crypttab — Configuration for encrypted block devices
/etc/crypttab
The /etc/crypttab file describes
encrypted block devices that are set up during system boot.
Empty lines and lines starting with the "#"
character are ignored. Each of the remaining lines describes one
encrypted block device. Fields are delimited by white space.
Each line is in the form
volume-nameencrypted-devicekey-fileoptions
The first two fields are mandatory, the remaining two are optional.
Setting up encrypted block devices using this file supports three encryption modes: LUKS, TrueCrypt and plain. See cryptsetup(8) for more information about each mode. When no mode is specified in the options field and the block device contains a LUKS signature, it is opened as a LUKS device; otherwise, it is assumed to be in raw dm-crypt (plain mode) format.
The first field contains the name of the resulting encrypted volume; its block device is set up
below /dev/mapper/.
The second field contains a path to the underlying block
device or file, or a specification of a block device via
"UUID=" followed by the UUID.
The third field specifies an absolute path to a file with the encryption key. Optionally,
the path may be followed by ":" and an fstab device specification (e.g. starting with
"LABEL=" or similar); in which case the path is taken relative to the device file system
root. If the field is not present or is "none" or "-", a key file
named after the volume to unlock (i.e. the first column of the line), suffixed with
.key is automatically loaded from the /etc/cryptsetup-keys.d/
and /run/cryptsetup-keys.d/ directories, if present. Otherwise, the password has to
be manually entered during system boot. For swap encryption, /dev/urandom may be
used as key file.
The fourth field, if present, is a comma-delimited list of options. The following options are recognized:
cipher=¶Specifies the cipher to use. See cryptsetup(8)
for possible values and the default value of this option. A cipher with unpredictable IV values, such
as "aes-cbc-essiv:sha256", is recommended. Embedded commas in the cipher
specification need to be escaped by preceding them with a backslash, see example below.
discard¶Allow discard requests to be passed through the encrypted block device. This improves performance on SSD storage but has security implications.
hash=¶Specifies the hash to use for password hashing. See cryptsetup(8) for possible values and the default value of this option.
header=¶Use a detached (separated) metadata device or file where the LUKS header is stored. This option is only relevant for LUKS devices. See cryptsetup(8) for possible values and the default value of this option.
Optionally, the path may be followed by ":" and an fstab device specification
(e.g. starting with "UUID=" or similar); in which case, the path is relative to the
device file system root. The device gets mounted automatically for LUKS device activation duration only.
keyfile-offset=¶Specifies the number of bytes to skip at the start of the key file. See cryptsetup(8) for possible values and the default value of this option.
keyfile-size=¶Specifies the maximum number of bytes to read from the key file. See cryptsetup(8) for possible values and the default value of this option. This option is ignored in plain encryption mode, as the key file size is then given by the key size.
keyfile-erase¶If enabled, the specified key file is erased after the volume is activated or when
activation fails. This is in particular useful when the key file is only acquired transiently before
activation (e.g. via a file in /run/, generated by a service running before
activation), and shall be removed after use. Defaults to off.
key-slot=¶Specifies the key slot to compare the
passphrase or key against. If the key slot does not match the
given passphrase or key, but another would, the setup of the
device will fail regardless. This option implies
luks. See
cryptsetup(8)
for possible values. The default is to try all key slots in
sequential order.
keyfile-timeout=¶Specifies the timeout for the device on which the key file resides and falls back to a password if it could not be mounted. See systemd-cryptsetup-generator(8) for key files on external devices.
luks¶Force LUKS mode. When this mode is used, the
following options are ignored since they are provided by the
LUKS header on the device: cipher=,
hash=,
size=.
bitlk¶Decrypt Bitlocker drive. Encryption parameters are deduced by cryptsetup from Bitlocker header.
_netdev¶Marks this cryptsetup device as requiring network. It will be
started after the network is available, similarly to
systemd.mount(5)
units marked with _netdev. The service unit to set up this device
will be ordered between remote-fs-pre.target and
remote-cryptsetup.target, instead of
cryptsetup-pre.target and
cryptsetup.target.
Hint: if this device is used for a mount point that is specified in
fstab(5),
the _netdev option should also be used for the mount
point. Otherwise, a dependency loop might be created where the mount point
will be pulled in by local-fs.target, while the
service to configure the network is usually only started after
the local file system has been mounted.
noauto¶This device will not be added to cryptsetup.target.
This means that it will not be automatically unlocked on boot, unless something else pulls
it in. In particular, if the device is used for a mount point, it'll be unlocked
automatically during boot, unless the mount point itself is also disabled with
noauto.
nofail¶This device will not be a hard dependency of
cryptsetup.target. It'll still be pulled in and started, but the system
will not wait for the device to show up and be unlocked, and boot will not fail if this is
unsuccessful. Note that other units that depend on the unlocked device may still fail. In
particular, if the device is used for a mount point, the mount point itself also needs to
have the nofail option, or the boot will fail if the device is not unlocked
successfully.
offset=¶Start offset in the backend device, in 512-byte sectors. This option is only relevant for plain devices.
plain¶Force plain encryption mode.
read-only, readonly¶Set up the encrypted block device in read-only mode.
same-cpu-crypt¶Perform encryption using the same cpu that IO was submitted on. The default is to use an unbound workqueue so that encryption work is automatically balanced between available CPUs.
