Lines Matching +full:master +full:- +full:kernel
2 Filesystem-level encryption (fscrypt)
11 Note: "fscrypt" in this document refers to the kernel-level portion,
14 covers the kernel-level portion. For command-line examples of how to
20 <https://source.android.com/security/encryption/file-based>`_, over
21 using the kernel's API directly. Using existing tools reduces the
23 completeness this documentation covers the kernel's API anyway.)
25 Unlike dm-crypt, fscrypt operates at the filesystem level rather than
28 filesystem. This is useful for multi-user systems where each user's
29 data-at-rest needs to be cryptographically isolated from the others.
34 directly into supported filesystems --- currently ext4, F2FS, and
44 fscrypt does not support encrypting files in-place. Instead, it
54 ---------------
58 event of a single point-in-time permanent offline compromise of the
60 non-filename metadata, e.g. file sizes, file permissions, file
70 --------------
75 Side-channel attacks
78 fscrypt is only resistant to side-channel attacks, such as timing or
81 vulnerable algorithm is used, such as a table-based implementation of
97 encryption but rather only by the correctness of the kernel.
98 Therefore, any encryption-specific access control checks would merely
99 be enforced by kernel *code* and therefore would be largely redundant
102 Kernel memory compromise
106 memory, e.g. by mounting a physical attack or by exploiting a kernel
110 However, fscrypt allows encryption keys to be removed from the kernel,
114 FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS ioctl) can wipe a master
115 encryption key from kernel memory. If it does so, it will also try to
117 thereby wiping their per-file keys and making them once again appear
122 - Per-file keys for in-use files will *not* be removed or wiped.
124 encrypted files and directories before removing a master key, as
128 - The kernel cannot magically wipe copies of the master key(s) that
130 copies of the master key(s) it makes as well; normally this should
137 - In general, decrypted contents and filenames in the kernel VFS
141 CONFIG_PAGE_POISONING=y in your kernel config and add page_poison=1
142 to your kernel command line. However, this has a performance cost.
144 - Secret keys might still exist in CPU registers, in crypto
154 - There is no verification that the provided master key is correct.
156 with another user's encrypted files to which they have read-only
160 meaning of "read-only access".
162 - A compromise of a per-file key also compromises the master key from
165 - Non-root users cannot securely remove encryption keys.
174 Master Keys
175 -----------
177 Each encrypted directory tree is protected by a *master key*. Master
180 encryption modes being used. For example, if any AES-256 mode is
181 used, the master key must be at least 256 bits, i.e. 32 bytes. A
183 policy and AES-256-XTS is used; such keys must be 64 bytes.
186 appropriate master key. There can be any number of master keys, each
190 Master keys must be real cryptographic keys, i.e. indistinguishable
192 **must not** directly use a password as a master key, zero-pad a
197 Instead, users should generate master keys either using a
199 (Key Derivation Function). The kernel does not do any key stretching;
200 therefore, if userspace derives the key from a low-entropy secret such
205 -----------------------
207 With one exception, fscrypt never uses the master key(s) for
211 The KDF used for a particular master key differs depending on whether
214 encryption policies. (No real-world attack is currently known on this
218 For v1 encryption policies, the KDF only supports deriving per-file
219 encryption keys. It works by encrypting the master key with
220 AES-128-ECB, using the file's 16-byte nonce as the AES key. The
224 For v2 encryption policies, the KDF is HKDF-SHA512. The master key is
226 "application-specific information string" is used for each distinct
227 key to be derived. For example, when a per-file encryption key is
228 derived, the application-specific information string is the file's
232 HKDF-SHA512 is preferred to the original AES-128-ECB based KDF because
234 entropy from the master key. HKDF is also standardized and widely
235 used by other software, whereas the AES-128-ECB based KDF is ad-hoc.
