONFIG_CRYPTO_SHA256 (or
353 <https://en.wikipedia.org/wiki/Disk_encryption !! 346 another SHA-256 implementation) must be enabled so that ESSIV can be
354 `CBC-ESSIV mode !! 347 used.
355 <https://en.wikipedia.org/wiki/Disk_encryption !! 348
356 or a wide-block cipher. Filenames encryption !! 349 Adiantum is a (primarily) stream cipher-based mode that is fast even
357 block cipher in `CBC-CTS mode !! 350 on CPUs without dedicated crypto instructions. It's also a true
358 <https://en.wikipedia.org/wiki/Ciphertext_stea !! 351 wide-block mode, unlike XTS. It can also eliminate the need to derive
359 cipher. !! 352 per-file encryption keys. However, it depends on the security of two
360 !! 353 primitives, XChaCha12 and AES-256, rather than just one. See the
361 The (AES-256-XTS, AES-256-CBC-CTS) pair is the !! 354 paper "Adiantum: length-preserving encryption for entry-level
362 It is also the only option that is *guaranteed !! 355 processors" (https://eprint.iacr.org/2018/720.pdf) for more details.
363 if the kernel supports fscrypt at all; see `Ke !! 356 To use Adiantum, CONFIG_CRYPTO_ADIANTUM must be enabled. Also, fast
364 !! 357 implementations of ChaCha and NHPoly1305 should be enabled, e.g.
365 The (AES-256-XTS, AES-256-HCTR2) pair is also !! 358 CONFIG_CRYPTO_CHACHA20_NEON and CONFIG_CRYPTO_NHPOLY1305_NEON for ARM.
366 upgrades the filenames encryption to use a wid !! 359
367 *wide-block cipher*, also called a tweakable s !! 360 New encryption modes can be added relatively easily, without changes
368 permutation, has the property that changing on !! 361 to individual filesystems. However, authenticated encryption (AE)
369 entire result.) As described in `Filenames en !! 362 modes are not currently supported because of the difficulty of dealing
370 cipher is the ideal mode for the problem domai !! 363 with ciphertext expansion.
371 "least bad" choice among the alternatives. Fo <<
372 HCTR2, see `the HCTR2 paper <https://eprint.ia <<
373 <<
374 Adiantum is recommended on systems where AES i <<
375 of hardware acceleration for AES. Adiantum is <<
376 that uses XChaCha12 and AES-256 as its underly <<
377 the work is done by XChaCha12, which is much f <<
378 acceleration is unavailable. For more informa <<
379 `the Adiantum paper <https://eprint.iacr.org/2 <<
380 <<
381 The (AES-128-CBC-ESSIV, AES-128-CBC-CTS) pair <<
382 systems whose only form of AES acceleration is <<
383 accelerator such as CAAM or CESA that does not <<
384 <<
385 The remaining mode pairs are the "national pri <<
386 <<
387 - (SM4-XTS, SM4-CBC-CTS) <<
388 <<
389 Generally speaking, these ciphers aren't "bad" <<
390 receive limited security review compared to th <<
391 AES and ChaCha. They also don't bring much ne <<
392 suggested to only use these ciphers where thei <<
393 <<
394 Kernel config options <<
395 --------------------- <<
396 <<
397 Enabling fscrypt support (CONFIG_FS_ENCRYPTION <<
398 only the basic support from the crypto API nee <<
399 and AES-256-CBC-CTS encryption. For optimal p <<
400 strongly recommended to also enable any availa <<
401 kconfig options that provide acceleration for <<
402 wish to use. Support for any "non-default" en <<
403 requires extra kconfig options as well. <<
404 <<
405 Below, some relevant options are listed by enc <<
406 acceleration options not listed below may be a <<
407 platform; refer to the kconfig menus. File co <<
408 also be configured to use inline encryption ha <<
409 kernel crypto API (see `Inline encryption supp <<
410 the file contents mode doesn't need to support <<
411 API, but the filenames mode still does. <<
412 <<
413 - AES-256-XTS and AES-256-CBC-CTS <<
414 - Recommended: <<
415 - arm64: CONFIG_CRYPTO_AES_ARM64_CE_BL <<
416 - x86: CONFIG_CRYPTO_AES_NI_INTEL <<
417 <<
418 - AES-256-HCTR2 <<
419 - Mandatory: <<
420 - CONFIG_CRYPTO_HCTR2 <<
421 - Recommended: <<
422 - arm64: CONFIG_CRYPTO_AES_ARM64_CE_BL <<
423 - arm64: CONFIG_CRYPTO_POLYVAL_ARM64_C <<
424 - x86: CONFIG_CRYPTO_AES_NI_INTEL <<
425 - x86: CONFIG_CRYPTO_POLYVAL_CLMUL_NI <<
426 <<
427 - Adiantum <<
428 - Mandatory: <<
429 - CONFIG_CRYPTO_ADIANTUM <<
430 - Recommended: <<
431 - arm32: CONFIG_CRYPTO_CHACHA20_NEON <<
432 - arm32: CONFIG_CRYPTO_NHPOLY1305_NEON <<
433 - arm64: CONFIG_CRYPTO_CHACHA20_NEON <<
434 - arm64: CONFIG_CRYPTO_NHPOLY1305_NEON <<
435 - x86: CONFIG_CRYPTO_CHACHA20_X86_64 <<
436 - x86: CONFIG_CRYPTO_NHPOLY1305_SSE2 <<
437 - x86: CONFIG_CRYPTO_NHPOLY1305_AVX2 <<
438 <<
439 - AES-128-CBC-ESSIV and AES-128-CBC-CTS: <<
440 - Mandatory: <<
441 - CONFIG_CRYPTO_ESSIV <<
442 - CONFIG_CRYPTO_SHA256 or another SHA- <<
443 - Recommended: <<
444 - AES-CBC acceleration <<
445 <<
446 fscrypt also uses HMAC-SHA512 for key derivati <<
447 acceleration is recommended: <<
448 <<
449 - SHA-512 <<
450 - Recommended: <<
451 - arm64: CONFIG_CRYPTO_SHA512_ARM64_CE <<
452 - x86: CONFIG_CRYPTO_SHA512_SSSE3 <<
453 364
454 Contents encryption 365 Contents encryption
455 ------------------- 366 -------------------
456 367
457 For contents encryption, each file's contents !! 368 For file contents, each filesystem block is encrypted independently.
458 units". Each data unit is encrypted independe !! 369 Starting from Linux kernel 5.5, encryption of filesystems with block
459 data unit incorporates the zero-based index of !! 370 size less than system's page size is supported.
460 the file. This ensures that each data unit wi !! 371
461 differently, which is essential to prevent lea !! 372 Each block's IV is set to the logical block number within the file as
462 !! 373 a little endian number, except that:
463 Note: the encryption depending on the offset i !! 374
464 operations like "collapse range" and "insert r !! 375 - With CBC mode encryption, ESSIV is also used. Specifically, each IV
465 extent mapping of files are not supported on e !! 376 is encrypted with AES-256 where the AES-256 key is the SHA-256 hash
466 !! 377 of the file's data encryption key.
467 There are two cases for the sizes of the data !! 378
468 !! 379 - With `DIRECT_KEY policies`_, the file's nonce is appended to the IV.
469 * Fixed-size data units. This is how all file !! 380 Currently this is only allowed with the Adiantum encryption mode.
470 work. A file's data units are all the same !! 381
471 is zero-padded if needed. By default, the d !! 382 - With `IV_INO_LBLK_64 policies`_, the logical block number is limited
472 to the filesystem block size. On some files !! 383 to 32 bits and is placed in bits 0-31 of the IV. The inode number
473 a sub-block data unit size via the ``log2_da !! 384 (which is also limited to 32 bits) is placed in bits 32-63.
474 the encryption policy; see `FS_IOC_SET_ENCRY !! 385
475 !! 386 - With `IV_INO_LBLK_32 policies`_, the logical block number is limited
476 * Variable-size data units. This is what UBIF !! 387 to 32 bits and is placed in bits 0-31 of the IV. The inode number
477 data node" is treated as a crypto data unit. !! 388 is then hashed and added mod 2^32.
478 length, possibly compressed data, zero-padde !! 389
479 boundary. Users cannot select a sub-block d !! 390 Note that because file logical block numbers are included in the IVs,
480 !! 391 filesystems must enforce that blocks are never shifted around within
481 In the case of compression + encryption, the c !! 392 encrypted files, e.g. via "collapse range" or "insert range".
482 encrypted. UBIFS compression works as describ <<
483 compression works a bit differently; it compre <<
484 filesystem blocks into a smaller number of fil <<
485 Therefore a f2fs-compressed file still uses fi <<
486 it is encrypted in a similar way to a file con <<
487 <<
488 As mentioned in `Key hierarchy`_, the default <<
489 per-file keys. In this case, the IV for each <<
490 index of the data unit in the file. However, <<
491 encryption setting that does not use per-file <<
492 kind of file identifier is incorporated into t <<
493 <<
494 - With `DIRECT_KEY policies`_, the data unit i <<
495 0-63 of the IV, and the file's nonce is plac <<
496 <<
497 - With `IV_INO_LBLK_64 policies`_, the data un <<
498 bits 0-31 of the IV, and the file's inode nu <<
499 32-63. This setting is only allowed when da <<
500 inode numbers fit in 32 bits. <<
501 <<
502 - With `IV_INO_LBLK_32 policies`_, the file's <<
503 and added to the data unit index. The resul <<
504 to 32 bits and placed in bits 0-31 of the IV <<
505 allowed when data unit indices and inode num <<
506 <<
507 The byte order of the IV is always little endi <<
508 <<
509 If the user selects FSCRYPT_MODE_AES_128_CBC f <<
510 ESSIV layer is automatically included. In thi <<
511 passed to AES-128-CBC, it is encrypted with AE <<
512 key is the SHA-256 hash of the file's contents <<
513 393
514 Filenames encryption 394 Filenames encryption
515 -------------------- 395 --------------------
516 396
517 For filenames, each full filename is encrypted 397 For filenames, each full filename is encrypted at once. Because of
518 the requirements to retain support for efficie 398 the requirements to retain support for efficient directory lookups and
519 filenames of up to 255 bytes, the same IV is u 399 filenames of up to 255 bytes, the same IV is used for every filename
520 in a directory. 400 in a directory.
521 401
522 However, each encrypted directory still uses a 402 However, each encrypted directory still uses a unique key, or
523 alternatively has the file's nonce (for `DIREC 403 alternatively has the file's nonce (for `DIRECT_KEY policies`_) or
524 inode number (for `IV_INO_LBLK_64 policies`_) 404 inode number (for `IV_INO_LBLK_64 policies`_) included in the IVs.
525 Thus, IV reuse is limited to within a single d 405 Thus, IV reuse is limited to within a single directory.
526 406
527 With CBC-CTS, the IV reuse means that when the !! 407 With CTS-CBC, the IV reuse means that when the plaintext filenames
528 common prefix at least as long as the cipher b !! 408 share a common prefix at least as long as the cipher block size (16
529 corresponding encrypted filenames will also sh !! 409 bytes for AES), the corresponding encrypted filenames will also share
530 undesirable. Adiantum and HCTR2 do not have t !! 410 a common prefix. This is undesirable. Adiantum does not have this
531 wide-block encryption modes. !! 411 weakness, as it is a wide-block encryption mode.
