dm-crypt

From ArchWiki
Revision as of 20:39, 14 September 2013 by Indigo (Talk | contribs) (Securing the unencrypted boot partition: adding aide and potts/stark as alternatives to chkboot)

Jump to: navigation, search

Merge-arrows-2.pngThis article or section is a candidate for merging with Plain dm-crypt without LUKS.Merge-arrows-2.png

Notes: Assess the possibility of merging the common content between the two articles in order to avoid duplication. (Discuss in Talk:Plain dm-crypt without LUKS#Merge)
Summary help replacing me
This tutorial will show you how to set up system encryption with LUKS for dm-crypt.
Related
Disk Encryption
Removing System Encryption
Plain dm-crypt without LUKS

This article focuses on how to set up full system encryption on Arch Linux, using dm-crypt with LUKS.

dm-crypt is the standard device-mapper encryption functionality provided by the Linux kernel. It can be used directly by those who like to have full control over all aspects of partition and key management.

LUKS is an additional convenience layer which stores all of the needed setup information for dm-crypt on the disk itself and abstracts partition and key management in an attempt to improve ease of use.

For more details on how dm-crypt+LUKS compares to other disk encryption solution, see Disk Encryption#Comparison table.

Contents

Caveats

Warning: Encrypting a disk or partition will erase everything currently on that disk or partition:
  • Please make appropriate data backups prior to starting

Also be aware that encrypting a system might not only make the life of laptop thieves more miserable, but also yours if you don't plan ahead on

  • how to make secure backups of the encrypted system/-setup/data and
  • how to access the encrypted system manually for maintenance.
Keeping those points in mind while deciding on how to use encryption may help to decide on method and tools as well.

Initial Setup

Overview and Preparation

The installation of a LUKS-encrypted system is largely the same as installing an unencrypted system. Routine creation of an encrypted system follows these general steps:

  • Secure erasure of the hard disk drive(s)
  • Partitioning and setup of encryption (LVM optional)
  • Routine package selection and installation
  • System configuration to handle the encryption

This page covers the first two points in a general way for different configuration options available with LUKS.

The third and fourth point are covered in the later sections. Since the Arch installation media comes with all the tools required for system encryption, you can follow the Installation Guide or the Beginners' Guide after the encrypted partitions are set up. You will have to adjust the system configuration to be able to boot from your LUKS-volumes though, which is also covered in setup examples in the later sections.

Secure erasure of the hard disk drive

Before encrypting a drive, you should perform a secure erase of the disk by overwriting the entire drive with random data. To prevent cryptographic attacks or unwanted File Recovery, this data should be completely indistinguishable from all data later written by dm-crypt.

In deciding which method to use for secure erasure of a hard disk drive, remember that this needs only to be performed once for as long as the drive is used as an encrypted drive. For detailed instructions on how to erase and prepare a drive consult Securely wipe disk. The following methods are specific for dm-crypt/LUKS and are mentioned complementary, because they can be performed after a partition setup too.

Use LUKS container as pseudorandom number generator (alternate)

The cryptsetup FAQ mentions a very simple procedure to use an existing dm-crypt-volume to wipe all free space accessible on the underlying block device with random data by acting as a simple pseudorandom number generator. It is also claimed to protect against disclosure of usage patterns.

# dd if=/dev/zero of=/dev/mapper/luks-container
Wipe free space with encrypted file after Installation

The same effect can be achieved if a file is created on each encrypted partition that fills the partition completely after the system is installed, booted and filesystems mounted. That is because encrypted data is indistinguishable from random.

# dd if=/dev/zero of=/file/in/luks-container
# rm /file/in/luks-container

The above process has to be repeated for every container created.

Wipe LUKS keyslots

#cryptsetup luksKillSlot <device> <key slot number>

This will only wipe a single keyslot.

Wipe LUKS header

The partitions formatted with dm-crypt/LUKS contain a header with the cipher and crypt-options used, which is referred to dm-mod when opening the blockdevice. After the header the actual random data partition starts. Hence, when de-commissioning a drive (e.g. sale of PC, switch of drives, etc.) it may be just enough to wipe the header of the partition, rather than overwriting the whole drive - which can be a lengthy process.

Wiping the LUKS header will delete the PBKDF2-encrypted (AES) master key, salts and so on.

Note: It is crucial to write to the LUKS encrypted partition (/dev/sda1 in this example) and not directly to the disks device node. If you did set up encryption as a device-mapper layer on top of others, e.g. LVM on LUKS on RAID then write to RAID respectively.

A header with one single default 256 bit size keyslot is 1024KB in size. It is advised to also overwrite the first 4KB written by dm-crypt, so 1028KB have to be wiped. That is 1052672 Byte.

For zero offset use:

#head -c 1052672 /dev/zero > /dev/sda1; sync

For 512 bit key length (e.g. for aes-xts-plain with 512 bit key) the header is 2MB.

If in doubt, just be generous and overwrite the first 10MB or so.

#dd if=/dev/zero of=/dev/sda1 bs=512 count=20480
Note: With a backup-copy of the header data can get rescued but the filesystem was likely damaged as the first encrypted sectors were overwritten. See further sections on how to make a backup of the crucial header blocks.

When wiping the header with random data everything left on the device is encrypted data. An exception to this may occur for an SSD, because of cache blocks SSDs employ. In theory it may happen that the header was cached in these some time before and that copy may consequently be still available after wiping the original header. For strong security concerns, a secure ATA erase of the SSD should be done (procedure please see the cryptsetup FAQ 5.19).

Discard/TRIM support for solid state disks (SSD)

Solid state disk users should be aware that by default, Linux's full-disk encryption mechanisms will not forward TRIM commands from the filesystem to the underlying disk. The device-mapper maintainers have made it clear that TRIM support will never be enabled by default on dm-crypt devices because of the potential security implications.

