Lines Matching +full:kernel +full:- +full:policy
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,
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.
128 - The kernel cannot magically wipe copies of the master key(s) that
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.
175 -----------
180 encryption modes being used. For example, if any AES-256 mode is
183 policy and AES-256-XTS is used; such keys must be 64 bytes.
192 **must not** directly use a password as a master key, zero-pad 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 -----------------------
214 encryption policies. (No real-world attack is currently known on this
218 For v1 encryption policies, the KDF only supports deriving per-file
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
235 used by other software, whereas the AES-128-ECB based KDF is ad-hoc.
237 Per-file encryption keys
238 ------------------------
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
251 require larger xattrs which would be less likely to fit in-line in 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.
270 this by setting FSCRYPT_POLICY_FLAG_DIRECT_KEY in the fscrypt policy,
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
279 - For v2 encryption policies, the encryption is done with a per-mode
284 -----------------------
286 When FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64 is set in the fscrypt policy,
299 -----------------------
302 IV_INO_LBLK_32, the inode number is hashed with SipHash-2-4 (where the
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 ---------------
341 - AES-256-XTS for contents and AES-256-CTS-CBC for filenames
342 - AES-256-XTS for contents and AES-256-HCTR2 for filenames
343 - Adiantum for both contents and filenames
344 - AES-128-CBC-ESSIV for contents and AES-128-CTS-CBC for filenames
345 - SM4-XTS for contents and SM4-CTS-CBC for filenames
351 `CBC-ESSIV mode
352 <https://en.wikipedia.org/wiki/Disk_encryption_theory#Encrypted_salt-sector_initialization_vector_(…
353 or a wide-block cipher. Filenames encryption uses a
354 block cipher in `CTS-CBC mode
355 <https://en.wikipedia.org/wiki/Ciphertext_stealing>`_ or a wide-block
358 The (AES-256-XTS, AES-256-CTS-CBC) pair is the recommended default.
360 if the kernel supports fscrypt at all; see `Kernel config options`_.
362 The (AES-256-XTS, AES-256-HCTR2) pair is also a good choice that
363 upgrades the filenames encryption to use a wide-block cipher. (A
364 *wide-block cipher*, also called a tweakable super-pseudorandom
366 entire result.) As described in `Filenames encryption`_, a wide-block
367 cipher is the ideal mode for the problem domain, though CTS-CBC is the
372 of hardware acceleration for AES. Adiantum is a wide-block cipher
373 that uses XChaCha12 and AES-256 as its underlying components. Most of
378 The (AES-128-CBC-ESSIV, AES-128-CTS-CBC) pair exists only to support
379 systems whose only form of AES acceleration is an off-CPU crypto
384 - (SM4-XTS, SM4-CTS-CBC)
391 Kernel config options
392 ---------------------
395 only the basic support from the crypto API needed to use AES-256-XTS
396 and AES-256-CTS-CBC encryption. For optimal performance, it is
397 strongly recommended to also enable any available platform-specific
399 wish to use. Support for any "non-default" encryption modes typically
406 kernel crypto API (see `Inline encryption support`_); in that case,
407 the file contents mode doesn't need to supported in the kernel crypto
410 - AES-256-XTS and AES-256-CTS-CBC