532 412
533 All supported filenames encryption modes accep 413 All supported filenames encryption modes accept any plaintext length
534 >= 16 bytes; cipher block alignment is not req 414 >= 16 bytes; cipher block alignment is not required. However,
535 filenames shorter than 16 bytes are NUL-padded 415 filenames shorter than 16 bytes are NUL-padded to 16 bytes before
536 being encrypted. In addition, to reduce leaka 416 being encrypted. In addition, to reduce leakage of filename lengths
537 via their ciphertexts, all filenames are NUL-p 417 via their ciphertexts, all filenames are NUL-padded to the next 4, 8,
538 16, or 32-byte boundary (configurable). 32 is 418 16, or 32-byte boundary (configurable). 32 is recommended since this
539 provides the best confidentiality, at the cost 419 provides the best confidentiality, at the cost of making directory
540 entries consume slightly more space. Note tha 420 entries consume slightly more space. Note that since NUL (``\0``) is
541 not otherwise a valid character in filenames, 421 not otherwise a valid character in filenames, the padding will never
542 produce duplicate plaintexts. 422 produce duplicate plaintexts.
543 423
544 Symbolic link targets are considered a type of 424 Symbolic link targets are considered a type of filename and are
545 encrypted in the same way as filenames in dire 425 encrypted in the same way as filenames in directory entries, except
546 that IV reuse is not a problem as each symlink 426 that IV reuse is not a problem as each symlink has its own inode.
547 427
548 User API 428 User API
549 ======== 429 ========
550 430
551 Setting an encryption policy 431 Setting an encryption policy
552 ---------------------------- 432 ----------------------------
553 433
554 FS_IOC_SET_ENCRYPTION_POLICY 434 FS_IOC_SET_ENCRYPTION_POLICY
555 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 435 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
556 436
557 The FS_IOC_SET_ENCRYPTION_POLICY ioctl sets an 437 The FS_IOC_SET_ENCRYPTION_POLICY ioctl sets an encryption policy on an
558 empty directory or verifies that a directory o 438 empty directory or verifies that a directory or regular file already
559 has the specified encryption policy. It takes 439 has the specified encryption policy. It takes in a pointer to
560 struct fscrypt_policy_v1 or struct fscrypt_pol 440 struct fscrypt_policy_v1 or struct fscrypt_policy_v2, defined as
561 follows:: 441 follows::
562 442
563 #define FSCRYPT_POLICY_V1 0 443 #define FSCRYPT_POLICY_V1 0
564 #define FSCRYPT_KEY_DESCRIPTOR_SIZE 8 444 #define FSCRYPT_KEY_DESCRIPTOR_SIZE 8
565 struct fscrypt_policy_v1 { 445 struct fscrypt_policy_v1 {
566 __u8 version; 446 __u8 version;
567 __u8 contents_encryption_mode; 447 __u8 contents_encryption_mode;
568 __u8 filenames_encryption_mode; 448 __u8 filenames_encryption_mode;
569 __u8 flags; 449 __u8 flags;
570 __u8 master_key_descriptor[FSCRYPT 450 __u8 master_key_descriptor[FSCRYPT_KEY_DESCRIPTOR_SIZE];
571 }; 451 };
572 #define fscrypt_policy fscrypt_policy_v1 452 #define fscrypt_policy fscrypt_policy_v1
573 453
574 #define FSCRYPT_POLICY_V2 2 454 #define FSCRYPT_POLICY_V2 2
575 #define FSCRYPT_KEY_IDENTIFIER_SIZE 16 455 #define FSCRYPT_KEY_IDENTIFIER_SIZE 16
576 struct fscrypt_policy_v2 { 456 struct fscrypt_policy_v2 {
577 __u8 version; 457 __u8 version;
578 __u8 contents_encryption_mode; 458 __u8 contents_encryption_mode;
579 __u8 filenames_encryption_mode; 459 __u8 filenames_encryption_mode;
580 __u8 flags; 460 __u8 flags;
581 __u8 log2_data_unit_size; !! 461 __u8 __reserved[4];
582 __u8 __reserved[3]; <<
583 __u8 master_key_identifier[FSCRYPT 462 __u8 master_key_identifier[FSCRYPT_KEY_IDENTIFIER_SIZE];
584 }; 463 };
585 464
586 This structure must be initialized as follows: 465 This structure must be initialized as follows:
587 466
588 - ``version`` must be FSCRYPT_POLICY_V1 (0) if 467 - ``version`` must be FSCRYPT_POLICY_V1 (0) if
589 struct fscrypt_policy_v1 is used or FSCRYPT_ 468 struct fscrypt_policy_v1 is used or FSCRYPT_POLICY_V2 (2) if
590 struct fscrypt_policy_v2 is used. (Note: we 469 struct fscrypt_policy_v2 is used. (Note: we refer to the original
591 policy version as "v1", though its version c 470 policy version as "v1", though its version code is really 0.)
592 For new encrypted directories, use v2 polici 471 For new encrypted directories, use v2 policies.
593 472
594 - ``contents_encryption_mode`` and ``filenames 473 - ``contents_encryption_mode`` and ``filenames_encryption_mode`` must
595 be set to constants from ``<linux/fscrypt.h> 474 be set to constants from ``<linux/fscrypt.h>`` which identify the
596 encryption modes to use. If unsure, use FSC 475 encryption modes to use. If unsure, use FSCRYPT_MODE_AES_256_XTS
597 (1) for ``contents_encryption_mode`` and FSC 476 (1) for ``contents_encryption_mode`` and FSCRYPT_MODE_AES_256_CTS
598 (4) for ``filenames_encryption_mode``. For !! 477 (4) for ``filenames_encryption_mode``.
599 modes and usage`_. <<
600 <<
601 v1 encryption policies only support three co <<
602 (FSCRYPT_MODE_AES_256_XTS, FSCRYPT_MODE_AES_ <<
603 (FSCRYPT_MODE_AES_128_CBC, FSCRYPT_MODE_AES_ <<
604 (FSCRYPT_MODE_ADIANTUM, FSCRYPT_MODE_ADIANTU <<
605 all combinations documented in `Supported mo <<
606 478
607 - ``flags`` contains optional flags from ``<li 479 - ``flags`` contains optional flags from ``<linux/fscrypt.h>``:
608 480
609 - FSCRYPT_POLICY_FLAGS_PAD_*: The amount of 481 - FSCRYPT_POLICY_FLAGS_PAD_*: The amount of NUL padding to use when
610 encrypting filenames. If unsure, use FSCR 482 encrypting filenames. If unsure, use FSCRYPT_POLICY_FLAGS_PAD_32
611 (0x3). 483 (0x3).
612 - FSCRYPT_POLICY_FLAG_DIRECT_KEY: See `DIREC 484 - FSCRYPT_POLICY_FLAG_DIRECT_KEY: See `DIRECT_KEY policies`_.
613 - FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64: See `I 485 - FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64: See `IV_INO_LBLK_64
614 policies`_. 486 policies`_.
615 - FSCRYPT_POLICY_FLAG_IV_INO_LBLK_32: See `I 487 - FSCRYPT_POLICY_FLAG_IV_INO_LBLK_32: See `IV_INO_LBLK_32
616 policies`_. 488 policies`_.
617 489
618 v1 encryption policies only support the PAD_ 490 v1 encryption policies only support the PAD_* and DIRECT_KEY flags.
619 The other flags are only supported by v2 enc 491 The other flags are only supported by v2 encryption policies.
620 492
621 The DIRECT_KEY, IV_INO_LBLK_64, and IV_INO_L 493 The DIRECT_KEY, IV_INO_LBLK_64, and IV_INO_LBLK_32 flags are
622 mutually exclusive. 494 mutually exclusive.
623 495
624 - ``log2_data_unit_size`` is the log2 of the d <<
625 or 0 to select the default data unit size. <<
626 the granularity of file contents encryption. <<
627 ``log2_data_unit_size`` to 12 causes file co <<
628 underlying encryption algorithm (such as AES <<
629 data units, each with its own IV. <<
630 <<
631 Not all filesystems support setting ``log2_d <<
632 and f2fs support it since Linux v6.7. On fi <<
633 it, the supported nonzero values are 9 throu <<
634 filesystem block size, inclusively. The def <<
635 the filesystem block size. <<
636 <<
637 The main use case for ``log2_data_unit_size` <<
638 data unit size smaller than the filesystem b <<
639 compatibility with inline encryption hardwar <<
640 smaller data unit sizes. ``/sys/block/$disk <<
641 useful for checking which data unit sizes ar <<
642 particular system's inline encryption hardwa <<
643 <<
644 Leave this field zeroed unless you are certa <<
645 an unnecessarily small data unit size reduce <<
646 <<
647 - For v2 encryption policies, ``__reserved`` m 496 - For v2 encryption policies, ``__reserved`` must be zeroed.
648 497
649 - For v1 encryption policies, ``master_key_des 498 - For v1 encryption policies, ``master_key_descriptor`` specifies how
650 to find the master key in a keyring; see `Ad 499 to find the master key in a keyring; see `Adding keys`_. It is up
651 to userspace to choose a unique ``master_key 500 to userspace to choose a unique ``master_key_descriptor`` for each
652 master key. The e4crypt and fscrypt tools u 501 master key. The e4crypt and fscrypt tools use the first 8 bytes of
653 ``SHA-512(SHA-512(master_key))``, but this p 502 ``SHA-512(SHA-512(master_key))``, but this particular scheme is not
654 required. Also, the master key need not be 503 required. Also, the master key need not be in the keyring yet when
655 FS_IOC_SET_ENCRYPTION_POLICY is executed. H 504 FS_IOC_SET_ENCRYPTION_POLICY is executed. However, it must be added
656 before any files can be created in the encry 505 before any files can be created in the encrypted directory.
657 506
658 For v2 encryption policies, ``master_key_des 507 For v2 encryption policies, ``master_key_descriptor`` has been
659 replaced with ``master_key_identifier``, whi 508 replaced with ``master_key_identifier``, which is longer and cannot
660 be arbitrarily chosen. Instead, the key mus 509 be arbitrarily chosen. Instead, the key must first be added using
661 `FS_IOC_ADD_ENCRYPTION_KEY`_. Then, the ``k 510 `FS_IOC_ADD_ENCRYPTION_KEY`_. Then, the ``key_spec.u.identifier``
662 the kernel returned in the struct fscrypt_ad 511 the kernel returned in the struct fscrypt_add_key_arg must
663 be used as the ``master_key_identifier`` in 512 be used as the ``master_key_identifier`` in
664 struct fscrypt_policy_v2. 513 struct fscrypt_policy_v2.
665 514
666 If the file is not yet encrypted, then FS_IOC_ 515 If the file is not yet encrypted, then FS_IOC_SET_ENCRYPTION_POLICY
667 verifies that the file is an empty directory. 516 verifies that the file is an empty directory. If so, the specified
668 encryption policy is assigned to the directory 517 encryption policy is assigned to the directory, turning it into an
669 encrypted directory. After that, and after pr 518 encrypted directory. After that, and after providing the
670 corresponding master key as described in `Addi 519 corresponding master key as described in `Adding keys`_, all regular
671 files, directories (recursively), and symlinks 520 files, directories (recursively), and symlinks created in the
672 directory will be encrypted, inheriting the sa 521 directory will be encrypted, inheriting the same encryption policy.
673 The filenames in the directory's entries will 522 The filenames in the directory's entries will be encrypted as well.
674 523
675 Alternatively, if the file is already encrypte 524 Alternatively, if the file is already encrypted, then
676 FS_IOC_SET_ENCRYPTION_POLICY validates that th 525 FS_IOC_SET_ENCRYPTION_POLICY validates that the specified encryption
677 policy exactly matches the actual one. If the 526 policy exactly matches the actual one. If they match, then the ioctl
678 returns 0. Otherwise, it fails with EEXIST. 527 returns 0. Otherwise, it fails with EEXIST. This works on both
679 regular files and directories, including nonem 528 regular files and directories, including nonempty directories.