Most users will still want to use TRIM on their encrypted SSDs. Minimal data leakage in the form of freed block information, perhaps sufficient to determine the filesystem in use, may occur on devices with TRIM enabled. An illustration and discussion of the issues arising from activating TRIM is available in the blog of a cryptsetup developer.

As a semi-tangential caveat, it is worth noting that because TRIM provides information to the disk firmware about which blocks contain data, encryption schemes that rely on plausible deniability, like TrueCrypt's hidden volumes, should never be used on a device that utilizes TRIM. This is probably also valid for TC containers within a LUKS encrypted device that uses TRIM.

TrueCrypt's developers also recommend against using any TC volume on a device that performs wear-leveling techniques to extend the life of the disk; most flash devices, including SSDs and USB flash drives, use mandatory wear-leveling at the firmware level. LUKS devices are probably not vulnerable to problems with wear-leveling if the entire device is blanked before the LUKS partition is initialized. See here and here for more information.

In linux 3.1 and up, support for dm-crypt TRIM pass-through can be toggled upon device creation or mount with dmsetup. Support for this option also exists in cryptsetup version 1.4.0 and up. To add support during boot, you will need to add :allow-discards to the cryptdevice option. The TRIM option may look like this:

cryptdevice=/dev/mapper/root:root:allow-discards

For the main cryptdevice configuration options before the :allow-discards please refer to the sections following. Besides the kernel option, it is also required to mount the filesystem (e.g. /dev/mapper/root in this example) with the discard option in /etc/fstab. For details, please refer to the SSD page. For LUKS devices unlocked manually on the console or via /etc/crypttab the respective option to be used is allow-discards again.

Partitioning

After the drive has been securely overwritten, it is time to create partitions and begin setting up an encrypted system.

There are multiple ways to create disk partitions:

LUKS can be used with systems that require LVM and/or RAID and is compatible with all regular partitioning standards in Linux.

Creating Disk Partitions

There are two required partitions for a basic encrypted system setup:

root file system
/ Will be encrypted and store all system and user files (/usr, /bin, /var, /home, etc.)
initial boot partition
/boot Will not be encrypted; the bootloader needs to access the /boot directory where it will load the initramfs/encryption modules needed to load the rest of the system which is encrypted (see Mkinitcpio for details). For this reason, /boot needs to reside on its own, unencrypted partition.

This partition layout is encrypted in this example and can be refined according to needs, e.g. by separating partitions or adding an encrypted swap-partition.

Single Disk Systems

If there are additional partitions desired, these can be individually created by defining separate primary or extended/logical partitions. It is possible to encrypt separate partitions (e.g. /home, /var) while leaving others (e.g. /usr) unencrypted as required. However, a standard install would also require separate passphrases or keys to open each encrypted partition during boot.

Multiple Disk Systems

In systems with multiple hard disk drives, the same options exist as a single disk system. After the creation of the /boot partition, the remaining free space on the physical disks can be divided up into their respective partitions at this level. For encrypted partitions that span multiple disks, LUKS must be used with RAID or LVM.

LVM: Logical Volume Manager

The LVM allows for systems that require complex hard disk configurations. Knowledge of using LVM is a requisite to continue with setting up LUKS with LVM later in this article.

Tip: Btrfs has a built-in Subvolume-Feature that fully replaces the need for LVM if no other filesystems are required. An encrypted swap is not possible this way and swap files are not supported by btrfs up to now.
LVM on LUKS

There is a growing preference towards logical volume management of LUKS encrypted physical media (LVM on LUKS). A reason for this is LVM's flexibility, e.g. for resizing volumes as needed. In a LVM on LUKS scenario, the LUKS-partition has to be opened and mapped before LVM can access the underlaying setup volumes. But once LVM has activated the volumes, the encryption is transparent and LVM tools can be used to change volumes without affecting the LUKS-blockdevice.

This setup is using LUKS as the lowest level of infrastructure most closely approximating the deployment of physical disks with built-in hardware encryption. In that case, logical volume management would be layered on top of the hardware encryption - usage of LUKS may be superfluous depending on requirements (e.g. open-source).

LUKS on LVM

Other usage scenarios exist where encrypting logical volumes rather than a physical disk is required (LUKS on LVM). One may be a requirement for utmost flexibility of assigning available diskspace across multiple disks, which is not possible with LVM on LUKS. Another scenario may be a mix of unencrypted and encrypted volumes.

Upon boot the LVM is setup and assigned before the LUKS-encrypted volumes are opened. In order to manage changes of volumes in a LUKS on LVM setup, both layers' setup have to be taken into account, i.e. shrinking or expanding an encrypted volume has to include the resizing of the filesystem, which is hidden for LVM inside the encrypted LUKS blockdevice, to ensure filesystem integrity.

Configuring LUKS

This section covers how to manually utilize LUKS from the command line to encrypt a system.

Mapping Physical Partitions to LUKS

After writing the partition table to the MBR (optionally set up LVM thereafter) the next step is to create the LUKS and dm-crypt magic and make device mapper mount it to the filesystem of the installation system.

When creating LUKS partitions they must be associated with a key. The key is used to unlock the header of the LUKS-encrypted partitions.

A key is either a:

  • Passphrase
  • Keyfile

It is possible to define up to 8 different keys per LUKS partition. This enables the user to create access keys for save backup storage. Also a different key-slot could be used to grant access to a partition to a user by issuing a second key and later revoking it again without the need to re-encrypt the partition. Having in mind that further passphrases or keyfiles can be added later easily at any time might make the choice for the initial key easier.

Using LUKS to Format Partitions with a Passphrase

Note: Using a passphrase to decrypt LUKS partitions automatically from /etc/crypttab is deprecated.

Cryptsetup is used to interface with LUKS for formatting, mounting and unmounting encrypted partitions.