237 Per-file encryption keys
238 ------------------------
240 Since each master key can protect many files, it is necessary to
243 cases, fscrypt does this by deriving per-file keys. When a new
245 fscrypt randomly generates a 16-byte nonce and stores it in the
247 derivation function`_) to derive the file's key from the master key
251 require larger xattrs which would be less likely to fit in-line in the
255 alternative master keys or to support rotating master keys. Instead,
256 the master keys may be wrapped in userspace, e.g. as is done by the
260 -------------------
264 long IVs --- long enough to hold both an 8-byte logical block number
265 and a 16-byte per-file nonce. Also, the overhead of each Adiantum key
266 is greater than that of an AES-256-XTS key.
271 per-file encryption keys are not used. Instead, whenever any data
272 (contents or filenames) is encrypted, the file's 16-byte nonce is
275 - For v1 encryption policies, the encryption is done directly with the
276 master key. Because of this, users **must not** use the same master
279 - For v2 encryption policies, the encryption is done with a per-mode
280 key derived using the KDF. Users may use the same master key for
284 -----------------------
287 the encryption keys are derived from the master key, encryption mode
289 protected by the same master key sharing a single contents encryption
299 -----------------------
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) and added to the file
304 logical block number mod 2^32 to produce a 32-bit IV.
313 ---------------
315 For master keys used for v2 encryption policies, a unique 16-byte "key
320 ------------
322 For directories that are indexed using a secret-keyed dirhash over the
323 plaintext filenames, the KDF is also used to derive a 128-bit
324 SipHash-2-4 key per directory in order to hash filenames. This works
325 just like deriving a per-file encryption key, except that a different
326 KDF context is used. Currently, only casefolded ("case-insensitive")
337 - AES-256-XTS for contents and AES-256-CTS-CBC for filenames
338 - AES-128-CBC for contents and AES-128-CTS-CBC for filenames
339 - Adiantum for both contents and filenames
340 - AES-256-XTS for contents and AES-256-HCTR2 for filenames (v2 policies only)
342 If unsure, you should use the (AES-256-XTS, AES-256-CTS-CBC) pair.
344 AES-128-CBC was added only for low-powered embedded devices with
346 use AES-128-CBC, CONFIG_CRYPTO_ESSIV and CONFIG_CRYPTO_SHA256 (or
347 another SHA-256 implementation) must be enabled so that ESSIV can be
350 Adiantum is a (primarily) stream cipher-based mode that is fast even
352 wide-block mode, unlike XTS. It can also eliminate the need to derive
353 per-file encryption keys. However, it depends on the security of two
354 primitives, XChaCha12 and AES-256, rather than just one. See the
355 paper "Adiantum: length-preserving encryption for entry-level
361 AES-256-HCTR2 is another true wide-block encryption mode that is intended for
362 use on CPUs with dedicated crypto instructions. AES-256-HCTR2 has the property
365 reused within a directory. For more details on AES-256-HCTR2, see the paper
366 "Length-preserving encryption with HCTR2"
367 (https://eprint.iacr.org/2021/1441.pdf). To use AES-256-HCTR2,
378 -------------------
381 Starting from Linux kernel 5.5, encryption of filesystems with block
387 - With CBC mode encryption, ESSIV is also used. Specifically, each IV
388 is encrypted with AES-256 where the AES-256 key is the SHA-256 hash
391 - With `DIRECT_KEY policies`_, the file's nonce is appended to the IV.
394 - With `IV_INO_LBLK_64 policies`_, the logical block number is limited
395 to 32 bits and is placed in bits 0-31 of the IV. The inode number
396 (which is also limited to 32 bits) is placed in bits 32-63.
398 - With `IV_INO_LBLK_32 policies`_, the logical block number is limited
399 to 32 bits and is placed in bits 0-31 of the IV. The inode number
407 --------------------
419 With CTS-CBC, the IV reuse means that when the plaintext filenames share a
423 wide-block encryption modes.