680 529
681 When a v2 encryption policy is assigned to a d 530 When a v2 encryption policy is assigned to a directory, it is also
682 required that either the specified key has bee 531 required that either the specified key has been added by the current
683 user or that the caller has CAP_FOWNER in the 532 user or that the caller has CAP_FOWNER in the initial user namespace.
684 (This is needed to prevent a user from encrypt 533 (This is needed to prevent a user from encrypting their data with
685 another user's key.) The key must remain adde 534 another user's key.) The key must remain added while
686 FS_IOC_SET_ENCRYPTION_POLICY is executing. Ho 535 FS_IOC_SET_ENCRYPTION_POLICY is executing. However, if the new
687 encrypted directory does not need to be access 536 encrypted directory does not need to be accessed immediately, then the
688 key can be removed right away afterwards. 537 key can be removed right away afterwards.
689 538
690 Note that the ext4 filesystem does not allow t 539 Note that the ext4 filesystem does not allow the root directory to be
691 encrypted, even if it is empty. Users who wan 540 encrypted, even if it is empty. Users who want to encrypt an entire
692 filesystem with one key should consider using 541 filesystem with one key should consider using dm-crypt instead.
693 542
694 FS_IOC_SET_ENCRYPTION_POLICY can fail with the 543 FS_IOC_SET_ENCRYPTION_POLICY can fail with the following errors:
695 544
696 - ``EACCES``: the file is not owned by the pro 545 - ``EACCES``: the file is not owned by the process's uid, nor does the
697 process have the CAP_FOWNER capability in a 546 process have the CAP_FOWNER capability in a namespace with the file
698 owner's uid mapped 547 owner's uid mapped
699 - ``EEXIST``: the file is already encrypted wi 548 - ``EEXIST``: the file is already encrypted with an encryption policy
700 different from the one specified 549 different from the one specified
701 - ``EINVAL``: an invalid encryption policy was 550 - ``EINVAL``: an invalid encryption policy was specified (invalid
702 version, mode(s), or flags; or reserved bits 551 version, mode(s), or flags; or reserved bits were set); or a v1
703 encryption policy was specified but the dire 552 encryption policy was specified but the directory has the casefold
704 flag enabled (casefolding is incompatible wi 553 flag enabled (casefolding is incompatible with v1 policies).
705 - ``ENOKEY``: a v2 encryption policy was speci 554 - ``ENOKEY``: a v2 encryption policy was specified, but the key with
706 the specified ``master_key_identifier`` has 555 the specified ``master_key_identifier`` has not been added, nor does
707 the process have the CAP_FOWNER capability i 556 the process have the CAP_FOWNER capability in the initial user
708 namespace 557 namespace
709 - ``ENOTDIR``: the file is unencrypted and is 558 - ``ENOTDIR``: the file is unencrypted and is a regular file, not a
710 directory 559 directory
711 - ``ENOTEMPTY``: the file is unencrypted and i 560 - ``ENOTEMPTY``: the file is unencrypted and is a nonempty directory
712 - ``ENOTTY``: this type of filesystem does not 561 - ``ENOTTY``: this type of filesystem does not implement encryption
713 - ``EOPNOTSUPP``: the kernel was not configure 562 - ``EOPNOTSUPP``: the kernel was not configured with encryption
714 support for filesystems, or the filesystem s 563 support for filesystems, or the filesystem superblock has not
715 had encryption enabled on it. (For example, 564 had encryption enabled on it. (For example, to use encryption on an
716 ext4 filesystem, CONFIG_FS_ENCRYPTION must b 565 ext4 filesystem, CONFIG_FS_ENCRYPTION must be enabled in the
717 kernel config, and the superblock must have 566 kernel config, and the superblock must have had the "encrypt"
718 feature flag enabled using ``tune2fs -O encr 567 feature flag enabled using ``tune2fs -O encrypt`` or ``mkfs.ext4 -O
719 encrypt``.) 568 encrypt``.)
720 - ``EPERM``: this directory may not be encrypt 569 - ``EPERM``: this directory may not be encrypted, e.g. because it is
721 the root directory of an ext4 filesystem 570 the root directory of an ext4 filesystem
722 - ``EROFS``: the filesystem is readonly 571 - ``EROFS``: the filesystem is readonly
723 572
724 Getting an encryption policy 573 Getting an encryption policy
725 ---------------------------- 574 ----------------------------
726 575
727 Two ioctls are available to get a file's encry 576 Two ioctls are available to get a file's encryption policy:
728 577
729 - `FS_IOC_GET_ENCRYPTION_POLICY_EX`_ 578 - `FS_IOC_GET_ENCRYPTION_POLICY_EX`_
730 - `FS_IOC_GET_ENCRYPTION_POLICY`_ 579 - `FS_IOC_GET_ENCRYPTION_POLICY`_
731 580
732 The extended (_EX) version of the ioctl is mor 581 The extended (_EX) version of the ioctl is more general and is
733 recommended to use when possible. However, on 582 recommended to use when possible. However, on older kernels only the
734 original ioctl is available. Applications sho 583 original ioctl is available. Applications should try the extended
735 version, and if it fails with ENOTTY fall back 584 version, and if it fails with ENOTTY fall back to the original
736 version. 585 version.
737 586
738 FS_IOC_GET_ENCRYPTION_POLICY_EX 587 FS_IOC_GET_ENCRYPTION_POLICY_EX
739 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 588 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
740 589
741 The FS_IOC_GET_ENCRYPTION_POLICY_EX ioctl retr 590 The FS_IOC_GET_ENCRYPTION_POLICY_EX ioctl retrieves the encryption
742 policy, if any, for a directory or regular fil 591 policy, if any, for a directory or regular file. No additional
743 permissions are required beyond the ability to 592 permissions are required beyond the ability to open the file. It
744 takes in a pointer to struct fscrypt_get_polic 593 takes in a pointer to struct fscrypt_get_policy_ex_arg,
745 defined as follows:: 594 defined as follows::
746 595
747 struct fscrypt_get_policy_ex_arg { 596 struct fscrypt_get_policy_ex_arg {
748 __u64 policy_size; /* input/output 597 __u64 policy_size; /* input/output */
749 union { 598 union {
750 __u8 version; 599 __u8 version;
751 struct fscrypt_policy_v1 v 600 struct fscrypt_policy_v1 v1;
752 struct fscrypt_policy_v2 v 601 struct fscrypt_policy_v2 v2;
753 } policy; /* output */ 602 } policy; /* output */
754 }; 603 };
755 604
756 The caller must initialize ``policy_size`` to 605 The caller must initialize ``policy_size`` to the size available for
757 the policy struct, i.e. ``sizeof(arg.policy)`` 606 the policy struct, i.e. ``sizeof(arg.policy)``.
758 607
759 On success, the policy struct is returned in ` 608 On success, the policy struct is returned in ``policy``, and its
760 actual size is returned in ``policy_size``. ` 609 actual size is returned in ``policy_size``. ``policy.version`` should
761 be checked to determine the version of policy 610 be checked to determine the version of policy returned. Note that the
762 version code for the "v1" policy is actually 0 611 version code for the "v1" policy is actually 0 (FSCRYPT_POLICY_V1).
763 612
764 FS_IOC_GET_ENCRYPTION_POLICY_EX can fail with 613 FS_IOC_GET_ENCRYPTION_POLICY_EX can fail with the following errors:
765 614
766 - ``EINVAL``: the file is encrypted, but it us 615 - ``EINVAL``: the file is encrypted, but it uses an unrecognized
767 encryption policy version 616 encryption policy version
768 - ``ENODATA``: the file is not encrypted 617 - ``ENODATA``: the file is not encrypted
769 - ``ENOTTY``: this type of filesystem does not 618 - ``ENOTTY``: this type of filesystem does not implement encryption,
770 or this kernel is too old to support FS_IOC_ 619 or this kernel is too old to support FS_IOC_GET_ENCRYPTION_POLICY_EX
771 (try FS_IOC_GET_ENCRYPTION_POLICY instead) 620 (try FS_IOC_GET_ENCRYPTION_POLICY instead)
772 - ``EOPNOTSUPP``: the kernel was not configure 621 - ``EOPNOTSUPP``: the kernel was not configured with encryption
773 support for this filesystem, or the filesyst 622 support for this filesystem, or the filesystem superblock has not
774 had encryption enabled on it 623 had encryption enabled on it
775 - ``EOVERFLOW``: the file is encrypted and use 624 - ``EOVERFLOW``: the file is encrypted and uses a recognized
776 encryption policy version, but the policy st 625 encryption policy version, but the policy struct does not fit into
777 the provided buffer 626 the provided buffer
778 627
779 Note: if you only need to know whether a file 628 Note: if you only need to know whether a file is encrypted or not, on
780 most filesystems it is also possible to use th 629 most filesystems it is also possible to use the FS_IOC_GETFLAGS ioctl
781 and check for FS_ENCRYPT_FL, or to use the sta 630 and check for FS_ENCRYPT_FL, or to use the statx() system call and
782 check for STATX_ATTR_ENCRYPTED in stx_attribut 631 check for STATX_ATTR_ENCRYPTED in stx_attributes.
783 632
784 FS_IOC_GET_ENCRYPTION_POLICY 633 FS_IOC_GET_ENCRYPTION_POLICY
785 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 634 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
786 635
787 The FS_IOC_GET_ENCRYPTION_POLICY ioctl can als 636 The FS_IOC_GET_ENCRYPTION_POLICY ioctl can also retrieve the
788 encryption policy, if any, for a directory or 637 encryption policy, if any, for a directory or regular file. However,
789 unlike `FS_IOC_GET_ENCRYPTION_POLICY_EX`_, 638 unlike `FS_IOC_GET_ENCRYPTION_POLICY_EX`_,
790 FS_IOC_GET_ENCRYPTION_POLICY only supports the 639 FS_IOC_GET_ENCRYPTION_POLICY only supports the original policy
791 version. It takes in a pointer directly to st 640 version. It takes in a pointer directly to struct fscrypt_policy_v1
792 rather than struct fscrypt_get_policy_ex_arg. 641 rather than struct fscrypt_get_policy_ex_arg.
793 642
794 The error codes for FS_IOC_GET_ENCRYPTION_POLI 643 The error codes for FS_IOC_GET_ENCRYPTION_POLICY are the same as those
795 for FS_IOC_GET_ENCRYPTION_POLICY_EX, except th 644 for FS_IOC_GET_ENCRYPTION_POLICY_EX, except that
796 FS_IOC_GET_ENCRYPTION_POLICY also returns ``EI 645 FS_IOC_GET_ENCRYPTION_POLICY also returns ``EINVAL`` if the file is
797 encrypted using a newer encryption policy vers 646 encrypted using a newer encryption policy version.
798 647
799 Getting the per-filesystem salt 648 Getting the per-filesystem salt
800 ------------------------------- 649 -------------------------------
801 650
802 Some filesystems, such as ext4 and F2FS, also 651 Some filesystems, such as ext4 and F2FS, also support the deprecated
803 ioctl FS_IOC_GET_ENCRYPTION_PWSALT. This ioct 652 ioctl FS_IOC_GET_ENCRYPTION_PWSALT. This ioctl retrieves a randomly
804 generated 16-byte value stored in the filesyst 653 generated 16-byte value stored in the filesystem superblock. This
805 value is intended to used as a salt when deriv 654 value is intended to used as a salt when deriving an encryption key
806 from a passphrase or other low-entropy user cr 655 from a passphrase or other low-entropy user credential.