Usage:

# cryptsetup [OPTION...] <action> <action-specific>

Example:

# cryptsetup --cipher aes-xts-plain64 --key-size 512 --hash sha512 --iter-time 5000 --use-random --verify-passphrase luksFormat <device> 

Common options used with luksFormat:

Available options Cryptsetup (<1.6.0) defaults Comment Example Comment
--cipher, -c/ aes-cbc-essiv:sha256 Use the AES-cipher with CBC/ESSIV. aes-xts-plain64 XTS. For volumes >2TiB use aes-xts-plain64 (requires kernel >= 2.6.33).
--key-size, -s 256 The cipher is used with 256 bit key-size. 512 XTS splits the supplied key into fraternal twins. For an effective AES-256 the XTS key-size must be 512.
--hash, -h sha1 Hash algorithm used for PBKDF2. sha512
--iter-time, -i 1000 Number of milliseconds to spend with PBKDF2 passphrase processing. 5000 Using a hash stronger than sha1 results in less iterations if iter-time is not increased.
--use-random --use-urandom /dev/urandom is used as randomness source for the (long-term) volume master key. --use-random Avoid generating an insecure master key if low on entropy. Will block if the entropy pool is used up.
--verify-passphrase, -y Yes Default only for luksFormat and luksAddKey. - No need to type for archlinux at the moment.

Please note that the above compares historic cryptsetup defaults in the left column. With release 1.6.0 the defaults have changed to an AES cipher in XTS mode, but with other options than the right column example (e.g. an effective key-size of 128-bit). The defaults can be checked with the tail output of

# cryptsetup --help

When deciding on the encryption cipher to use during blockdevice creation, it has to be taken into account that a number of the options stated affect system performance just for creating the initial crypt-blockdevice or opening it, but not the crypto and disk-io operations when the system is running. The throughput and security of the crypted data itself depends on the cipher and key-size. The used hash, iteration-time and random source options affect the cryptographic security of the master-key creation and processing time needed to unlock it in the future.

A full list of options cryptsetup accepts can be found in the manpage. Furthermore, cryptsetup now has a feature to benchmark the crypto performance of the processor:

# cryptsetup benchmark 

can give guidance on deciding for a cipher to use prior to installation. If certain AES ciphers appear in the output to have a considerably (e.g. tenfold) higher throughput, these are probably the ones with hardware support in the CPU.

In the following examples for creating LUKS partitions, we will use the AES cipher in XTS mode; at present this is not only the default, but also a most generally used preferred cipher. More information on this and other ciphers used with cryptsetup can be found here: Wikipedia:Block_cipher

Formatting LUKS Partitions

First of all make sure the device mapper kernel module is loaded by executing the following: # modprobe dm_mod

In order to format a desired partition as an encrypted LUKS partition execute:

# cryptsetup -c <cipher> -y -s <key size> luksFormat /dev/<partition name>
Enter passphrase: <password>
Verify passphrase: <password>

Check results:

# cryptsetup luksDump /dev/<drive>

This should be repeated for all partitions except for /boot and possibly swap. You will note that the dump not only shows the cipher header information, but also the key-slots in use for the LUKS partition.

The example below will create an encrypted root partition using the AES cipher in XTS mode (generally referred to as XTS-AES).

# cryptsetup -c aes-xts-plain -y -s 512 luksFormat /dev/sda2
Note: If hibernation usage is planned, swap must be encrypted in this fashion; otherwise, if hibernation is not a planned feature for the system, encrypting the swap file will be performed in a alternative manner.
Warning: Irrespective of the chosen partitioning method, the /boot partition must remain separate and unencrypted in order to load the kernel and boot the system.

Unlocking/Mapping LUKS Partitions with the Device Mapper

Once the LUKS partitions have been created it is time to unlock them.

The unlocking process will map the partitions to a new device name using the device mapper. This alerts the kernel that /dev/<partition name> is actually an encrypted device and should be addressed through LUKS using the /dev/mapper/<name> so as not to overwrite the encrypted data. To guard against accidental overwriting, read about the possibilities to backup the cryptheader after finishing setup.

In order to open an encrypted LUKS partition execute:

# cryptsetup open --type luks /dev/<partition name> <device-mapper name>
Enter any LUKS passphrase: <password>
key slot 0 unlocked.
Command successful.

Usually the device mapped name is descriptive of the function of the partition that is mapped, example:

cryptsetup open --type luks /dev/sda2 swap
Once opened, the swap partition device address would be /dev/mapper/swap instead of /dev/sda2.
cryptsetup open --type luks /dev/sda3 root
Once opened, the root partition device address would be /dev/mapper/root instead of /dev/sda3.
cryptsetup open --type luks /dev/sda3 lvmpool (alternate)
For setting up LVM ontop the encryption layer the device file for the decrypted volume group would be anything like /dev/mapper/lvmpool instead of /dev/sda3. LVM will then give additional names to all logical volumes created, e.g. /dev/mapper/lvmpool-root and /dev/mapper/lvmpool-swap.

In order to write encrypted data into the partition it must be accessed through the device mapped name.

Note: Since /boot is not encrypted, it does not need a device mapped name and will be addressed as /dev/sda1.

Using LUKS to Format Partitions with a Keyfile

When creating a new LUKS encrypted partition, a keyfile may be associated with the partition on its creation using:

# cryptsetup -c <desired cipher> -s <key size> luksFormat /dev/<volume to encrypt> /path/to/mykeyfile

This is accomplished by appending the bold area to the standard cryptsetup command which defines where the keyfile is located.

Adding Additional Passphrases or Keyfiles to a LUKS Encrypted Partition

LUKS supports the association of up to 8 keyslots with any single encrypted volume. Keyslots can be either keyfiles or passphrases.

Once an encrypted partition has been created, the initial keyslot 0 is created. Additional keyslots are numbered from 1 to 7.

Adding new keyslots is accomplished using cryptsetup with the luksAddKey action.

Don't forget wiping unused keyslots with luksKillSlot as described in #Wipe LUKS keyslots.)

# cryptsetup luksAddKey /dev/<volume> (/path/to/<additionalkeyfile>)
Enter any passphrase:
Enter new passphrase for key slot:
Verify passphrase:

Where <device> is the volume containing the LUKS header to wich the new keyslot is added. This works on header backup files as well.