427 filenames shorter than 16 bytes are NUL-padded to 16 bytes before
429 via their ciphertexts, all filenames are NUL-padded to the next 4, 8,
430 16, or 32-byte boundary (configurable). 32 is recommended since this
444 ----------------------------
479 - ``version`` must be FSCRYPT_POLICY_V1 (0) if
485 - ``contents_encryption_mode`` and ``filenames_encryption_mode`` must
491 - ``flags`` contains optional flags from ``<linux/fscrypt.h>``:
493 - FSCRYPT_POLICY_FLAGS_PAD_*: The amount of NUL padding to use when
496 - FSCRYPT_POLICY_FLAG_DIRECT_KEY: See `DIRECT_KEY policies`_.
497 - FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64: See `IV_INO_LBLK_64
499 - FSCRYPT_POLICY_FLAG_IV_INO_LBLK_32: See `IV_INO_LBLK_32
508 - For v2 encryption policies, ``__reserved`` must be zeroed.
510 - For v1 encryption policies, ``master_key_descriptor`` specifies how
511 to find the master key in a keyring; see `Adding keys`_. It is up
513 master key. The e4crypt and fscrypt tools use the first 8 bytes of
514 ``SHA-512(SHA-512(master_key))``, but this particular scheme is not
515 required. Also, the master key need not be in the keyring yet when
523 the kernel returned in the struct fscrypt_add_key_arg must
531 corresponding master key as described in `Adding keys`_, all regular
553 filesystem with one key should consider using dm-crypt instead.
557 - ``EACCES``: the file is not owned by the process's uid, nor does the
560 - ``EEXIST``: the file is already encrypted with an encryption policy
562 - ``EINVAL``: an invalid encryption policy was specified (invalid
566 - ``ENOKEY``: a v2 encryption policy was specified, but the key with
570 - ``ENOTDIR``: the file is unencrypted and is a regular file, not a
572 - ``ENOTEMPTY``: the file is unencrypted and is a nonempty directory
573 - ``ENOTTY``: this type of filesystem does not implement encryption
574 - ``EOPNOTSUPP``: the kernel was not configured with encryption
578 kernel config, and the superblock must have had the "encrypt"
579 feature flag enabled using ``tune2fs -O encrypt`` or ``mkfs.ext4 -O
581 - ``EPERM``: this directory may not be encrypted, e.g. because it is
583 - ``EROFS``: the filesystem is readonly
586 ----------------------------
590 - `FS_IOC_GET_ENCRYPTION_POLICY_EX`_
591 - `FS_IOC_GET_ENCRYPTION_POLICY`_
627 - ``EINVAL``: the file is encrypted, but it uses an unrecognized
629 - ``ENODATA``: the file is not encrypted
630 - ``ENOTTY``: this type of filesystem does not implement encryption,
631 or this kernel is too old to support FS_IOC_GET_ENCRYPTION_POLICY_EX
633 - ``EOPNOTSUPP``: the kernel was not configured with encryption
636 - ``EOVERFLOW``: the file is encrypted and uses a recognized
660 Getting the per-filesystem salt
661 -------------------------------
665 generated 16-byte value stored in the filesystem superblock. This
667 from a passphrase or other low-entropy user credential.
673 ---------------------------------
676 On encrypted files and directories it gets the inode's 16-byte nonce.
684 -----------
689 The FS_IOC_ADD_ENCRYPTION_KEY ioctl adds a master encryption key to
726 - If the key is being added for use by v1 encryption policies, then
737 an *output* field which the kernel fills in with a cryptographic
743 - ``raw_size`` must be the size of the ``raw`` key provided, in bytes.