807 656
808 FS_IOC_GET_ENCRYPTION_PWSALT is deprecated. I 657 FS_IOC_GET_ENCRYPTION_PWSALT is deprecated. Instead, prefer to
809 generate and manage any needed salt(s) in user 658 generate and manage any needed salt(s) in userspace.
810 659
811 Getting a file's encryption nonce 660 Getting a file's encryption nonce
812 --------------------------------- 661 ---------------------------------
813 662
814 Since Linux v5.7, the ioctl FS_IOC_GET_ENCRYPT 663 Since Linux v5.7, the ioctl FS_IOC_GET_ENCRYPTION_NONCE is supported.
815 On encrypted files and directories it gets the 664 On encrypted files and directories it gets the inode's 16-byte nonce.
816 On unencrypted files and directories, it fails 665 On unencrypted files and directories, it fails with ENODATA.
817 666
818 This ioctl can be useful for automated tests w 667 This ioctl can be useful for automated tests which verify that the
819 encryption is being done correctly. It is not 668 encryption is being done correctly. It is not needed for normal use
820 of fscrypt. 669 of fscrypt.
821 670
822 Adding keys 671 Adding keys
823 ----------- 672 -----------
824 673
825 FS_IOC_ADD_ENCRYPTION_KEY 674 FS_IOC_ADD_ENCRYPTION_KEY
826 ~~~~~~~~~~~~~~~~~~~~~~~~~ 675 ~~~~~~~~~~~~~~~~~~~~~~~~~
827 676
828 The FS_IOC_ADD_ENCRYPTION_KEY ioctl adds a mas 677 The FS_IOC_ADD_ENCRYPTION_KEY ioctl adds a master encryption key to
829 the filesystem, making all files on the filesy 678 the filesystem, making all files on the filesystem which were
830 encrypted using that key appear "unlocked", i. 679 encrypted using that key appear "unlocked", i.e. in plaintext form.
831 It can be executed on any file or directory on 680 It can be executed on any file or directory on the target filesystem,
832 but using the filesystem's root directory is r 681 but using the filesystem's root directory is recommended. It takes in
833 a pointer to struct fscrypt_add_key_arg, defin 682 a pointer to struct fscrypt_add_key_arg, defined as follows::
834 683
835 struct fscrypt_add_key_arg { 684 struct fscrypt_add_key_arg {
836 struct fscrypt_key_specifier key_s 685 struct fscrypt_key_specifier key_spec;
837 __u32 raw_size; 686 __u32 raw_size;
838 __u32 key_id; 687 __u32 key_id;
839 __u32 __reserved[8]; 688 __u32 __reserved[8];
840 __u8 raw[]; 689 __u8 raw[];
841 }; 690 };
842 691
843 #define FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR 692 #define FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR 1
844 #define FSCRYPT_KEY_SPEC_TYPE_IDENTIFIER 693 #define FSCRYPT_KEY_SPEC_TYPE_IDENTIFIER 2
845 694
846 struct fscrypt_key_specifier { 695 struct fscrypt_key_specifier {
847 __u32 type; /* one of FSCRYPT_ 696 __u32 type; /* one of FSCRYPT_KEY_SPEC_TYPE_* */
848 __u32 __reserved; 697 __u32 __reserved;
849 union { 698 union {
850 __u8 __reserved[32]; /* re 699 __u8 __reserved[32]; /* reserve some extra space */
851 __u8 descriptor[FSCRYPT_KE 700 __u8 descriptor[FSCRYPT_KEY_DESCRIPTOR_SIZE];
852 __u8 identifier[FSCRYPT_KE 701 __u8 identifier[FSCRYPT_KEY_IDENTIFIER_SIZE];
853 } u; 702 } u;
854 }; 703 };
855 704
856 struct fscrypt_provisioning_key_payload { 705 struct fscrypt_provisioning_key_payload {
857 __u32 type; 706 __u32 type;
858 __u32 __reserved; 707 __u32 __reserved;
859 __u8 raw[]; 708 __u8 raw[];
860 }; 709 };
861 710
862 struct fscrypt_add_key_arg must be zeroed, the 711 struct fscrypt_add_key_arg must be zeroed, then initialized
863 as follows: 712 as follows:
864 713
865 - If the key is being added for use by v1 encr 714 - If the key is being added for use by v1 encryption policies, then
866 ``key_spec.type`` must contain FSCRYPT_KEY_S 715 ``key_spec.type`` must contain FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR, and
867 ``key_spec.u.descriptor`` must contain the d 716 ``key_spec.u.descriptor`` must contain the descriptor of the key
868 being added, corresponding to the value in t 717 being added, corresponding to the value in the
869 ``master_key_descriptor`` field of struct fs 718 ``master_key_descriptor`` field of struct fscrypt_policy_v1.
870 To add this type of key, the calling process 719 To add this type of key, the calling process must have the
871 CAP_SYS_ADMIN capability in the initial user 720 CAP_SYS_ADMIN capability in the initial user namespace.
872 721
873 Alternatively, if the key is being added for 722 Alternatively, if the key is being added for use by v2 encryption
874 policies, then ``key_spec.type`` must contai 723 policies, then ``key_spec.type`` must contain
875 FSCRYPT_KEY_SPEC_TYPE_IDENTIFIER, and ``key_ 724 FSCRYPT_KEY_SPEC_TYPE_IDENTIFIER, and ``key_spec.u.identifier`` is
876 an *output* field which the kernel fills in 725 an *output* field which the kernel fills in with a cryptographic
877 hash of the key. To add this type of key, t 726 hash of the key. To add this type of key, the calling process does
878 not need any privileges. However, the numbe 727 not need any privileges. However, the number of keys that can be
879 added is limited by the user's quota for the 728 added is limited by the user's quota for the keyrings service (see
880 ``Documentation/security/keys/core.rst``). 729 ``Documentation/security/keys/core.rst``).
881 730
882 - ``raw_size`` must be the size of the ``raw`` 731 - ``raw_size`` must be the size of the ``raw`` key provided, in bytes.
883 Alternatively, if ``key_id`` is nonzero, thi 732 Alternatively, if ``key_id`` is nonzero, this field must be 0, since
884 in that case the size is implied by the spec 733 in that case the size is implied by the specified Linux keyring key.
885 734
886 - ``key_id`` is 0 if the raw key is given dire 735 - ``key_id`` is 0 if the raw key is given directly in the ``raw``
887 field. Otherwise ``key_id`` is the ID of a 736 field. Otherwise ``key_id`` is the ID of a Linux keyring key of
888 type "fscrypt-provisioning" whose payload is 737 type "fscrypt-provisioning" whose payload is
889 struct fscrypt_provisioning_key_payload whos 738 struct fscrypt_provisioning_key_payload whose ``raw`` field contains
890 the raw key and whose ``type`` field matches 739 the raw key and whose ``type`` field matches ``key_spec.type``.
891 Since ``raw`` is variable-length, the total 740 Since ``raw`` is variable-length, the total size of this key's
892 payload must be ``sizeof(struct fscrypt_prov 741 payload must be ``sizeof(struct fscrypt_provisioning_key_payload)``
893 plus the raw key size. The process must hav 742 plus the raw key size. The process must have Search permission on
894 this key. 743 this key.
895 744
896 Most users should leave this 0 and specify t 745 Most users should leave this 0 and specify the raw key directly.
897 The support for specifying a Linux keyring k 746 The support for specifying a Linux keyring key is intended mainly to
898 allow re-adding keys after a filesystem is u 747 allow re-adding keys after a filesystem is unmounted and re-mounted,
899 without having to store the raw keys in user 748 without having to store the raw keys in userspace memory.
900 749
901 - ``raw`` is a variable-length field which mus 750 - ``raw`` is a variable-length field which must contain the actual
902 key, ``raw_size`` bytes long. Alternatively 751 key, ``raw_size`` bytes long. Alternatively, if ``key_id`` is
903 nonzero, then this field is unused. 752 nonzero, then this field is unused.
904 753
905 For v2 policy keys, the kernel keeps track of 754 For v2 policy keys, the kernel keeps track of which user (identified
906 by effective user ID) added the key, and only 755 by effective user ID) added the key, and only allows the key to be
907 removed by that user --- or by "root", if they 756 removed by that user --- or by "root", if they use
908 `FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS`_. 757 `FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS`_.
909 758
910 However, if another user has added the key, it 759 However, if another user has added the key, it may be desirable to
911 prevent that other user from unexpectedly remo 760 prevent that other user from unexpectedly removing it. Therefore,
912 FS_IOC_ADD_ENCRYPTION_KEY may also be used to 761 FS_IOC_ADD_ENCRYPTION_KEY may also be used to add a v2 policy key
913 *again*, even if it's already added by other u 762 *again*, even if it's already added by other user(s). In this case,
914 FS_IOC_ADD_ENCRYPTION_KEY will just install a 763 FS_IOC_ADD_ENCRYPTION_KEY will just install a claim to the key for the
915 current user, rather than actually add the key 764 current user, rather than actually add the key again (but the raw key
916 must still be provided, as a proof of knowledg 765 must still be provided, as a proof of knowledge).
917 766
918 FS_IOC_ADD_ENCRYPTION_KEY returns 0 if either 767 FS_IOC_ADD_ENCRYPTION_KEY returns 0 if either the key or a claim to
919 the key was either added or already exists. 768 the key was either added or already exists.
920 769
921 FS_IOC_ADD_ENCRYPTION_KEY can fail with the fo 770 FS_IOC_ADD_ENCRYPTION_KEY can fail with the following errors:
922 771
923 - ``EACCES``: FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR 772 - ``EACCES``: FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR was specified, but the
924 caller does not have the CAP_SYS_ADMIN capab 773 caller does not have the CAP_SYS_ADMIN capability in the initial
925 user namespace; or the raw key was specified 774 user namespace; or the raw key was specified by Linux key ID but the
926 process lacks Search permission on the key. 775 process lacks Search permission on the key.
927 - ``EDQUOT``: the key quota for this user woul 776 - ``EDQUOT``: the key quota for this user would be exceeded by adding
928 the key 777 the key
929 - ``EINVAL``: invalid key size or key specifie 778 - ``EINVAL``: invalid key size or key specifier type, or reserved bits
930 were set 779 were set
931 - ``EKEYREJECTED``: the raw key was specified 780 - ``EKEYREJECTED``: the raw key was specified by Linux key ID, but the
932 key has the wrong type 781 key has the wrong type
933 - ``ENOKEY``: the raw key was specified by Lin 782 - ``ENOKEY``: the raw key was specified by Linux key ID, but no key
934 exists with that ID 783 exists with that ID
935 - ``ENOTTY``: this type of filesystem does not 784 - ``ENOTTY``: this type of filesystem does not implement encryption
936 - ``EOPNOTSUPP``: the kernel was not configure 785 - ``EOPNOTSUPP``: the kernel was not configured with encryption
937 support for this filesystem, or the filesyst 786 support for this filesystem, or the filesystem superblock has not
938 had encryption enabled on it 787 had encryption enabled on it
939 788
940 Legacy method 789 Legacy method
941 ~~~~~~~~~~~~~ 790 ~~~~~~~~~~~~~
942 791
943 For v1 encryption policies, a master encryptio 792 For v1 encryption policies, a master encryption key can also be
944 provided by adding it to a process-subscribed 793 provided by adding it to a process-subscribed keyring, e.g. to a
945 session keyring, or to a user keyring if the u 794 session keyring, or to a user keyring if the user keyring is linked
946 into the session keyring. 795 into the session keyring.