If /path/to/<additionalkeyfile> is given, cryptsetup will add a new keyslot for <additionalkeyfile>. Otherwise a new passphrase will be prompted for twice.

For adding keyslots cryptsetup has to decrypt the master key from an existing keyslot so it first asks for "any passphrase" (of an existing keyslot).

For getting the master key from an existing keyfile keyslot the --key-file or -d option followed by the "old" <keyfile> will try to unlock all available keyfile keyslots.

# cryptsetup luksAddKey /dev/mapper/<device> (/path/to/<additionalkeyfile>) -d /path/to/<keyfile>

Using Cryptsetup with a Keyfile

Note: This section describes using a plaintext keyfile. If you want to encrypt your keyfile giving you two factor authentication see Using GPG or OpenSSL Encrypted Keyfiles for details, but please still read this section.

What is a Keyfile?

A keyfile is any file in which the data contained within it is used as the passphrase to unlock an encrypted volume. Therefore if these files are lost or changed, decrypting the volume will no longer be possible.

Tip: Define a passphrase in addition to the keyfile for backup access to encrypted volumes in the event the defined keyfile is lost or changed.

Why use a Keyfile?

There are many kinds of keyfile. Each type of keyfile used has benefits and disadvantages summarized below:

keyfile.passphrase:
this is my passphrase I would have typed during boot but I have placed it in a file instead

This is a keyfile containing a simple passphrase. The benefit of this type of keyfile is that if the file is lost the data it contained is known and hopefully easily remembered by the owner of the encrypted volume. However the disadvantage is that this does not add any security over entering a passphrase during the initial system start.

keyfile.randomtext:
fjqweifj830149-57 819y4my1- 38t1934yt8-91m 34co3;t8y;9p3y-

This is a keyfile containing a block of random characters. The benefit of this type of keyfile is that it is much more resistant to dictionary attacks than a simple passphrase. An additional strength of keyfiles can be utilized in this situation which is the length of data used. Since this is not a string meant to be memorized by a person for entry, it is trivial to create files containing thousands of random characters as the key. The disadvantage is that if this file is lost or changed, it will most likely not be possible to access the encrypted volume without a backup passphrase.

keyfile.binary:
where any binary file, images, text, video could be chosen as the keyfile

This is a binary file that has been defined as a keyfile. When identifying files as candidates for a keyfile, it is recommended to choose files that are relatively static such as photos, music, video clips. The benefit of these files is that they serve a dual function which can make them harder to identify as keyfiles. Instead of having a text file with a large amount of random text, the keyfile would look like a regular image file or music clip to the casual observer. The disadvantage is that if this file is lost or changed, it will most likely not be possible to access the encrypted volume without a backup passphrase. Additionally, there is a theoretical loss of randomness when compared to a randomly generated text file. This is due to the fact that images, videos and music have some intrinsic relationship between neighboring bits of data that does not exist for a text file. However this is controversial and has never been exploited publicly.

Creating a Keyfile with Random Characters

Here dd is used to generate a keyfile of 2048 random bytes.

# dd if=/dev/urandom of=mykeyfile bs=512 count=4

The usage of dd is similar to initially wiping the volume with random data prior to encryption.

Storing the Key File

Tango-view-fullscreen.pngThis article or section needs expansion.Tango-view-fullscreen.png

Reason: The default method to store a keyfile and reference it via crypttab should be added. (Discuss in Talk:Dm-crypt#)
External Storage on a USB Drive
Preparation for Persistent block device naming

For reading the file from an external storage device it is very convenient to access it through udev's Persistent block device naming features and not by ordinary device nodes like /dev/sdb1 whose naming depends on the order in which devices are plugged in. So in order to assure that the encrypt HOOK in the initcpio finds your keyfile, you must use a permanent device name.

Persistent symlinks

Merge-arrows-2.pngThis article or section is a candidate for merging with Persistent block device naming.Merge-arrows-2.png

Notes: Anything not specific to storing LUKS keyfiles should get merged there. (Discuss in Talk:Dm-crypt#)

A quick method (as opposed to setting up a udev rule) for doing so involves referencing the right partition by its UUID, id (based on hardware info and serial number) or filesystem label.

Plug the device in and print every file name under /dev/disk:

#ls -lR /dev/disk/
/dev/disk/:
total 0
drwxr-xr-x 2 root root 180 Feb 12 10:11 by-id
drwxr-xr-x 2 root root  60 Feb 12 10:11 by-label
drwxr-xr-x 2 root root 100 Feb 12 10:11 by-path
drwxr-xr-x 2 root root 180 Feb 12 10:11 by-uuid

/dev/disk/by-id:
total 0
lrwxrwxrwx 1 root root  9 Feb 12 10:11 usb-Generic_STORAGE_DEVICE_000000014583-0:0 -> ../../sdb
lrwxrwxrwx 1 root root 10 Feb 12 10:11 usb-Generic_STORAGE_DEVICE_000000014583-0:0-part1 -> ../../sdb1

/dev/disk/by-label:
total 0
lrwxrwxrwx 1 root root 10 Feb 12 10:11 Keys -> ../../sdb1

/dev/disk/by-path:
total 0
lrwxrwxrwx 1 root root  9 Feb 12 10:11 pci-0000:00:1d.7-usb-0:1:1.0-scsi-0:0:0:0 -> ../../sdb
lrwxrwxrwx 1 root root 10 Feb 12 10:11 pci-0000:00:1d.7-usb-0:1:1.0-scsi-0:0:0:0-part1 -> ../../sdb1

/dev/disk/by-uuid:
total 0
lrwxrwxrwx 1 root root 10 Feb 12 10:11 baa07781-2a10-43a7-b876-c1715aba9d54 -> ../../sdb1

UUID

Using the filesystem UUID for persistent block device naming is considered very reliable. Filesystem UUIDs are stored in the filesystem itself, meaning that the UUID will be the same if you plug it into any other computer, and that a dd backup of it will always have the same UUID since dd does a bitwise copy.