747 - ``key_id`` is 0 if the raw key is given directly in the ``raw``
749 type "fscrypt-provisioning" whose payload is
752 Since ``raw`` is variable-length, the total size of this key's
759 allow re-adding keys after a filesystem is unmounted and re-mounted,
762 - ``raw`` is a variable-length field which must contain the actual
766 For v2 policy keys, the kernel keeps track of which user (identified
768 removed by that user --- or by "root", if they use
784 - ``EACCES``: FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR was specified, but the
788 - ``EDQUOT``: the key quota for this user would be exceeded by adding
790 - ``EINVAL``: invalid key size or key specifier type, or reserved bits
792 - ``EKEYREJECTED``: the raw key was specified by Linux key ID, but the
794 - ``ENOKEY``: the raw key was specified by Linux key ID, but no key
796 - ``ENOTTY``: this type of filesystem does not implement encryption
797 - ``EOPNOTSUPP``: the kernel was not configured with encryption
804 For v1 encryption policies, a master encryption key can also be
805 provided by adding it to a process-subscribed keyring, e.g. to a
821 Nevertheless, to add a key to one of the process-subscribed keyrings,
824 "logon"; keys of this type are kept in kernel memory and cannot be
826 followed by the 16-character lower case hex representation of the
840 bytes ``raw[0..size-1]`` (inclusive) are the actual key.
843 with a filesystem-specific prefix such as "ext4:". However, the
844 filesystem-specific prefixes are deprecated and should not be used in
848 -------------
853 - `FS_IOC_REMOVE_ENCRYPTION_KEY`_
854 - `FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS`_
857 or removed by non-root users.
860 process-subscribed keyrings mechanism.
862 Before using these ioctls, read the `Kernel memory compromise`_
869 The FS_IOC_REMOVE_ENCRYPTION_KEY ioctl removes a claim to a master
886 - The key to remove is specified by ``key_spec``:
888 - To remove a key used by v1 encryption policies, set
894 - To remove a key used by v2 encryption policies, set
898 For v2 policy keys, this ioctl is usable by non-root users. However,
913 lock files that are still in-use, so this ioctl is expected to be used
925 - ``FSCRYPT_KEY_REMOVAL_STATUS_FLAG_FILES_BUSY``: set if some file(s)
926 are still in-use. Not guaranteed to be set in the case where only
928 - ``FSCRYPT_KEY_REMOVAL_STATUS_FLAG_OTHER_USERS``: set if only the
933 - ``EACCES``: The FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR key specifier type
936 - ``EINVAL``: invalid key specifier type, or reserved bits were set
937 - ``ENOKEY``: the key object was not found at all, i.e. it was never
941 - ``ENOTTY``: this type of filesystem does not implement encryption
942 - ``EOPNOTSUPP``: the kernel was not configured with encryption
954 only meaningful if non-root users are adding and removing keys.
961 ------------------
967 master encryption key. It can be executed on any file or directory on
990 - To get the status of a key for v1 encryption policies, set
994 - To get the status of a key for v2 encryption policies, set
998 On success, 0 is returned and the kernel fills in the output fields:
1000 - ``status`` indicates whether the key is absent, present, or
1001 incompletely removed. Incompletely removed means that the master
1006 - ``status_flags`` can contain the following flags:
1008 - ``FSCRYPT_KEY_STATUS_FLAG_ADDED_BY_SELF`` indicates that the key
1012 - ``user_count`` specifies the number of users who have added the key.
1018 - ``EINVAL``: invalid key specifier type, or reserved bits were set
1019 - ``ENOTTY``: this type of filesystem does not implement encryption
1020 - ``EOPNOTSUPP``: the kernel was not configured with encryption
1030 the filesystem-level keyring, i.e. the keyring managed by
1034 process-subscribed keyrings.
1040 ------------
1043 symlinks behave very similarly to their unencrypted counterparts ---
1047 - Unencrypted files, or files encrypted with a different encryption
1062 - Direct I/O is supported on encrypted files only under some
1065 - The fallocate operations FALLOC_FL_COLLAPSE_RANGE and
1069 - Online defragmentation of encrypted files is not supported. The
1073 - The ext4 filesystem does not support data journaling with encrypted
1076 - DAX (Direct Access) is not supported on encrypted files.
1078 - The maximum length of an encrypted symlink is 2 bytes shorter than
1088 ---------------
1094 - File metadata may be read, e.g. using stat().