947 796
948 This method is deprecated (and not supported f 797 This method is deprecated (and not supported for v2 encryption
949 policies) for several reasons. First, it cann 798 policies) for several reasons. First, it cannot be used in
950 combination with FS_IOC_REMOVE_ENCRYPTION_KEY 799 combination with FS_IOC_REMOVE_ENCRYPTION_KEY (see `Removing keys`_),
951 so for removing a key a workaround such as key 800 so for removing a key a workaround such as keyctl_unlink() in
952 combination with ``sync; echo 2 > /proc/sys/vm 801 combination with ``sync; echo 2 > /proc/sys/vm/drop_caches`` would
953 have to be used. Second, it doesn't match the 802 have to be used. Second, it doesn't match the fact that the
954 locked/unlocked status of encrypted files (i.e 803 locked/unlocked status of encrypted files (i.e. whether they appear to
955 be in plaintext form or in ciphertext form) is 804 be in plaintext form or in ciphertext form) is global. This mismatch
956 has caused much confusion as well as real prob 805 has caused much confusion as well as real problems when processes
957 running under different UIDs, such as a ``sudo 806 running under different UIDs, such as a ``sudo`` command, need to
958 access encrypted files. 807 access encrypted files.
959 808
960 Nevertheless, to add a key to one of the proce 809 Nevertheless, to add a key to one of the process-subscribed keyrings,
961 the add_key() system call can be used (see: 810 the add_key() system call can be used (see:
962 ``Documentation/security/keys/core.rst``). Th 811 ``Documentation/security/keys/core.rst``). The key type must be
963 "logon"; keys of this type are kept in kernel 812 "logon"; keys of this type are kept in kernel memory and cannot be
964 read back by userspace. The key description m 813 read back by userspace. The key description must be "fscrypt:"
965 followed by the 16-character lower case hex re 814 followed by the 16-character lower case hex representation of the
966 ``master_key_descriptor`` that was set in the 815 ``master_key_descriptor`` that was set in the encryption policy. The
967 key payload must conform to the following stru 816 key payload must conform to the following structure::
968 817
969 #define FSCRYPT_MAX_KEY_SIZE 64 818 #define FSCRYPT_MAX_KEY_SIZE 64
970 819
971 struct fscrypt_key { 820 struct fscrypt_key {
972 __u32 mode; 821 __u32 mode;
973 __u8 raw[FSCRYPT_MAX_KEY_SIZE]; 822 __u8 raw[FSCRYPT_MAX_KEY_SIZE];
974 __u32 size; 823 __u32 size;
975 }; 824 };
976 825
977 ``mode`` is ignored; just set it to 0. The ac 826 ``mode`` is ignored; just set it to 0. The actual key is provided in
978 ``raw`` with ``size`` indicating its size in b 827 ``raw`` with ``size`` indicating its size in bytes. That is, the
979 bytes ``raw[0..size-1]`` (inclusive) are the a 828 bytes ``raw[0..size-1]`` (inclusive) are the actual key.
980 829
981 The key description prefix "fscrypt:" may alte 830 The key description prefix "fscrypt:" may alternatively be replaced
982 with a filesystem-specific prefix such as "ext 831 with a filesystem-specific prefix such as "ext4:". However, the
983 filesystem-specific prefixes are deprecated an 832 filesystem-specific prefixes are deprecated and should not be used in
984 new programs. 833 new programs.
985 834
986 Removing keys 835 Removing keys
987 ------------- 836 -------------
988 837
989 Two ioctls are available for removing a key th 838 Two ioctls are available for removing a key that was added by
990 `FS_IOC_ADD_ENCRYPTION_KEY`_: 839 `FS_IOC_ADD_ENCRYPTION_KEY`_:
991 840
992 - `FS_IOC_REMOVE_ENCRYPTION_KEY`_ 841 - `FS_IOC_REMOVE_ENCRYPTION_KEY`_
993 - `FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS`_ 842 - `FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS`_
994 843
995 These two ioctls differ only in cases where v2 844 These two ioctls differ only in cases where v2 policy keys are added
996 or removed by non-root users. 845 or removed by non-root users.
997 846
998 These ioctls don't work on keys that were adde 847 These ioctls don't work on keys that were added via the legacy
999 process-subscribed keyrings mechanism. 848 process-subscribed keyrings mechanism.
1000 849
1001 Before using these ioctls, read the `Kernel m 850 Before using these ioctls, read the `Kernel memory compromise`_
1002 section for a discussion of the security goal 851 section for a discussion of the security goals and limitations of
1003 these ioctls. 852 these ioctls.
1004 853
1005 FS_IOC_REMOVE_ENCRYPTION_KEY 854 FS_IOC_REMOVE_ENCRYPTION_KEY
1006 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 855 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1007 856
1008 The FS_IOC_REMOVE_ENCRYPTION_KEY ioctl remove 857 The FS_IOC_REMOVE_ENCRYPTION_KEY ioctl removes a claim to a master
1009 encryption key from the filesystem, and possi 858 encryption key from the filesystem, and possibly removes the key
1010 itself. It can be executed on any file or di 859 itself. It can be executed on any file or directory on the target
1011 filesystem, but using the filesystem's root d 860 filesystem, but using the filesystem's root directory is recommended.
1012 It takes in a pointer to struct fscrypt_remov 861 It takes in a pointer to struct fscrypt_remove_key_arg, defined
1013 as follows:: 862 as follows::
1014 863
1015 struct fscrypt_remove_key_arg { 864 struct fscrypt_remove_key_arg {
1016 struct fscrypt_key_specifier key_ 865 struct fscrypt_key_specifier key_spec;
1017 #define FSCRYPT_KEY_REMOVAL_STATUS_FLAG_F 866 #define FSCRYPT_KEY_REMOVAL_STATUS_FLAG_FILES_BUSY 0x00000001
1018 #define FSCRYPT_KEY_REMOVAL_STATUS_FLAG_O 867 #define FSCRYPT_KEY_REMOVAL_STATUS_FLAG_OTHER_USERS 0x00000002
1019 __u32 removal_status_flags; / 868 __u32 removal_status_flags; /* output */
1020 __u32 __reserved[5]; 869 __u32 __reserved[5];
1021 }; 870 };
1022 871
1023 This structure must be zeroed, then initializ 872 This structure must be zeroed, then initialized as follows:
1024 873
1025 - The key to remove is specified by ``key_spe 874 - The key to remove is specified by ``key_spec``:
1026 875
1027 - To remove a key used by v1 encryption p 876 - To remove a key used by v1 encryption policies, set
1028 ``key_spec.type`` to FSCRYPT_KEY_SPEC_T 877 ``key_spec.type`` to FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR and fill
1029 in ``key_spec.u.descriptor``. To remov 878 in ``key_spec.u.descriptor``. To remove this type of key, the
1030 calling process must have the CAP_SYS_A 879 calling process must have the CAP_SYS_ADMIN capability in the
1031 initial user namespace. 880 initial user namespace.
1032 881
1033 - To remove a key used by v2 encryption p 882 - To remove a key used by v2 encryption policies, set
1034 ``key_spec.type`` to FSCRYPT_KEY_SPEC_T 883 ``key_spec.type`` to FSCRYPT_KEY_SPEC_TYPE_IDENTIFIER and fill
1035 in ``key_spec.u.identifier``. 884 in ``key_spec.u.identifier``.
1036 885
1037 For v2 policy keys, this ioctl is usable by n 886 For v2 policy keys, this ioctl is usable by non-root users. However,
1038 to make this possible, it actually just remov 887 to make this possible, it actually just removes the current user's
1039 claim to the key, undoing a single call to FS 888 claim to the key, undoing a single call to FS_IOC_ADD_ENCRYPTION_KEY.
1040 Only after all claims are removed is the key 889 Only after all claims are removed is the key really removed.
1041 890
1042 For example, if FS_IOC_ADD_ENCRYPTION_KEY was 891 For example, if FS_IOC_ADD_ENCRYPTION_KEY was called with uid 1000,
1043 then the key will be "claimed" by uid 1000, a 892 then the key will be "claimed" by uid 1000, and
1044 FS_IOC_REMOVE_ENCRYPTION_KEY will only succee 893 FS_IOC_REMOVE_ENCRYPTION_KEY will only succeed as uid 1000. Or, if
1045 both uids 1000 and 2000 added the key, then f 894 both uids 1000 and 2000 added the key, then for each uid
1046 FS_IOC_REMOVE_ENCRYPTION_KEY will only remove 895 FS_IOC_REMOVE_ENCRYPTION_KEY will only remove their own claim. Only
1047 once *both* are removed is the key really rem 896 once *both* are removed is the key really removed. (Think of it like
1048 unlinking a file that may have hard links.) 897 unlinking a file that may have hard links.)
1049 898
1050 If FS_IOC_REMOVE_ENCRYPTION_KEY really remove 899 If FS_IOC_REMOVE_ENCRYPTION_KEY really removes the key, it will also
1051 try to "lock" all files that had been unlocke 900 try to "lock" all files that had been unlocked with the key. It won't
1052 lock files that are still in-use, so this ioc 901 lock files that are still in-use, so this ioctl is expected to be used
1053 in cooperation with userspace ensuring that n 902 in cooperation with userspace ensuring that none of the files are
1054 still open. However, if necessary, this ioct 903 still open. However, if necessary, this ioctl can be executed again
1055 later to retry locking any remaining files. 904 later to retry locking any remaining files.
1056 905
1057 FS_IOC_REMOVE_ENCRYPTION_KEY returns 0 if eit 906 FS_IOC_REMOVE_ENCRYPTION_KEY returns 0 if either the key was removed
1058 (but may still have files remaining to be loc 907 (but may still have files remaining to be locked), the user's claim to
1059 the key was removed, or the key was already r 908 the key was removed, or the key was already removed but had files
1060 remaining to be the locked so the ioctl retri 909 remaining to be the locked so the ioctl retried locking them. In any
1061 of these cases, ``removal_status_flags`` is f 910 of these cases, ``removal_status_flags`` is filled in with the
1062 following informational status flags: 911 following informational status flags:
1063 912
1064 - ``FSCRYPT_KEY_REMOVAL_STATUS_FLAG_FILES_BUS 913 - ``FSCRYPT_KEY_REMOVAL_STATUS_FLAG_FILES_BUSY``: set if some file(s)
1065 are still in-use. Not guaranteed to be set 914 are still in-use. Not guaranteed to be set in the case where only
1066 the user's claim to the key was removed. 915 the user's claim to the key was removed.
1067 - ``FSCRYPT_KEY_REMOVAL_STATUS_FLAG_OTHER_USE 916 - ``FSCRYPT_KEY_REMOVAL_STATUS_FLAG_OTHER_USERS``: set if only the
1068 user's claim to the key was removed, not th 917 user's claim to the key was removed, not the key itself
1069 918
1070 FS_IOC_REMOVE_ENCRYPTION_KEY can fail with th 919 FS_IOC_REMOVE_ENCRYPTION_KEY can fail with the following errors:
1071 920
1072 - ``EACCES``: The FSCRYPT_KEY_SPEC_TYPE_DESCR 921 - ``EACCES``: The FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR key specifier type
1073 was specified, but the caller does not have 922 was specified, but the caller does not have the CAP_SYS_ADMIN
1074 capability in the initial user namespace 923 capability in the initial user namespace
1075 - ``EINVAL``: invalid key specifier type, or 924 - ``EINVAL``: invalid key specifier type, or reserved bits were set
1076 - ``ENOKEY``: the key object was not found at 925 - ``ENOKEY``: the key object was not found at all, i.e. it was never
1077 added in the first place or was already ful 926 added in the first place or was already fully removed including all
1078 files locked; or, the user does not have a 927 files locked; or, the user does not have a claim to the key (but
1079 someone else does). 928 someone else does).