The right device node for what is now /dev/sdb1 will always get symlinked by /dev/disk/by-uuid/baa07781-2a10-43a7-b876-c1715aba9d54. Symlinks can be used in the bootloaders "cryptkey" kernel option or anywhere else.

For legacy filesystems like FAT the UUID will be much shorter but collision is still unlikely to happen if not mounting many different FAT filesystems at once.

Label

In the following example a FAT partition is labeled as "Keys" and will always get symlinked by /dev/disk/by-label/Keys:

#mkdosfs -n >volume-name< /dev/sdb1
#blkid -o list
device     fs_type label    mount point    UUID
-------------------------------------------------------
/dev/sdb1  vfat    Keys     (not mounted)  221E-09C0
Persistent udev rule

Optionally you may choose to set up your flash drive with a udev rule. There is some documentation in the Arch wiki about that already; if you want more in-depth, structural info, read this guide. Here is quickly how it goes.

Get the serial number from your USB flash drive:

lsusb -v | grep -A 5 Vendor

Create a udev rule for it by adding the following to a file in /etc/udev/rules.d/, such as 8-usbstick.rules:

KERNEL=="sd*", ATTRS{serial}=="$SERIAL", SYMLINK+="$SYMLINK%n"

Replace $SYMLINK and $SERIAL with their respective values. %n will expand to the partition (just like sda is subdivided into sda1, sda2, ...). You do not need to go with the 'serial' attribute. If you have a custom rule of your own, you can put it in as well (e.g. using the vendor name).

Rescan your sysfs:

udevadm trigger

Now check the contents of /dev:

ls /dev

It should show your device with your desired name.

Generating the keyfile

Optionally you can mount a tmpfs for storing the temporary keyfile.

# mkdir ./mytmpfs
# mount tmpfs ./mytmpfs -t tmpfs -o size=32m
# cd ./mytmpfs

The advantage is that it resides in RAM and not on a physical disk, so after unmounting your keyfile is securly gone. So copy your keyfile to some place you consider as secure before unmounting. If you are planning to store the keyfile as a plain file on your USB device, you can also simply execute the following command in the corresponding directory, e.g. /media/sdb1

The keyfile can be of arbitrary content and size. We will generate a random temporary keyfile of 2048 bytes:

# dd if=/dev/urandom of=secretkey bs=512 count=4

If you stored your temporary keyfile on a physical storage device, remember to not just (re)move the keyfile later on, but use something like

cp secretkey /destination/path
shred --remove --zero secretkey

to securely overwrite it. For overaged filesystems like FAT or ext2 this will suffice while in the case of journaling filesystems, flash memory hardware and other cases it is highly recommended to wipe the entire device or at least the keyfiles partition.

Add a keyslot for the temporary keyfile to the LUKS header:

# cryptsetup luksAddKey /dev/sda2 secretkey
Enter any LUKS passphrase:
key slot 0 unlocked.
Command successful.
Storing the keyfile

The following uses an USB-stick to store the key file and modifies the initramfs to load and use it on boot to unlock the root partition.

Configuration of initcpio

You have to add two extra modules in your /etc/mkinitcpio.conf, one for the drive's file system and one for the codepage. Further if you created a udev rule, you should tell mkinitcpio about it:

MODULES="ata_generic ata_piix nls_cp437 vfat"
FILES="/etc/udev/rules.d/8-usbstick.rules"

In this example it is assumed that you use a FAT formatted USB drive. Replace those module names if you use another file system on your USB stick (e.g. ext2) or another codepage. Users running the stock Arch kernel should stick to the codepage mentioned here.

If you have a non-US keyboard, it might prove useful to load your keyboard layout before you are prompted to enter the password to unlock the root partition at boot. For this, you will need the keymap hook before encrypt.

Generate a new image (maybe you should backup a copy of your old /boot/initramfs-linux.img first):

# mkinitcpio -p linux
Storing the key as a plain (visible) file

Be sure to choose a plain name for your key – a bit of 'security through obscurity' is always nice ;-). Avoid using dotfiles (hidden files) – the encrypt hook will fail to find the keyfile during the boot process.

You have to add cryptdevice=/dev/sda3:root cryptkey=/dev/usbstick:vfat:/secretkey to your kernel parameters. This assumes /dev/usbstick is the FAT partition of your choice. Replace it with /dev/disk/by-... or whatever your device is.

That is all, reboot and have fun!

Encrypting the Swap partition

A swap partition may be added to an encrypted system, if required. The swap partition must be encrypted as well to protect any data swapped out by the system. Depending on the requirements, different methods may be used which are described in the following. A setup where the swap encryption is re-initialised on reboot (with a new encryption) provides higher data protection. However, re-encrypting swap also forbids using a suspend-to-disk feature generally.

Without suspend-to-disk support

In systems where suspend to disk is not a desired feature, it is possible to create a swap file that will have a random master key with each boot. This is accomplished by using dm-crypt directly without LUKS extensions.

The /etc/crypttab is well commented and you can basically just uncomment the swap line and change <device> to a persistent symlink.

/etc/crypttab
# <name>       <device>         <password>              <options>
# swap         /dev/hdx4        /dev/urandom            swap,cipher=aes-cbc-essiv:sha256,size=256

Where:

<name>
Represents the name (/dev/mapper/<name>) to list in /etc/fstab.
<device>
Should be the symlink to the actual partition's device file.
<password>
/dev/urandom sets the dm-crypt master key to be randomized on every volume recreation.
<options>
The swap option runs mkswap after cryptographic's are setup.
Warning: You should use persistent block device naming (in example ID's) for <device> because if there are multiple hard drives installed in the system, their naming order (sda, sdb,...) can occasionally be scrambled upon boot and thus the swap would be created over a valuable file system, destroying its content.