411 - Recommended:
412 - arm64: CONFIG_CRYPTO_AES_ARM64_CE_BLK
413 - x86: CONFIG_CRYPTO_AES_NI_INTEL
415 - AES-256-HCTR2
416 - Mandatory:
417 - CONFIG_CRYPTO_HCTR2
418 - Recommended:
419 - arm64: CONFIG_CRYPTO_AES_ARM64_CE_BLK
420 - arm64: CONFIG_CRYPTO_POLYVAL_ARM64_CE
421 - x86: CONFIG_CRYPTO_AES_NI_INTEL
422 - x86: CONFIG_CRYPTO_POLYVAL_CLMUL_NI
424 - Adiantum
425 - Mandatory:
426 - CONFIG_CRYPTO_ADIANTUM
427 - Recommended:
428 - arm32: CONFIG_CRYPTO_CHACHA20_NEON
429 - arm32: CONFIG_CRYPTO_NHPOLY1305_NEON
430 - arm64: CONFIG_CRYPTO_CHACHA20_NEON
431 - arm64: CONFIG_CRYPTO_NHPOLY1305_NEON
432 - x86: CONFIG_CRYPTO_CHACHA20_X86_64
433 - x86: CONFIG_CRYPTO_NHPOLY1305_SSE2
434 - x86: CONFIG_CRYPTO_NHPOLY1305_AVX2
436 - AES-128-CBC-ESSIV and AES-128-CTS-CBC:
437 - Mandatory:
438 - CONFIG_CRYPTO_ESSIV
439 - CONFIG_CRYPTO_SHA256 or another SHA-256 implementation
440 - Recommended:
441 - AES-CBC acceleration
443 fscrypt also uses HMAC-SHA512 for key derivation, so enabling SHA-512
446 - SHA-512
447 - Recommended:
448 - arm64: CONFIG_CRYPTO_SHA512_ARM64_CE
449 - x86: CONFIG_CRYPTO_SHA512_SSSE3
452 -------------------
455 Starting from Linux kernel 5.5, encryption of filesystems with block
461 - With CBC mode encryption, ESSIV is also used. Specifically, each IV
462 is encrypted with AES-256 where the AES-256 key is the SHA-256 hash
465 - With `DIRECT_KEY policies`_, the file's nonce is appended to the IV.
468 - With `IV_INO_LBLK_64 policies`_, the logical block number is limited
469 to 32 bits and is placed in bits 0-31 of the IV. The inode number
470 (which is also limited to 32 bits) is placed in bits 32-63.
472 - With `IV_INO_LBLK_32 policies`_, the logical block number is limited
473 to 32 bits and is placed in bits 0-31 of the IV. The inode number
481 --------------------
493 With CTS-CBC, the IV reuse means that when the plaintext filenames share a
497 wide-block encryption modes.
501 filenames shorter than 16 bytes are NUL-padded to 16 bytes before
503 via their ciphertexts, all filenames are NUL-padded to the next 4, 8,
504 16, or 32-byte boundary (configurable). 32 is recommended since this
517 Setting an encryption policy
518 ----------------------------
523 The FS_IOC_SET_ENCRYPTION_POLICY ioctl sets an encryption policy on an
525 has the specified encryption policy. It takes in a pointer to
553 - ``version`` must be FSCRYPT_POLICY_V1 (0) if
556 policy version as "v1", though its version code is really 0.)
559 - ``contents_encryption_mode`` and ``filenames_encryption_mode`` must
572 - ``flags`` contains optional flags from ``<linux/fscrypt.h>``:
574 - FSCRYPT_POLICY_FLAGS_PAD_*: The amount of NUL padding to use when
577 - FSCRYPT_POLICY_FLAG_DIRECT_KEY: See `DIRECT_KEY policies`_.
578 - FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64: See `IV_INO_LBLK_64
580 - FSCRYPT_POLICY_FLAG_IV_INO_LBLK_32: See `IV_INO_LBLK_32
589 - For v2 encryption policies, ``__reserved`` must be zeroed.
591 - For v1 encryption policies, ``master_key_descriptor`` specifies how
595 ``SHA-512(SHA-512(master_key))``, but this particular scheme is not
604 the kernel returned in the struct fscrypt_add_key_arg must
610 encryption policy is assigned to the directory, turning it into an
614 directory will be encrypted, inheriting the same encryption policy.
619 policy exactly matches the actual one. If they match, then the ioctl
623 When a v2 encryption policy is assigned to a directory, it is also
634 filesystem with one key should consider using dm-crypt instead.