1080 - ``ENOTTY``: this type of filesystem does no 929 - ``ENOTTY``: this type of filesystem does not implement encryption
1081 - ``EOPNOTSUPP``: the kernel was not configur 930 - ``EOPNOTSUPP``: the kernel was not configured with encryption
1082 support for this filesystem, or the filesys 931 support for this filesystem, or the filesystem superblock has not
1083 had encryption enabled on it 932 had encryption enabled on it
1084 933
1085 FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS 934 FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS
1086 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 935 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1087 936
1088 FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS is exa 937 FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS is exactly the same as
1089 `FS_IOC_REMOVE_ENCRYPTION_KEY`_, except that 938 `FS_IOC_REMOVE_ENCRYPTION_KEY`_, except that for v2 policy keys, the
1090 ALL_USERS version of the ioctl will remove al 939 ALL_USERS version of the ioctl will remove all users' claims to the
1091 key, not just the current user's. I.e., the 940 key, not just the current user's. I.e., the key itself will always be
1092 removed, no matter how many users have added 941 removed, no matter how many users have added it. This difference is
1093 only meaningful if non-root users are adding 942 only meaningful if non-root users are adding and removing keys.
1094 943
1095 Because of this, FS_IOC_REMOVE_ENCRYPTION_KEY 944 Because of this, FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS also requires
1096 "root", namely the CAP_SYS_ADMIN capability i 945 "root", namely the CAP_SYS_ADMIN capability in the initial user
1097 namespace. Otherwise it will fail with EACCE 946 namespace. Otherwise it will fail with EACCES.
1098 947
1099 Getting key status 948 Getting key status
1100 ------------------ 949 ------------------
1101 950
1102 FS_IOC_GET_ENCRYPTION_KEY_STATUS 951 FS_IOC_GET_ENCRYPTION_KEY_STATUS
1103 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 952 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1104 953
1105 The FS_IOC_GET_ENCRYPTION_KEY_STATUS ioctl re 954 The FS_IOC_GET_ENCRYPTION_KEY_STATUS ioctl retrieves the status of a
1106 master encryption key. It can be executed on 955 master encryption key. It can be executed on any file or directory on
1107 the target filesystem, but using the filesyst 956 the target filesystem, but using the filesystem's root directory is
1108 recommended. It takes in a pointer to 957 recommended. It takes in a pointer to
1109 struct fscrypt_get_key_status_arg, defined as 958 struct fscrypt_get_key_status_arg, defined as follows::
1110 959
1111 struct fscrypt_get_key_status_arg { 960 struct fscrypt_get_key_status_arg {
1112 /* input */ 961 /* input */
1113 struct fscrypt_key_specifier key_ 962 struct fscrypt_key_specifier key_spec;
1114 __u32 __reserved[6]; 963 __u32 __reserved[6];
1115 964
1116 /* output */ 965 /* output */
1117 #define FSCRYPT_KEY_STATUS_ABSENT 966 #define FSCRYPT_KEY_STATUS_ABSENT 1
1118 #define FSCRYPT_KEY_STATUS_PRESENT 967 #define FSCRYPT_KEY_STATUS_PRESENT 2
1119 #define FSCRYPT_KEY_STATUS_INCOMPLETELY_R 968 #define FSCRYPT_KEY_STATUS_INCOMPLETELY_REMOVED 3
1120 __u32 status; 969 __u32 status;
1121 #define FSCRYPT_KEY_STATUS_FLAG_ADDED_BY_ 970 #define FSCRYPT_KEY_STATUS_FLAG_ADDED_BY_SELF 0x00000001
1122 __u32 status_flags; 971 __u32 status_flags;
1123 __u32 user_count; 972 __u32 user_count;
1124 __u32 __out_reserved[13]; 973 __u32 __out_reserved[13];
1125 }; 974 };
1126 975
1127 The caller must zero all input fields, then f 976 The caller must zero all input fields, then fill in ``key_spec``:
1128 977
1129 - To get the status of a key for v1 encry 978 - To get the status of a key for v1 encryption policies, set
1130 ``key_spec.type`` to FSCRYPT_KEY_SPEC_T 979 ``key_spec.type`` to FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR and fill
1131 in ``key_spec.u.descriptor``. 980 in ``key_spec.u.descriptor``.
1132 981
1133 - To get the status of a key for v2 encry 982 - To get the status of a key for v2 encryption policies, set
1134 ``key_spec.type`` to FSCRYPT_KEY_SPEC_T 983 ``key_spec.type`` to FSCRYPT_KEY_SPEC_TYPE_IDENTIFIER and fill
1135 in ``key_spec.u.identifier``. 984 in ``key_spec.u.identifier``.
1136 985
1137 On success, 0 is returned and the kernel fill 986 On success, 0 is returned and the kernel fills in the output fields:
1138 987
1139 - ``status`` indicates whether the key is abs 988 - ``status`` indicates whether the key is absent, present, or
1140 incompletely removed. Incompletely removed !! 989 incompletely removed. Incompletely removed means that the master
1141 been initiated, but some files are still in !! 990 secret has been removed, but some files are still in use; i.e.,
1142 `FS_IOC_REMOVE_ENCRYPTION_KEY`_ returned 0 991 `FS_IOC_REMOVE_ENCRYPTION_KEY`_ returned 0 but set the informational
1143 status flag FSCRYPT_KEY_REMOVAL_STATUS_FLAG 992 status flag FSCRYPT_KEY_REMOVAL_STATUS_FLAG_FILES_BUSY.
1144 993
1145 - ``status_flags`` can contain the following 994 - ``status_flags`` can contain the following flags:
1146 995
1147 - ``FSCRYPT_KEY_STATUS_FLAG_ADDED_BY_SELF 996 - ``FSCRYPT_KEY_STATUS_FLAG_ADDED_BY_SELF`` indicates that the key
1148 has added by the current user. This is 997 has added by the current user. This is only set for keys
1149 identified by ``identifier`` rather tha 998 identified by ``identifier`` rather than by ``descriptor``.
1150 999
1151 - ``user_count`` specifies the number of user 1000 - ``user_count`` specifies the number of users who have added the key.
1152 This is only set for keys identified by ``i 1001 This is only set for keys identified by ``identifier`` rather than
1153 by ``descriptor``. 1002 by ``descriptor``.
1154 1003
1155 FS_IOC_GET_ENCRYPTION_KEY_STATUS can fail wit 1004 FS_IOC_GET_ENCRYPTION_KEY_STATUS can fail with the following errors:
1156 1005
1157 - ``EINVAL``: invalid key specifier type, or 1006 - ``EINVAL``: invalid key specifier type, or reserved bits were set
1158 - ``ENOTTY``: this type of filesystem does no 1007 - ``ENOTTY``: this type of filesystem does not implement encryption
1159 - ``EOPNOTSUPP``: the kernel was not configur 1008 - ``EOPNOTSUPP``: the kernel was not configured with encryption
1160 support for this filesystem, or the filesys 1009 support for this filesystem, or the filesystem superblock has not
1161 had encryption enabled on it 1010 had encryption enabled on it
1162 1011
1163 Among other use cases, FS_IOC_GET_ENCRYPTION_ 1012 Among other use cases, FS_IOC_GET_ENCRYPTION_KEY_STATUS can be useful
1164 for determining whether the key for a given e 1013 for determining whether the key for a given encrypted directory needs
1165 to be added before prompting the user for the 1014 to be added before prompting the user for the passphrase needed to
1166 derive the key. 1015 derive the key.
1167 1016
1168 FS_IOC_GET_ENCRYPTION_KEY_STATUS can only get 1017 FS_IOC_GET_ENCRYPTION_KEY_STATUS can only get the status of keys in
1169 the filesystem-level keyring, i.e. the keyrin 1018 the filesystem-level keyring, i.e. the keyring managed by
1170 `FS_IOC_ADD_ENCRYPTION_KEY`_ and `FS_IOC_REMO 1019 `FS_IOC_ADD_ENCRYPTION_KEY`_ and `FS_IOC_REMOVE_ENCRYPTION_KEY`_. It
1171 cannot get the status of a key that has only 1020 cannot get the status of a key that has only been added for use by v1
1172 encryption policies using the legacy mechanis 1021 encryption policies using the legacy mechanism involving
1173 process-subscribed keyrings. 1022 process-subscribed keyrings.
1174 1023
1175 Access semantics 1024 Access semantics
1176 ================ 1025 ================
1177 1026
1178 With the key 1027 With the key
1179 ------------ 1028 ------------
1180 1029
1181 With the encryption key, encrypted regular fi 1030 With the encryption key, encrypted regular files, directories, and
1182 symlinks behave very similarly to their unenc 1031 symlinks behave very similarly to their unencrypted counterparts ---
1183 after all, the encryption is intended to be t 1032 after all, the encryption is intended to be transparent. However,
1184 astute users may notice some differences in b 1033 astute users may notice some differences in behavior:
1185 1034
1186 - Unencrypted files, or files encrypted with 1035 - Unencrypted files, or files encrypted with a different encryption
1187 policy (i.e. different key, modes, or flags 1036 policy (i.e. different key, modes, or flags), cannot be renamed or
1188 linked into an encrypted directory; see `En 1037 linked into an encrypted directory; see `Encryption policy
1189 enforcement`_. Attempts to do so will fail 1038 enforcement`_. Attempts to do so will fail with EXDEV. However,
1190 encrypted files can be renamed within an en 1039 encrypted files can be renamed within an encrypted directory, or
1191 into an unencrypted directory. 1040 into an unencrypted directory.
1192 1041
1193 Note: "moving" an unencrypted file into an 1042 Note: "moving" an unencrypted file into an encrypted directory, e.g.
1194 with the `mv` program, is implemented in us 1043 with the `mv` program, is implemented in userspace by a copy
1195 followed by a delete. Be aware that the or 1044 followed by a delete. Be aware that the original unencrypted data
1196 may remain recoverable from free space on t 1045 may remain recoverable from free space on the disk; prefer to keep
1197 all files encrypted from the very beginning 1046 all files encrypted from the very beginning. The `shred` program
1198 may be used to overwrite the source files b 1047 may be used to overwrite the source files but isn't guaranteed to be
1199 effective on all filesystems and storage de 1048 effective on all filesystems and storage devices.
1200 1049
1201 - Direct I/O is supported on encrypted files !! 1050 - Direct I/O is not supported on encrypted files. Attempts to use
1202 circumstances. For details, see `Direct I/ !! 1051 direct I/O on such files will fall back to buffered I/O.
1203 1052
1204 - The fallocate operations FALLOC_FL_COLLAPSE 1053 - The fallocate operations FALLOC_FL_COLLAPSE_RANGE and
1205 FALLOC_FL_INSERT_RANGE are not supported on 1054 FALLOC_FL_INSERT_RANGE are not supported on encrypted files and will
1206 fail with EOPNOTSUPP. 1055 fail with EOPNOTSUPP.
1207 1056
1208 - Online defragmentation of encrypted files i 1057 - Online defragmentation of encrypted files is not supported. The
1209 EXT4_IOC_MOVE_EXT and F2FS_IOC_MOVE_RANGE i 1058 EXT4_IOC_MOVE_EXT and F2FS_IOC_MOVE_RANGE ioctls will fail with
1210 EOPNOTSUPP. 1059 EOPNOTSUPP.
1211 1060
1212 - The ext4 filesystem does not support data j 1061 - The ext4 filesystem does not support data journaling with encrypted
1213 regular files. It will fall back to ordere 1062 regular files. It will fall back to ordered data mode instead.
1214 1063
1215 - DAX (Direct Access) is not supported on enc 1064 - DAX (Direct Access) is not supported on encrypted files.