Persistent block device naming is implemented with simple symlinks. Using UUID's or filesystem-labels is not possible as plain dm-crypt writes only encrypted data without a persistent header like LUKS. If you are not familar with one of the directories under /dev/disk/ read on in the section on #Preparation for Persistent block device naming

#ls -l /dev/disk/*/* | grep sda2
lrwxrwxrwx 1 root root 10 Oct 12 16:54 /dev/disk/by-id/ata-WDC_WD2500BEVT-22ZCT0_WD-WXE908VF0470-part2 -> ../../sda2

Example line for the /dev/sda2 symlink from above:

/etc/crypttab
# <name>                      <device>                                   <password>     <options>
  swap  /dev/disk/by-id/ata-WDC_WD2500BEVT-22ZCT0_WD-WXE908VF0470-part2  /dev/urandom   swap,cipher=aes-cbc-essiv:sha256,size=256

This will map /dev/sda2 to /dev/mapper/swap as a swap partition that can be added in /etc/fstab like a normal swap.

If the partition chosen for swap was previously a LUKS partition, crypttab will not overwrite the partition to create a swap partition. This is a safety measure to prevent data loss from accidental mis-identification of the swap partition in crypttab. In order to use such a partition the LUKS header must be overwritten once.

With suspend-to-disk support

To be able to resume after suspending the computer to disk (hibernate), it is required to keep the swap filesystem intact. Therefore, it is required to have a pre-existent LUKS swap partition, which can be stored on the disk or input manually at startup. Because the resume takes place before /etc/crypttab can be used, it is required to create a hook in /etc/mkinitcpio.conf to open the swap LUKS device before resuming.

If you want to use a partition which is currently used by the system, you have to disable it first:

# swapoff /dev/<device>

Also make sure you remove any line in /etc/crypttab pointing to this device.

A simple way to realize encrypted swap with suspend-to-disk support is by using LVM ontop the encryption layer, so one encrypted partition can contain infinite filesystems (root, swap, home, ...). Follow the instructions on #Encrypting a LVM setup.

The following setup has the disadvantage of having to insert an additional passphrase for the swap partition manually on every boot.

Warning: Do not use this setup with a key file. Please read about the issue reported here

To format the encrypted container for the swap partition, follow steps similar to those described in #Configuring LUKS above and create keyslot for a user-memorizable passphrase.

Open the partition in /dev/mapper:

# cryptsetup open --type luks /dev/<device> swapDevice

Create a swap filesystem inside the mapped partition:

# mkswap /dev/mapper/swapDevice

Now you have to create a hook to open the swap at boot time.

  • Create a hook file containing the open command:
/lib/initcpio/hooks/openswap
 # vim: set ft=sh:
 run_hook ()
 {
     cryptsetup open --type luks /dev/<device> swapDevice
 }

for opening the swap device by typing your password or

/lib/initcpio/hooks/openswap
 # vim: set ft=sh:
 run_hook ()
 {
     mkdir crypto_key_device
     mount /dev/mapper/<root-device> crypto_key_device
     cryptsetup open --type luks --key-file crypto_key_device/<path-to-the-key> /dev/<device> swapDevice
     umount crypto_key_device
 }

for opening the swap device by loading a keyfile from a crypted root device

Note: If swap is on a Solid State Disk (SSD) and Discard/TRIM is desired the option --allow-discards has to get added to the cryptsetup line in the openswap hook above. See Discard/TRIM support for solid state disks (SSD) or SSD for more information on discard. Additionally you have to add the mount option 'discard' to your fstab entry for the swap device.
  • Then create and edit the hook setup file:
/lib/initcpio/install/openswap
# vim: set ft=sh:
build ()
{
   add_runscript
}
help ()
{
cat<<HELPEOF
  This opens the swap encrypted partition /dev/<device> in /dev/mapper/swapDevice
HELPEOF
}
  • Add the hook openswap in the HOOKS array in /etc/mkinitcpio.conf, before filesystem but after encrypt. Do not forget to add the resume hook after openswap.
HOOKS="... encrypt openswap resume filesystems ..."
  • Regenerate the boot image:
# mkinitcpio -p linux
  • Add the mapped partition to /etc/fstab by adding the following line:
/dev/mapper/swapDevice swap swap defaults 0 0
  • Set up your system to resume from /dev/mapper/swapDevice. For example, if you use GRUB with kernel hibernation support, add resume=/dev/mapper/swapDevice to the kernel line in /boot/grub/grub.cfg. A line with encrypted root and swap partitions can look like this:
kernel /vmlinuz-linux cryptdevice=/dev/sda2:rootDevice root=/dev/mapper/rootDevice resume=/dev/mapper/swapDevice ro

To make the parameter persistent on kernel updates, add it to /etc/default/grub.

At boot time, the openswap hook will open the swap partition so the kernel resume may use it. If you use special hooks for resuming from hibernation, make sure they are placed after openswap in the HOOKS array. Please note that because of initrd opening swap, there is no entry for swapDevice in /etc/crypttab needed in this case.

Using a swap file for suspend-to-disk support

A swap file can be used to reserve swap-space within an existing partition and may also be setup inside an encrypted blockdevice's partition. When resuming from a swapfile the resume hook must be supplied with the passphrase to unlock the device where the swap file is located. To create it:

  • Choose a mapped partition (e.g. /dev/mapper/rootDevice) whose mounted filesystem (e.g. /) contains enough free space to create a swapfile with the desired size.
  • Create the swap file (e.g. /swapfile) inside the mounted filesystem of your chosen mapped partition. Be sure to activate it with swapon and also add it to your /etc/fstab file afterward. Note that the swapfile's previous contents remain transparent over reboots.
  • Set up your system to resume from your chosen mapped partition. For example, if you use GRUB with kernel hibernation support, add resume=your chosen mapped partition and resume_offset=see calculation command below to the kernel line in /boot/grub/grub.cfg. A line with encrypted root partition can look like this:
kernel /vmlinuz-linux cryptdevice=/dev/sda2:rootDevice root=/dev/mapper/rootDevice resume=/dev/mapper/rootDevice resume_offset=123456789 ro

The resume_offset of the swap-file points to the start (extent zero) of the file and can be identified like this:

# filefrag -v /swapfile | awk '{if($1==0){print $3}}'
  • Add the resume hook to your etc/mkinitcpio.conf file and rebuild the image afterward:
HOOKS="... encrypt resume ... filesystems ..."
  • If you use a USB keyboard to enter your decryption password, then the keyboard module must appear in front of the encrypt hook, as shown below. Otherwise, you will not be able to boot your computer because you couldn't enter your decryption password to decrypt your Linux root partition! (If you still have this problem after adding keyboard, try usbinput, though this is deprecated.)
HOOKS="... keyboard encrypt ..."