This requires kernel 4.0 or newer.
submit-from-crypt-cpus¶Disable offloading writes to a separate thread after encryption. There are some situations where offloading write requests from the encryption threads to a dedicated thread degrades performance significantly. The default is to offload write requests to a dedicated thread because it benefits the CFQ scheduler to have writes submitted using the same context.
This requires kernel 4.0 or newer.
no-read-workqueue¶Bypass dm-crypt internal workqueue and process read requests synchronously. The default is to queue these requests and process them asynchronously.
This requires kernel 5.9 or newer.
no-write-workqueue¶Bypass dm-crypt internal workqueue and process write requests synchronously. The default is to queue these requests and process them asynchronously.
This requires kernel 5.9 or newer.
skip=¶How many 512-byte sectors of the encrypted data to skip at the
beginning. This is different from the offset= option with respect
to the sector numbers used in initialization vector (IV) calculation. Using
offset= will shift the IV calculation by the same negative
amount. Hence, if offset= is given,
sector nn will get a sector number of 0 for the IV
calculation. Using skip= causes sector
n to also be the first sector of the mapped device, but
with its number for IV generation being n.
This option is only relevant for plain devices.
size=¶Specifies the key size in bits. See cryptsetup(8) for possible values and the default value of this option.
sector-size=¶Specifies the sector size in bytes. See cryptsetup(8) for possible values and the default value of this option.
swap¶The encrypted block device will be used as a
swap device, and will be formatted accordingly after setting
up the encrypted block device, with
mkswap(8).
This option implies plain.
WARNING: Using the swap option will
destroy the contents of the named partition during every boot,
so make sure the underlying block device is specified
correctly.
tcrypt¶Use TrueCrypt encryption mode. When this mode
is used, the following options are ignored since they are
provided by the TrueCrypt header on the device or do not
apply:
cipher=,
hash=,
keyfile-offset=,
keyfile-size=,
size=.
When this mode is used, the passphrase is read from the key file given in the third field. Only the first line of this file is read, excluding the new line character.
Note that the TrueCrypt format uses both passphrase and
key files to derive a password for the volume. Therefore, the
passphrase and all key files need to be provided. Use
tcrypt-keyfile= to provide the absolute path
to all key files. When using an empty passphrase in
combination with one or more key files, use
"/dev/null" as the password file in the third
field.
Use the hidden TrueCrypt volume. This option
implies tcrypt.
This will map the hidden volume that is inside of the volume provided in the second field. Please note that there is no protection for the hidden volume if the outer volume is mounted instead. See cryptsetup(8) for more information on this limitation.
tcrypt-keyfile=¶Specifies the absolute path to a key file to
use for a TrueCrypt volume. This implies
tcrypt and can be used more than once to
provide several key files.
See the entry for tcrypt on the
behavior of the passphrase and key files when using TrueCrypt
encryption mode.
tcrypt-system¶Use TrueCrypt in system encryption mode. This
option implies tcrypt.
tcrypt-veracrypt¶Check for a VeraCrypt volume. VeraCrypt is a fork of
TrueCrypt that is mostly compatible, but uses different, stronger key
derivation algorithms that cannot be detected without this flag.
Enabling this option could substantially slow down unlocking, because
VeraCrypt's key derivation takes much longer than TrueCrypt's. This
option implies tcrypt.
timeout=¶Specifies the timeout for querying for a password. If no unit is specified, seconds is used. Supported units are s, ms, us, min, h, d. A timeout of 0 waits indefinitely (which is the default).
tmp=¶The encrypted block device will be prepared for using it as
/tmp/; it will be formatted using mkfs(8). Takes
a file system type as argument, such as "ext4", "xfs" or
"btrfs". If no argument is specified defaults to "ext4". This
option implies plain.
WARNING: Using the tmp option will destroy the contents of the named partition
during every boot, so make sure the underlying block device is specified correctly.
tries=¶Specifies the maximum number of times the user is queried for a password. The default is 3. If set to 0, the user is queried for a password indefinitely.
verify¶If the encryption password is read from console, it has to be entered twice to prevent typos.
pkcs11-uri=¶Takes a RFC7512 PKCS#11 URI pointing to a private RSA key which is used to decrypt the key specified in the third column of the line. This is useful for unlocking encrypted volumes through security tokens or smartcards. See below for an example how to set up this mechanism for unlocking a LUKS volume with a YubiKey security token. The specified URI can refer directly to a private RSA key stored on a token or alternatively just to a slot or token, in which case a search for a suitable private RSA key will be performed. In this case if multiple suitable objects are found the token is refused. The key configured in the third column is passed as is to RSA decryption. The resulting decrypted key is then base64 encoded before it is used to unlock the LUKS volume.
try-empty-password=¶Takes a boolean argument. If enabled, right before asking the user for a password it is first attempted to unlock the volume with an empty password. This is useful for systems that are initialized with an encrypted volume with only an empty password set, which shall be replaced with a suitable password during first boot, but after activation.
x-systemd.device-timeout=¶Specifies how long systemd should wait for a device to show up
before giving up on the entry. The argument is a time in seconds or explicitly
specified units of
"s",
"min",
"h",
"ms".
x-initrd.attach¶Setup this encrypted block device in the initramfs, similarly to
systemd.mount(5)
units marked with x-initrd.mount.