1216 1065
>> 1066 - The st_size of an encrypted symlink will not necessarily give the
>> 1067 length of the symlink target as required by POSIX. It will actually
>> 1068 give the length of the ciphertext, which will be slightly longer
>> 1069 than the plaintext due to NUL-padding and an extra 2-byte overhead.
>> 1070
1217 - The maximum length of an encrypted symlink 1071 - The maximum length of an encrypted symlink is 2 bytes shorter than
1218 the maximum length of an unencrypted symlin 1072 the maximum length of an unencrypted symlink. For example, on an
1219 EXT4 filesystem with a 4K block size, unenc 1073 EXT4 filesystem with a 4K block size, unencrypted symlinks can be up
1220 to 4095 bytes long, while encrypted symlink 1074 to 4095 bytes long, while encrypted symlinks can only be up to 4093
1221 bytes long (both lengths excluding the term 1075 bytes long (both lengths excluding the terminating null).
1222 1076
1223 Note that mmap *is* supported. This is possi 1077 Note that mmap *is* supported. This is possible because the pagecache
1224 for an encrypted file contains the plaintext, 1078 for an encrypted file contains the plaintext, not the ciphertext.
1225 1079
1226 Without the key 1080 Without the key
1227 --------------- 1081 ---------------
1228 1082
1229 Some filesystem operations may be performed o 1083 Some filesystem operations may be performed on encrypted regular
1230 files, directories, and symlinks even before 1084 files, directories, and symlinks even before their encryption key has
1231 been added, or after their encryption key has 1085 been added, or after their encryption key has been removed:
1232 1086
1233 - File metadata may be read, e.g. using stat( 1087 - File metadata may be read, e.g. using stat().
1234 1088
1235 - Directories may be listed, in which case th 1089 - Directories may be listed, in which case the filenames will be
1236 listed in an encoded form derived from thei 1090 listed in an encoded form derived from their ciphertext. The
1237 current encoding algorithm is described in 1091 current encoding algorithm is described in `Filename hashing and
1238 encoding`_. The algorithm is subject to ch 1092 encoding`_. The algorithm is subject to change, but it is
1239 guaranteed that the presented filenames wil 1093 guaranteed that the presented filenames will be no longer than
1240 NAME_MAX bytes, will not contain the ``/`` 1094 NAME_MAX bytes, will not contain the ``/`` or ``\0`` characters, and
1241 will uniquely identify directory entries. 1095 will uniquely identify directory entries.
1242 1096
1243 The ``.`` and ``..`` directory entries are 1097 The ``.`` and ``..`` directory entries are special. They are always
1244 present and are not encrypted or encoded. 1098 present and are not encrypted or encoded.
1245 1099
1246 - Files may be deleted. That is, nondirector 1100 - Files may be deleted. That is, nondirectory files may be deleted
1247 with unlink() as usual, and empty directori 1101 with unlink() as usual, and empty directories may be deleted with
1248 rmdir() as usual. Therefore, ``rm`` and `` 1102 rmdir() as usual. Therefore, ``rm`` and ``rm -r`` will work as
1249 expected. 1103 expected.
1250 1104
1251 - Symlink targets may be read and followed, b 1105 - Symlink targets may be read and followed, but they will be presented
1252 in encrypted form, similar to filenames in 1106 in encrypted form, similar to filenames in directories. Hence, they
1253 are unlikely to point to anywhere useful. 1107 are unlikely to point to anywhere useful.
1254 1108
1255 Without the key, regular files cannot be open 1109 Without the key, regular files cannot be opened or truncated.
1256 Attempts to do so will fail with ENOKEY. Thi 1110 Attempts to do so will fail with ENOKEY. This implies that any
1257 regular file operations that require a file d 1111 regular file operations that require a file descriptor, such as
1258 read(), write(), mmap(), fallocate(), and ioc 1112 read(), write(), mmap(), fallocate(), and ioctl(), are also forbidden.
1259 1113
1260 Also without the key, files of any type (incl 1114 Also without the key, files of any type (including directories) cannot
1261 be created or linked into an encrypted direct 1115 be created or linked into an encrypted directory, nor can a name in an
1262 encrypted directory be the source or target o 1116 encrypted directory be the source or target of a rename, nor can an
1263 O_TMPFILE temporary file be created in an enc 1117 O_TMPFILE temporary file be created in an encrypted directory. All
1264 such operations will fail with ENOKEY. 1118 such operations will fail with ENOKEY.
1265 1119
1266 It is not currently possible to backup and re 1120 It is not currently possible to backup and restore encrypted files
1267 without the encryption key. This would requi 1121 without the encryption key. This would require special APIs which
1268 have not yet been implemented. 1122 have not yet been implemented.
1269 1123
1270 Encryption policy enforcement 1124 Encryption policy enforcement
1271 ============================= 1125 =============================
1272 1126
1273 After an encryption policy has been set on a 1127 After an encryption policy has been set on a directory, all regular
1274 files, directories, and symbolic links create 1128 files, directories, and symbolic links created in that directory
1275 (recursively) will inherit that encryption po 1129 (recursively) will inherit that encryption policy. Special files ---
1276 that is, named pipes, device nodes, and UNIX 1130 that is, named pipes, device nodes, and UNIX domain sockets --- will
1277 not be encrypted. 1131 not be encrypted.
1278 1132
1279 Except for those special files, it is forbidd 1133 Except for those special files, it is forbidden to have unencrypted
1280 files, or files encrypted with a different en 1134 files, or files encrypted with a different encryption policy, in an
1281 encrypted directory tree. Attempts to link o 1135 encrypted directory tree. Attempts to link or rename such a file into
1282 an encrypted directory will fail with EXDEV. 1136 an encrypted directory will fail with EXDEV. This is also enforced
1283 during ->lookup() to provide limited protecti 1137 during ->lookup() to provide limited protection against offline
1284 attacks that try to disable or downgrade encr 1138 attacks that try to disable or downgrade encryption in known locations
1285 where applications may later write sensitive 1139 where applications may later write sensitive data. It is recommended
1286 that systems implementing a form of "verified 1140 that systems implementing a form of "verified boot" take advantage of
1287 this by validating all top-level encryption p 1141 this by validating all top-level encryption policies prior to access.
1288 1142
1289 Inline encryption support <<
1290 ========================= <<
1291 <<
1292 By default, fscrypt uses the kernel crypto AP <<
1293 operations (other than HKDF, which fscrypt pa <<
1294 itself). The kernel crypto API supports hard <<
1295 but only ones that work in the traditional wa <<
1296 outputs (e.g. plaintexts and ciphertexts) are <<
1297 take advantage of such hardware, but the trad <<
1298 model isn't particularly efficient and fscryp <<
1299 for it. <<
1300 <<
1301 Instead, many newer systems (especially mobil <<
1302 encryption hardware* that can encrypt/decrypt <<
1303 way to/from the storage device. Linux suppor <<
1304 through a set of extensions to the block laye <<
1305 blk-crypto allows filesystems to attach encry <<
1306 (I/O requests) to specify how the data will b <<
1307 in-line. For more information about blk-cryp <<
1308 :ref:`Documentation/block/inline-encryption.r <<
1309 <<
1310 On supported filesystems (currently ext4 and <<
1311 blk-crypto instead of the kernel crypto API t <<
1312 contents. To enable this, set CONFIG_FS_ENCR <<
1313 the kernel configuration, and specify the "in <<
1314 when mounting the filesystem. <<
1315 <<
1316 Note that the "inlinecrypt" mount option just <<
1317 encryption when possible; it doesn't force it <<
1318 still fall back to using the kernel crypto AP <<
1319 inline encryption hardware doesn't have the n <<
1320 (e.g. support for the needed encryption algor <<
1321 and where blk-crypto-fallback is unusable. ( <<
1322 to be usable, it must be enabled in the kerne <<
1323 CONFIG_BLK_INLINE_ENCRYPTION_FALLBACK=y.) <<
1324 <<
1325 Currently fscrypt always uses the filesystem <<
1326 usually 4096 bytes) as the data unit size. T <<
1327 inline encryption hardware that supports that <<
1328 <<
1329 Inline encryption doesn't affect the cipherte <<
1330 the on-disk format, so users may freely switc <<
1331 using "inlinecrypt" and not using "inlinecryp <<
1332 <<
1333 Direct I/O support <<
1334 ================== <<
1335 <<
1336 For direct I/O on an encrypted file to work, <<
1337 must be met (in addition to the conditions fo <<
1338 unencrypted file): <<
1339 <<
1340 * The file must be using inline encryption. <<
1341 the filesystem must be mounted with ``-o in <<
1342 encryption hardware must be present. Howev <<
1343 is also available. For details, see `Inlin <<
1344 <<
1345 * The I/O request must be fully aligned to th <<
1346 This means that the file position the I/O i <<
1347 of all I/O segments, and the memory address <<
1348 must be multiples of this value. Note that <<
1349 size may be greater than the logical block <<
1350 <<
1351 If either of the above conditions is not met, <<
1352 encrypted file will fall back to buffered I/O <<
1353 <<
1354 Implementation details 1143 Implementation details
1355 ====================== 1144 ======================
1356 1145
1357 Encryption context 1146 Encryption context
1358 ------------------ 1147 ------------------
1359 1148
1360 An encryption policy is represented on-disk b 1149 An encryption policy is represented on-disk by
1361 struct fscrypt_context_v1 or struct fscrypt_c 1150 struct fscrypt_context_v1 or struct fscrypt_context_v2. It is up to
1362 individual filesystems to decide where to sto 1151 individual filesystems to decide where to store it, but normally it
1363 would be stored in a hidden extended attribut 1152 would be stored in a hidden extended attribute. It should *not* be
1364 exposed by the xattr-related system calls suc 1153 exposed by the xattr-related system calls such as getxattr() and
1365 setxattr() because of the special semantics o 1154 setxattr() because of the special semantics of the encryption xattr.
1366 (In particular, there would be much confusion 1155 (In particular, there would be much confusion if an encryption policy
1367 were to be added to or removed from anything 1156 were to be added to or removed from anything other than an empty
1368 directory.) These structs are defined as fol 1157 directory.) These structs are defined as follows::
1369 1158
1370 #define FSCRYPT_FILE_NONCE_SIZE 16 1159 #define FSCRYPT_FILE_NONCE_SIZE 16
1371 1160
1372 #define FSCRYPT_KEY_DESCRIPTOR_SIZE 8 1161 #define FSCRYPT_KEY_DESCRIPTOR_SIZE 8
1373 struct fscrypt_context_v1 { 1162 struct fscrypt_context_v1 {
1374 u8 version; 1163 u8 version;
1375 u8 contents_encryption_mode; 1164 u8 contents_encryption_mode;
1376 u8 filenames_encryption_mode; 1165 u8 filenames_encryption_mode;
1377 u8 flags; 1166 u8 flags;
1378 u8 master_key_descriptor[FSCRYPT_ 1167 u8 master_key_descriptor[FSCRYPT_KEY_DESCRIPTOR_SIZE];
1379 u8 nonce[FSCRYPT_FILE_NONCE_SIZE] 1168 u8 nonce[FSCRYPT_FILE_NONCE_SIZE];
1380 }; 1169 };
1381 1170
1382 #define FSCRYPT_KEY_IDENTIFIER_SIZE 16 1171 #define FSCRYPT_KEY_IDENTIFIER_SIZE 16
1383 struct fscrypt_context_v2 { 1172 struct fscrypt_context_v2 {
1384 u8 version; 1173 u8 version;
1385 u8 contents_encryption_mode; 1174 u8 contents_encryption_mode;
1386 u8 filenames_encryption_mode; 1175 u8 filenames_encryption_mode;
1387 u8 flags; 1176 u8 flags;
1388 u8 log2_data_unit_size; !! 1177 u8 __reserved[4];
1389 u8 __reserved[3]; <<
1390 u8 master_key_identifier[FSCRYPT_ 1178 u8 master_key_identifier[FSCRYPT_KEY_IDENTIFIER_SIZE];
1391 u8 nonce[FSCRYPT_FILE_NONCE_SIZE] 1179 u8 nonce[FSCRYPT_FILE_NONCE_SIZE];
1392 }; 1180 };
1393 1181
1394 The context structs contain the same informat 1182 The context structs contain the same information as the corresponding
1395 policy structs (see `Setting an encryption po 1183 policy structs (see `Setting an encryption policy`_), except that the
1396 context structs also contain a nonce. The no 1184 context structs also contain a nonce. The nonce is randomly generated
1397 by the kernel and is used as KDF input or as 1185 by the kernel and is used as KDF input or as a tweak to cause
1398 different files to be encrypted differently; 1186 different files to be encrypted differently; see `Per-file encryption
1399 keys`_ and `DIRECT_KEY policies`_. 1187 keys`_ and `DIRECT_KEY policies`_.