1096 - Directories may be listed, in which case the filenames will be
1107 - Files may be deleted. That is, nondirectory files may be deleted
1109 rmdir() as usual. Therefore, ``rm`` and ``rm -r`` will work as
1112 - Symlink targets may be read and followed, but they will be presented
1136 (recursively) will inherit that encryption policy. Special files ---
1137 that is, named pipes, device nodes, and UNIX domain sockets --- will
1144 during ->lookup() to provide limited protection against offline
1148 this by validating all top-level encryption policies prior to access.
1153 By default, fscrypt uses the kernel crypto API for all cryptographic
1155 itself). The kernel crypto API supports hardware crypto accelerators,
1165 through a set of extensions to the block layer called *blk-crypto*.
1166 blk-crypto allows filesystems to attach encryption contexts to bios
1168 in-line. For more information about blk-crypto, see
1169 :ref:`Documentation/block/inline-encryption.rst <inline_encryption>`.
1172 blk-crypto instead of the kernel crypto API to encrypt/decrypt file
1174 the kernel configuration, and specify the "inlinecrypt" mount option
1179 still fall back to using the kernel crypto API on files where the
1182 and where blk-crypto-fallback is unusable. (For blk-crypto-fallback
1183 to be usable, it must be enabled in the kernel configuration with
1191 the on-disk format, so users may freely switch back and forth between
1202 the filesystem must be mounted with ``-o inlinecrypt`` and inline
1219 ------------------
1221 An encryption policy is represented on-disk by
1225 exposed by the xattr-related system calls such as getxattr() and
1257 by the kernel and is used as KDF input or as a tweak to cause
1258 different files to be encrypted differently; see `Per-file encryption
1262 -----------------
1271 For the read path (->read_folio()) of regular files, filesystems can
1272 read the ciphertext into the page cache and decrypt it in-place. The
1276 For the write path (->writepage()) of regular files, filesystems
1277 cannot encrypt data in-place in the page cache, since the cached
1285 -----------------------------
1289 filename hashes. When a ->lookup() is requested, the filesystem
1299 i.e. the bytes actually stored on-disk in the directory entries. When
1300 asked to do a ->lookup() with the key, the filesystem just encrypts
1301 the user-supplied name to get the ciphertext.
1305 filenames. Therefore, readdir() must base64url-encode the ciphertext
1306 for presentation. For most filenames, this works fine; on ->lookup(),
1307 the filesystem just base64url-decodes the user-supplied name to get
1314 filesystem-specific hash(es) needed for directory lookups. This
1316 the filename given in ->lookup() back to a particular directory entry
1323 ``rm -r`` work as expected on encrypted directories.
1333 f2fs encryption using `kvm-xfstests
1334 <https://github.com/tytso/xfstests-bld/blob/master/Documentation/kvm-quickstart.md>`_::
1336 kvm-xfstests -c ext4,f2fs -g encrypt
1337 kvm-xfstests -c ext4,f2fs -g encrypt -m inlinecrypt
1340 a separate command, and it takes some time for kvm-xfstests to set up
1343 kvm-xfstests -c ubifs -g encrypt
1345 No tests should fail. However, tests that use non-default encryption
1347 algorithms were not built into the kernel's crypto API. Also, tests
1356 kvm-xfstests, use the "encrypt" filesystem configuration::
1358 kvm-xfstests -c ext4/encrypt,f2fs/encrypt -g auto
1359 kvm-xfstests -c ext4/encrypt,f2fs/encrypt -g auto -m inlinecrypt
1361 Because this runs many more tests than "-g encrypt" does, it takes
1362 much longer to run; so also consider using `gce-xfstests
1363 <https://github.com/tytso/xfstests-bld/blob/master/Documentation/gce-xfstests.md>`_
1364 instead of kvm-xfstests::
1366 gce-xfstests -c ext4/encrypt,f2fs/encrypt -g auto
1367 gce-xfstests -c ext4/encrypt,f2fs/encrypt -g auto -m inlinecrypt