Although it's not necessary to mark the mount entry for the root file system with
x-initrd.mount, x-initrd.attach is still recommended with
the encrypted block device containing the root file system as otherwise systemd will
attempt to detach the device during the regular system shutdown while it's still in
use. With this option the device will still be detached but later after the root file
system is unmounted.
All other encrypted block devices that contain file systems mounted in the initramfs should use this option.
At early boot and when the system manager configuration is reloaded, this file is translated into native systemd units by systemd-cryptsetup-generator(8).
Example 1. /etc/crypttab example
Set up four encrypted block devices. One using LUKS for normal storage, another one for usage as
a swap device and two TrueCrypt volumes. For the fourth device, the option string is interpreted as two
options "cipher=xchacha12,aes-adiantum-plain64",
"keyfile-timeout=10s".
luks UUID=2505567a-9e27-4efe-a4d5-15ad146c258b swap /dev/sda7 /dev/urandom swap truecrypt /dev/sda2 /etc/container_password tcrypt hidden /mnt/tc_hidden /dev/null tcrypt-hidden,tcrypt-keyfile=/etc/keyfile external /dev/sda3 keyfile:LABEL=keydev keyfile-timeout=10s,cipher=xchacha12\,aes-adiantum-plain64
Example 2. Yubikey-based Volume Unlocking Example
The PKCS#11 logic allows hooking up any compatible security token that is capable of storing RSA decryption keys. Here's an example how to set up a Yubikey security token for this purpose, using ykmap(1) from the yubikey-manager project:
# Make sure no one can read the files we generate but us umask 077 # Destroy any old key on the Yubikey (careful!) ykman piv reset # Generate a new private/public key pair on the device, store the public key in 'pubkey.pem'. ykman piv generate-key -a RSA2048 9d pubkey.pem # Create a self-signed certificate from this public key, and store it on the # device. The "subject" should be an arbitrary string to identify the token in # the p11tool output below. ykman piv generate-certificate --subject "Knobelei" 9d pubkey.pem # Check if the newly create key on the Yubikey shows up as token in PKCS#11. Have a look at the output, and # copy the resulting token URI to the clipboard. p11tool --list-tokens # Generate a (secret) random key to use as LUKS decryption key. dd if=/dev/urandom of=plaintext.bin bs=128 count=1 # Encode the secret key also as base64 text (with all whitespace removed) base64 < plaintext.bin | tr -d '\n\r\t ' > plaintext.base64 # Encrypt this newly generated (binary) LUKS decryption key using the public key whose private key is on the # Yubikey, store the result in /etc/cryptsetup-keys.d/mytest.key, where we'll look for it during boot. mkdir -p /etc/cryptsetup-keys.d sudo openssl rsautl -encrypt -pubin -inkey pubkey.pem -in plaintext.bin -out /etc/cryptsetup-keys.d/mytest.key # Configure the LUKS decryption key on the LUKS device. We use very low pbkdf settings since the key already # has quite a high quality (it comes directly from /dev/urandom after all), and thus we don't need to do much # key derivation. Replace /dev/sdXn by the partition to use (e.g. sda1) sudo cryptsetup luksAddKey /dev/sdXn plaintext.base64 --pbkdf=pbkdf2 --pbkdf-force-iterations=1000 # Now securely delete the plain text LUKS key, we don't need it anymore, and since it contains secret key # material it should be removed from disk thoroughly. shred -u plaintext.bin plaintext.base64 # We don't need the public key anymore either, let's remove it too. Since this one is not security # sensitive we just do a regular "rm" here. rm pubkey.pem # Test: Let's run systemd-cryptsetup to test if this all worked. The option string should contain the full # PKCS#11 URI we have in the clipboard; it tells the tool how to decipher the encrypted LUKS key. Note that # systemd-cryptsetup automatically searches for the encrypted key in /etc/cryptsetup-keys.d/, hence we do # not need to specify the key file path explicitly here. sudo systemd-cryptsetup attach mytest /dev/sdXn - 'pkcs11-uri=pkcs11:…' # If that worked, let's now add the same line persistently to /etc/crypttab, for the future. sudo bash -c 'echo "mytest /dev/sdXn - \'pkcs11-uri=pkcs11:…\'" >> /etc/crypttab'
A few notes on the above:
We use RSA (and not ECC), since Yubikeys support PKCS#11 Decrypt() only for RSA keys
We use RSA2048, which is the longest key size current Yubikeys support
LUKS key size must be shorter than 2048bit due to RSA padding, hence we use 128 bytes
We use Yubikey key slot 9d, since that's apparently the keyslot to use for decryption purposes, see documentation.