Backup the cryptheader

If the header of your encrypted partition gets destroyed, you will not be able to decrypt your data. It is just as much as a dilemma as forgetting the passphrase or damaging a key-file used to unlock the partition. A damage may occur by your own fault while re-partitioning the disk later or by third-party programs misinterpreting the partition table.

Therefore, having a backup of the headers and storing them on another disk might be a good idea.

Attention: Many people recommend NOT backing up the cryptheader, but even so it's a single point of failure! In short, the problem is that LUKS is not aware of the duplicated cryptheader, which contains the master key which is used to encrypt all files on your partition. Of course this master key is encrypted with your passphrases or keyfiles. But if one of those gets compromised and you want to revoke it you have to do this on all copies of the cryptheader! I.e. if someone has got your cryptheader and one of your keys he can decrypt the master key and access all your data. Of course the same is true for all backups you create of your partions. So you decide if you are one of those paranoids brave enough to go without a backup for the sake of security or not. See also the LUKS FAQ for further details on this.

Backup

Using cryptsetup

Cryptsetups luksHeaderBackup action stores a binary backup of the LUKS header and keyslot area:

# cryptsetup luksHeaderBackup /dev/<device> --header-backup-file /mnt/<backup>/<file>.img

where <device> is the partition containing the LUKS volume.

Note: Using - as header backup file writes to a file named -.
Tip: You can also back up the plaintext header into ramfs and encrypt it in example with gpg before writing to persistent backup storage by executing the following commands.
# mkdir /root/<tmp>/
# mount ramfs /root/<tmp>/ -t ramfs
# cryptsetup luksHeaderBackup /dev/<device> --header-backup-file /root/<tmp>/<file>.img
# gpg2 --recipient <User ID> --encrypt /root/<tmp>/<file>.img 
# cp /root/<tmp>/<file>.img.gpg /mnt/<backup>/
# umount /root/<tmp>
Warning: Tmpfs can swap to harddisk if low on memory so it is not recommended here.

Manually

First you have to find out the payload offset of the crypted partition:

# cryptsetup luksDump /dev/<device> | grep "Payload offset"
 Payload offset:	4040

Second check the sector size of the drive

# fdisk -l /dev/<device> |grep "Sector size"
Sector size (logical/physical): 512 bytes / 512 bytes

Now that you know the values, you can backup the header with a simple dd command:

# dd if=/dev/<device> of=/path/to/<file>.img bs=512 count=4040

and store it safely.

Restore

Be careful before restore: make sure that you chose the right partition (again replace sdaX with the corresponding partition). Restoring the wrong header or restoring to an unencrypted partition will cause data loss.

Using cryptsetup

Or you can use the luksHeaderRestore command:

cryptsetup luksHeaderRestore /dev/sdaX --header-backup-file ./backup.img

Note: All the keyslot areas are overwritten; only active keyslots from the backup file are available after issuing this command.

Manually

Again, you will need to the same values as when backing up:

dd if=./backup.img of=/dev/sdX bs=512 count=4040

Encrypting a system partition

Note: This install example covers the a full system encryption with dmcrypt+LUKS in a simple partition layout. For other partition layouts, e.g. LVM, please see the other examples.

Most of the installation of an encrypted system partition can be carried out normally. However, there are a few areas where it is important to make certain selections. These are marked below.

Prepare hard drive for Arch Install Scripts

This assumes you want to install an encrypted system with the Arch Install Scripts, have created partitions for / (e.g. /dev/sdaX) and /boot (/dev/sdaY) at least, following the Installation Guide and deciding against using LVM. Prior to creating the partitions you have done a preparation of the disk for encryption according to your necessities (the necessary tools are on the installation-ISO).

First check, if the blockdevice mapper dm_mod is loaded with

# lsmod | grep mod 

If one wants to use the default LUKS-cipher algorithm, there is no need to specify one for the luksFormat. You may want to check the defaults used by the cryptsetup version at time of installation and decide yourself. With defaults a dm-crypt/LUKS blockdevice for the crypted root can be created

# cryptsetup -y -v luksFormat /dev/sdaX

opened

# cryptsetup open /dev/sdaX cryptroot

formatted with your desired filesystem

# mkfs -t ext4 /dev/mapper/cryptroot

and mounted

# mount -t ext4 /dev/mapper/cryptroot /mnt

At this point, just before installing the base system, it might be useful to check the mapping works as intended:

# umount /mnt
# cryptsetup close cryptroot

and mount it again to check.

If you created a separate /home partition, the steps have to be adapted and repeated for that. However, this may be inconvenient, because it would result in a separate passphrase to be input as well. An alternative is to use a key-file stored in the system partition to unlock the separate partition via crypttab. How that is done may be derived from the LVM example below. If you leave unpartitioned disk space, the necessary configuration may be done later too along with migration data.

What you do have to setup is a non-encrypted /boot partition, which is needed for a crypted root. For a standard MBR/non-EFI /boot partition that may be achieved by formatting

# mkfs -t ext2 /dev/sdaY

creating a mount-point for installation

# mkdir /mnt/boot

and mounting it

# mount -t ext2 /dev/sdaY /mnt/boot

That is basically what is necessary at this point before installing the base system with the Arch Install Scripts. Take care to install the bootloader to /mnt/boot with the pacstrap script. Additional configuration steps must be followed before booting the installed system.