638 - ``EACCES``: the file is not owned by the process's uid, nor does the
641 - ``EEXIST``: the file is already encrypted with an encryption policy
643 - ``EINVAL``: an invalid encryption policy was specified (invalid
645 encryption policy was specified but the directory has the casefold
647 - ``ENOKEY``: a v2 encryption policy was specified, but the key with
651 - ``ENOTDIR``: the file is unencrypted and is a regular file, not a
653 - ``ENOTEMPTY``: the file is unencrypted and is a nonempty directory
654 - ``ENOTTY``: this type of filesystem does not implement encryption
655 - ``EOPNOTSUPP``: the kernel was not configured with encryption
659 kernel config, and the superblock must have had the "encrypt"
660 feature flag enabled using ``tune2fs -O encrypt`` or ``mkfs.ext4 -O
662 - ``EPERM``: this directory may not be encrypted, e.g. because it is
664 - ``EROFS``: the filesystem is readonly
666 Getting an encryption policy
667 ----------------------------
669 Two ioctls are available to get a file's encryption policy:
671 - `FS_IOC_GET_ENCRYPTION_POLICY_EX`_
672 - `FS_IOC_GET_ENCRYPTION_POLICY`_
684 policy, if any, for a directory or regular file. No additional
695 } policy; /* output */
699 the policy struct, i.e. ``sizeof(arg.policy)``.
701 On success, the policy struct is returned in ``policy``, and its
702 actual size is returned in ``policy_size``. ``policy.version`` should
703 be checked to determine the version of policy returned. Note that the
704 version code for the "v1" policy is actually 0 (FSCRYPT_POLICY_V1).
708 - ``EINVAL``: the file is encrypted, but it uses an unrecognized
709 encryption policy version
710 - ``ENODATA``: the file is not encrypted
711 - ``ENOTTY``: this type of filesystem does not implement encryption,
712 or this kernel is too old to support FS_IOC_GET_ENCRYPTION_POLICY_EX
714 - ``EOPNOTSUPP``: the kernel was not configured with encryption
717 - ``EOVERFLOW``: the file is encrypted and uses a recognized
718 encryption policy version, but the policy struct does not fit into
730 encryption policy, if any, for a directory or regular file. However,
732 FS_IOC_GET_ENCRYPTION_POLICY only supports the original policy
739 encrypted using a newer encryption policy version.
741 Getting the per-filesystem salt
742 -------------------------------
746 generated 16-byte value stored in the filesystem superblock. This
748 from a passphrase or other low-entropy user credential.
754 ---------------------------------
757 On encrypted files and directories it gets the inode's 16-byte nonce.
765 -----------
807 - If the key is being added for use by v1 encryption policies, then
818 an *output* field which the kernel fills in with a cryptographic
824 - ``raw_size`` must be the size of the ``raw`` key provided, in bytes.
828 - ``key_id`` is 0 if the raw key is given directly in the ``raw``
830 type "fscrypt-provisioning" whose payload is
833 Since ``raw`` is variable-length, the total size of this key's
840 allow re-adding keys after a filesystem is unmounted and re-mounted,
843 - ``raw`` is a variable-length field which must contain the actual
847 For v2 policy keys, the kernel keeps track of which user (identified
849 removed by that user --- or by "root", if they use
854 FS_IOC_ADD_ENCRYPTION_KEY may also be used to add a v2 policy key
865 - ``EACCES``: FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR was specified, but the
869 - ``EDQUOT``: the key quota for this user would be exceeded by adding
871 - ``EINVAL``: invalid key size or key specifier type, or reserved bits
873 - ``EKEYREJECTED``: the raw key was specified by Linux key ID, but the
875 - ``ENOKEY``: the raw key was specified by Linux key ID, but no key
877 - ``ENOTTY``: this type of filesystem does not implement encryption
878 - ``EOPNOTSUPP``: the kernel was not configured with encryption
886 provided by adding it to a process-subscribed keyring, e.g. to a
902 Nevertheless, to add a key to one of the process-subscribed keyrings,
905 "logon"; keys of this type are kept in kernel memory and cannot be
907 followed by the 16-character lower case hex representation of the
908 ``master_key_descriptor`` that was set in the encryption policy. The
921 bytes ``raw[0..size-1]`` (inclusive) are the actual key.