1400 1188
1401 Data path changes 1189 Data path changes
1402 ----------------- 1190 -----------------
1403 1191
1404 When inline encryption is used, filesystems j !! 1192 For the read path (->readpage()) of regular files, filesystems can
1405 encryption contexts with bios to specify how <<
1406 inline encryption hardware will encrypt/decry <<
1407 <<
1408 When inline encryption isn't used, filesystem <<
1409 the file contents themselves, as described be <<
1410 <<
1411 For the read path (->read_folio()) of regular <<
1412 read the ciphertext into the page cache and d 1193 read the ciphertext into the page cache and decrypt it in-place. The
1413 folio lock must be held until decryption has !! 1194 page lock must be held until decryption has finished, to prevent the
1414 folio from becoming visible to userspace prem !! 1195 page from becoming visible to userspace prematurely.
1415 1196
1416 For the write path (->writepage()) of regular 1197 For the write path (->writepage()) of regular files, filesystems
1417 cannot encrypt data in-place in the page cach 1198 cannot encrypt data in-place in the page cache, since the cached
1418 plaintext must be preserved. Instead, filesy 1199 plaintext must be preserved. Instead, filesystems must encrypt into a
1419 temporary buffer or "bounce page", then write 1200 temporary buffer or "bounce page", then write out the temporary
1420 buffer. Some filesystems, such as UBIFS, alr 1201 buffer. Some filesystems, such as UBIFS, already use temporary
1421 buffers regardless of encryption. Other file 1202 buffers regardless of encryption. Other filesystems, such as ext4 and
1422 F2FS, have to allocate bounce pages specially 1203 F2FS, have to allocate bounce pages specially for encryption.
1423 1204
>> 1205 Fscrypt is also able to use inline encryption hardware instead of the
>> 1206 kernel crypto API for en/decryption of file contents. When possible,
>> 1207 and if directed to do so (by specifying the 'inlinecrypt' mount option
>> 1208 for an ext4/F2FS filesystem), it adds encryption contexts to bios and
>> 1209 uses blk-crypto to perform the en/decryption instead of making use of
>> 1210 the above read/write path changes. Of course, even if directed to
>> 1211 make use of inline encryption, fscrypt will only be able to do so if
>> 1212 either hardware inline encryption support is available for the
>> 1213 selected encryption algorithm or CONFIG_BLK_INLINE_ENCRYPTION_FALLBACK
>> 1214 is selected. If neither is the case, fscrypt will fall back to using
>> 1215 the above mentioned read/write path changes for en/decryption.
>> 1216
1424 Filename hashing and encoding 1217 Filename hashing and encoding
1425 ----------------------------- 1218 -----------------------------
1426 1219
1427 Modern filesystems accelerate directory looku 1220 Modern filesystems accelerate directory lookups by using indexed
1428 directories. An indexed directory is organiz 1221 directories. An indexed directory is organized as a tree keyed by
1429 filename hashes. When a ->lookup() is reques 1222 filename hashes. When a ->lookup() is requested, the filesystem
1430 normally hashes the filename being looked up 1223 normally hashes the filename being looked up so that it can quickly
1431 find the corresponding directory entry, if an 1224 find the corresponding directory entry, if any.
1432 1225
1433 With encryption, lookups must be supported an 1226 With encryption, lookups must be supported and efficient both with and
1434 without the encryption key. Clearly, it woul 1227 without the encryption key. Clearly, it would not work to hash the
1435 plaintext filenames, since the plaintext file 1228 plaintext filenames, since the plaintext filenames are unavailable
1436 without the key. (Hashing the plaintext file 1229 without the key. (Hashing the plaintext filenames would also make it
1437 impossible for the filesystem's fsck tool to 1230 impossible for the filesystem's fsck tool to optimize encrypted
1438 directories.) Instead, filesystems hash the 1231 directories.) Instead, filesystems hash the ciphertext filenames,
1439 i.e. the bytes actually stored on-disk in the 1232 i.e. the bytes actually stored on-disk in the directory entries. When
1440 asked to do a ->lookup() with the key, the fi 1233 asked to do a ->lookup() with the key, the filesystem just encrypts
1441 the user-supplied name to get the ciphertext. 1234 the user-supplied name to get the ciphertext.
1442 1235
1443 Lookups without the key are more complicated. 1236 Lookups without the key are more complicated. The raw ciphertext may
1444 contain the ``\0`` and ``/`` characters, whic 1237 contain the ``\0`` and ``/`` characters, which are illegal in
1445 filenames. Therefore, readdir() must base64u !! 1238 filenames. Therefore, readdir() must base64-encode the ciphertext for
1446 for presentation. For most filenames, this w !! 1239 presentation. For most filenames, this works fine; on ->lookup(), the
1447 the filesystem just base64url-decodes the use !! 1240 filesystem just base64-decodes the user-supplied name to get back to
1448 back to the raw ciphertext. !! 1241 the raw ciphertext.
1449 1242
1450 However, for very long filenames, base64url e !! 1243 However, for very long filenames, base64 encoding would cause the
1451 filename length to exceed NAME_MAX. To preve 1244 filename length to exceed NAME_MAX. To prevent this, readdir()
1452 actually presents long filenames in an abbrev 1245 actually presents long filenames in an abbreviated form which encodes
1453 a strong "hash" of the ciphertext filename, a 1246 a strong "hash" of the ciphertext filename, along with the optional
1454 filesystem-specific hash(es) needed for direc 1247 filesystem-specific hash(es) needed for directory lookups. This
1455 allows the filesystem to still, with a high d 1248 allows the filesystem to still, with a high degree of confidence, map
1456 the filename given in ->lookup() back to a pa 1249 the filename given in ->lookup() back to a particular directory entry
1457 that was previously listed by readdir(). See 1250 that was previously listed by readdir(). See
1458 struct fscrypt_nokey_name in the source for m 1251 struct fscrypt_nokey_name in the source for more details.
1459 1252
1460 Note that the precise way that filenames are 1253 Note that the precise way that filenames are presented to userspace
1461 without the key is subject to change in the f 1254 without the key is subject to change in the future. It is only meant
1462 as a way to temporarily present valid filenam 1255 as a way to temporarily present valid filenames so that commands like
1463 ``rm -r`` work as expected on encrypted direc 1256 ``rm -r`` work as expected on encrypted directories.
1464 1257
1465 Tests 1258 Tests
1466 ===== 1259 =====
1467 1260
1468 To test fscrypt, use xfstests, which is Linux 1261 To test fscrypt, use xfstests, which is Linux's de facto standard
1469 filesystem test suite. First, run all the te 1262 filesystem test suite. First, run all the tests in the "encrypt"
1470 group on the relevant filesystem(s). One can 1263 group on the relevant filesystem(s). One can also run the tests
1471 with the 'inlinecrypt' mount option to test t 1264 with the 'inlinecrypt' mount option to test the implementation for
1472 inline encryption support. For example, to t 1265 inline encryption support. For example, to test ext4 and
1473 f2fs encryption using `kvm-xfstests 1266 f2fs encryption using `kvm-xfstests
1474 <https://github.com/tytso/xfstests-bld/blob/m 1267 <https://github.com/tytso/xfstests-bld/blob/master/Documentation/kvm-quickstart.md>`_::
1475 1268
1476 kvm-xfstests -c ext4,f2fs -g encrypt 1269 kvm-xfstests -c ext4,f2fs -g encrypt
1477 kvm-xfstests -c ext4,f2fs -g encrypt -m i 1270 kvm-xfstests -c ext4,f2fs -g encrypt -m inlinecrypt
1478 1271
1479 UBIFS encryption can also be tested this way, 1272 UBIFS encryption can also be tested this way, but it should be done in
1480 a separate command, and it takes some time fo 1273 a separate command, and it takes some time for kvm-xfstests to set up
1481 emulated UBI volumes:: 1274 emulated UBI volumes::
1482 1275
1483 kvm-xfstests -c ubifs -g encrypt 1276 kvm-xfstests -c ubifs -g encrypt
1484 1277
1485 No tests should fail. However, tests that us 1278 No tests should fail. However, tests that use non-default encryption
1486 modes (e.g. generic/549 and generic/550) will 1279 modes (e.g. generic/549 and generic/550) will be skipped if the needed
1487 algorithms were not built into the kernel's c 1280 algorithms were not built into the kernel's crypto API. Also, tests
1488 that access the raw block device (e.g. generi 1281 that access the raw block device (e.g. generic/399, generic/548,
1489 generic/549, generic/550) will be skipped on 1282 generic/549, generic/550) will be skipped on UBIFS.
1490 1283
1491 Besides running the "encrypt" group tests, fo 1284 Besides running the "encrypt" group tests, for ext4 and f2fs it's also
1492 possible to run most xfstests with the "test_ 1285 possible to run most xfstests with the "test_dummy_encryption" mount
1493 option. This option causes all new files to 1286 option. This option causes all new files to be automatically
1494 encrypted with a dummy key, without having to 1287 encrypted with a dummy key, without having to make any API calls.
1495 This tests the encrypted I/O paths more thoro 1288 This tests the encrypted I/O paths more thoroughly. To do this with
1496 kvm-xfstests, use the "encrypt" filesystem co 1289 kvm-xfstests, use the "encrypt" filesystem configuration::
1497 1290
1498 kvm-xfstests -c ext4/encrypt,f2fs/encrypt 1291 kvm-xfstests -c ext4/encrypt,f2fs/encrypt -g auto
1499 kvm-xfstests -c ext4/encrypt,f2fs/encrypt 1292 kvm-xfstests -c ext4/encrypt,f2fs/encrypt -g auto -m inlinecrypt
1500 1293
1501 Because this runs many more tests than "-g en 1294 Because this runs many more tests than "-g encrypt" does, it takes
1502 much longer to run; so also consider using `g 1295 much longer to run; so also consider using `gce-xfstests
1503 <https://github.com/tytso/xfstests-bld/blob/m 1296 <https://github.com/tytso/xfstests-bld/blob/master/Documentation/gce-xfstests.md>`_
1504 instead of kvm-xfstests:: 1297 instead of kvm-xfstests::
1505 1298
1506 gce-xfstests -c ext4/encrypt,f2fs/encrypt 1299 gce-xfstests -c ext4/encrypt,f2fs/encrypt -g auto
1507 gce-xfstests -c ext4/encrypt,f2fs/encrypt 1300 gce-xfstests -c ext4/encrypt,f2fs/encrypt -g auto -m inlinecrypt
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.