924 with a filesystem-specific prefix such as "ext4:". However, the
925 filesystem-specific prefixes are deprecated and should not be used in
929 -------------
934 - `FS_IOC_REMOVE_ENCRYPTION_KEY`_
935 - `FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS`_
937 These two ioctls differ only in cases where v2 policy keys are added
938 or removed by non-root users.
941 process-subscribed keyrings mechanism.
943 Before using these ioctls, read the `Kernel memory compromise`_
967 - The key to remove is specified by ``key_spec``:
969 - To remove a key used by v1 encryption policies, set
975 - To remove a key used by v2 encryption policies, set
979 For v2 policy keys, this ioctl is usable by non-root users. However,
994 lock files that are still in-use, so this ioctl is expected to be used
1006 - ``FSCRYPT_KEY_REMOVAL_STATUS_FLAG_FILES_BUSY``: set if some file(s)
1007 are still in-use. Not guaranteed to be set in the case where only
1009 - ``FSCRYPT_KEY_REMOVAL_STATUS_FLAG_OTHER_USERS``: set if only the
1014 - ``EACCES``: The FSCRYPT_KEY_SPEC_TYPE_DESCRIPTOR key specifier type
1017 - ``EINVAL``: invalid key specifier type, or reserved bits were set
1018 - ``ENOKEY``: the key object was not found at all, i.e. it was never
1022 - ``ENOTTY``: this type of filesystem does not implement encryption
1023 - ``EOPNOTSUPP``: the kernel was not configured with encryption
1031 `FS_IOC_REMOVE_ENCRYPTION_KEY`_, except that for v2 policy keys, the
1035 only meaningful if non-root users are adding and removing keys.
1042 ------------------
1071 - To get the status of a key for v1 encryption policies, set
1075 - To get the status of a key for v2 encryption policies, set
1079 On success, 0 is returned and the kernel fills in the output fields:
1081 - ``status`` indicates whether the key is absent, present, or
1087 - ``status_flags`` can contain the following flags:
1089 - ``FSCRYPT_KEY_STATUS_FLAG_ADDED_BY_SELF`` indicates that the key
1093 - ``user_count`` specifies the number of users who have added the key.
1099 - ``EINVAL``: invalid key specifier type, or reserved bits were set
1100 - ``ENOTTY``: this type of filesystem does not implement encryption
1101 - ``EOPNOTSUPP``: the kernel was not configured with encryption
1111 the filesystem-level keyring, i.e. the keyring managed by
1115 process-subscribed keyrings.
1121 ------------
1124 symlinks behave very similarly to their unencrypted counterparts ---
1128 - Unencrypted files, or files encrypted with a different encryption
1129 policy (i.e. different key, modes, or flags), cannot be renamed or
1130 linked into an encrypted directory; see `Encryption policy
1143 - Direct I/O is supported on encrypted files only under some
1146 - The fallocate operations FALLOC_FL_COLLAPSE_RANGE and
1150 - Online defragmentation of encrypted files is not supported. The
1154 - The ext4 filesystem does not support data journaling with encrypted
1157 - DAX (Direct Access) is not supported on encrypted files.
1159 - The maximum length of an encrypted symlink is 2 bytes shorter than
1169 ---------------
1175 - File metadata may be read, e.g. using stat().
1177 - Directories may be listed, in which case the filenames will be
1188 - Files may be deleted. That is, nondirectory files may be deleted
1190 rmdir() as usual. Therefore, ``rm`` and ``rm -r`` will work as
1193 - Symlink targets may be read and followed, but they will be presented
1212 Encryption policy enforcement
1215 After an encryption policy has been set on a directory, all regular
1217 (recursively) will inherit that encryption policy. Special files ---
1218 that is, named pipes, device nodes, and UNIX domain sockets --- will
1222 files, or files encrypted with a different encryption policy, in an
1225 during ->lookup() to provide limited protection against offline
1229 this by validating all top-level encryption policies prior to access.
1234 By default, fscrypt uses the kernel crypto API for all cryptographic
1236 itself). The kernel crypto API supports hardware crypto accelerators,
1246 through a set of extensions to the block layer called *blk-crypto*.
1247 blk-crypto allows filesystems to attach encryption contexts to bios
1249 in-line. For more information about blk-crypto, see
1250 :ref:`Documentation/block/inline-encryption.rst <inline_encryption>`.
1253 blk-crypto instead of the kernel crypto API to encrypt/decrypt file
1255 the kernel configuration, and specify the "inlinecrypt" mount option
1260 still fall back to using the kernel crypto API on files where the
1263 and where blk-crypto-fallback is unusable. (For blk-crypto-fallback
1264 to be usable, it must be enabled in the kernel configuration with
1272 the on-disk format, so users may freely switch back and forth between
1283 the filesystem must be mounted with ``-o inlinecrypt`` and inline
1300 ------------------
1302 An encryption policy is represented on-disk by
1306 exposed by the xattr-related system calls such as getxattr() and
1308 (In particular, there would be much confusion if an encryption policy
1336 policy structs (see `Setting an encryption policy`_), except that the
1338 by the kernel and is used as KDF input or as a tweak to cause
1339 different files to be encrypted differently; see `Per-file encryption
1343 -----------------
1352 For the read path (->read_folio()) of regular files, filesystems can
1353 read the ciphertext into the page cache and decrypt it in-place. The
1357 For the write path (->writepage()) of regular files, filesystems
1358 cannot encrypt data in-place in the page cache, since the cached
1366 -----------------------------
1370 filename hashes. When a ->lookup() is requested, the filesystem
1380 i.e. the bytes actually stored on-disk in the directory entries. When
1381 asked to do a ->lookup() with the key, the filesystem just encrypts
1382 the user-supplied name to get the ciphertext.
1386 filenames. Therefore, readdir() must base64url-encode the ciphertext
1387 for presentation. For most filenames, this works fine; on ->lookup(),
1388 the filesystem just base64url-decodes the user-supplied name to get
1395 filesystem-specific hash(es) needed for directory lookups. This
1397 the filename given in ->lookup() back to a particular directory entry
1404 ``rm -r`` work as expected on encrypted directories.
1414 f2fs encryption using `kvm-xfstests
1415 <https://github.com/tytso/xfstests-bld/blob/master/Documentation/kvm-quickstart.md>`_::
1417 kvm-xfstests -c ext4,f2fs -g encrypt
1418 kvm-xfstests -c ext4,f2fs -g encrypt -m inlinecrypt
1421 a separate command, and it takes some time for kvm-xfstests to set up
1424 kvm-xfstests -c ubifs -g encrypt
1426 No tests should fail. However, tests that use non-default encryption
1428 algorithms were not built into the kernel's crypto API. Also, tests
1437 kvm-xfstests, use the "encrypt" filesystem configuration::
1439 kvm-xfstests -c ext4/encrypt,f2fs/encrypt -g auto
1440 kvm-xfstests -c ext4/encrypt,f2fs/encrypt -g auto -m inlinecrypt
1442 Because this runs many more tests than "-g encrypt" does, it takes
1443 much longer to run; so also consider using `gce-xfstests
1444 <https://github.com/tytso/xfstests-bld/blob/master/Documentation/gce-xfstests.md>`_
1445 instead of kvm-xfstests::
1447 gce-xfstests -c ext4/encrypt,f2fs/encrypt -g auto
1448 gce-xfstests -c ext4/encrypt,f2fs/encrypt -g auto -m inlinecrypt