Configure initramfs

One important point is to add the hooks relevant for your particular install in the correct order to /etc/mkinitcpio.conf. The one you have to add when encrypting the root filesystem is encrypt. A recommended hook for LUKS encrypted blockdevices is shutdown to ensure controlled unmounting during system shutdown. Others needed, e.g. keymap, should be clear from other manual steps you follow during the installation and further details in the following. For detailed information about initramfs configuration and available Hooks refer to Mkinitcpio#HOOKS.

Note: The encrypt hook is only needed if your root partition is a LUKS partition (or a LUKS partition that needs to be mounted before root). The encrypt hook is not needed for any other encrypted partitions (swap, for example). System initialization scripts (/etc/rc.sysinit and /etc/crypttab among others) take care of those.

It is important that the encrypt hook comes before the filesystems hook (in case you are using LVM on LUKS, the order should be: encrypt lvm2 filesystems), so make sure that your HOOKS array looks something like this:

etc/mkinitcpio.conf
HOOKS="(base udev) ... encrypt ... filesystems ..."

If you need support for foreign keymaps for your encryption password, you have to specify the hook keymap as well before encrypt.

If you have a USB keyboard, you will need the keyboard hook. Without it, no USB keyboard will work in early userspace. If you still have this problem after adding keyboard, try usbinput, though this is deprecated.

In the same file, you may want to add to "MODULES" dm_mod and the filesystem types used, e.g: MODULES="dm_mod ext4"

After you are done don't forget:

mkinitcpio -p linux

Kernel parameter configuration of the bootloader

In order to enable booting an encrypted root partition, it is passed to the encrypt hook with kernel parameters to be set up in the bootloader. The main parameter is cryptdevice, with the following syntax:

cryptdevice=<device>:<dmname>
<device>
The path to the raw encrypted device. Usage of Persistent block device naming is advisable.
<dmname>
The name given to the device after decryption, will be available as /dev/mapper/<dmname>. (<dmname> MUST NOT be set to a name already used for LVM partitions!)

So if the encrypted root device in the example is /dev/sda2 and the decrypted one should be mapped to /dev/mapper/cryptroot, the kernel parameter would be:

cryptdevice=/dev/sda2:cryptroot

This will make the system prompt for the passphrase to unlock the root device on a cold boot.

Depending on the setup other parameters are required as well:

cryptdevice=<device>:<dmname> root=<device> resume=<device> cryptkey=<device>:<fstype>:<path>
root=<device>
The device file of the actual (decrypted) root filesystem. If the filesystem is formatted directly on the decrypted device file this will be /dev/mapper/<dmname>. If LVM is used, the device must be addressed like /dev/mapper/<volgroup>-<pvol> or /dev/<volgroup>/<pvol>.
resume=<device>
The device file of the decrypted (swap) filesystem used for suspend2disk.
cryptkey=<device>:<fstype>:<path>
Required for reading a keyfile from a filesystem.

The syntax for the optional cryptkey parameter is:

cryptkey=<device>:<fstype>:<path>
<device>
The raw block device where the key exists.
<fstype>
The filesystem type of <device> (or auto).
<path>
The absolute path of the keyfile within the device.

Examples on configuration for GRUB and Syslinux are available on the respective pages.

Fstab

Further, double-check the genfstab scripts result for your /dev/mapper/cryptroot and other mounts.

Encrypting a loopback filesystem

A loop device enables to map a blockdevice to a file with the standard util-linux tool losetup. The file can then contain a filesystem, which can be used quite like any other filesystem. A lot of users know Truecrypt as a tool to create encrypted containers. Just about the same functionality can be achieved with a loopback filesystem encrypted with LUKS and is shown in the following example.

Preparation and mapping

First, start by creating an encrypted container!

dd if=/dev/urandom of=/bigsecret bs=1M count=10

This will create the file bigsecret with a size of 10 megabytes.

losetup /dev/loop0 /bigsecret

This will create the device node /dev/loop0, so that we can mount/use our container.

Note: If it gives you the error /dev/loop0: No such file or directory, you need to first load the kernel module with modprobe loop. These days (Kernel 3.2) loop devices are created on demand. Ask for a new loop device with losetup -f.
cryptsetup luksFormat /dev/loop0

This will ask you for a password for your new container file.

Note: If you get an error like Command failed: Failed to setup dm-crypt key mapping. Check kernel for support for the aes-cbc-essiv:sha256 cipher spec and verify that /dev/loop0 contains at least 133 sectors, then run modprobe dm-mod.
cryptsetup open --type luks /dev/loop0 secret

The encrypted container is now available through the device file /dev/mapper/secret. Now we are able to create a partition in the container:

mkfs.ext2 /dev/mapper/secret

and mount it...

mkdir /mnt/secret
mount -t ext2 /dev/mapper/secret /mnt/secret

We can now use the container as if it was a normal partition! To unmount the container:

umount /mnt/secret
cryptsetup luksClose secret
losetup -d /dev/loop0 # free the loopdevice.

so, if you want to mount the container again, you just apply the following commands:

losetup /dev/loop0 /bigsecret
cryptsetup open --type luks /dev/loop0 secret
mount -t ext2 /dev/mapper/secret /mnt/secret

Encrypt using a key-file

Let us first generate a 2048 byte random keyfile:

dd if=/dev/urandom of=keyfile bs=1k count=2

We can now format our container using this key

cryptsetup luksFormat /dev/loop0 keyfile

or our partition :

cryptsetup luksFormat /dev/hda2 keyfile

Once formatted, we can now open the LUKS device using the key:

cryptsetup -d keyfile open --type luks /dev/loop0 container

You can now like before format the device /dev/mapper/container with your favorite filesystem and then mount it just as easily.

The keyfile is now the only key to your file. I personally advise encrypting your keyfile using your private GPG key and storing an off-site secured copy of the file.

Resizing the loopback filesystem

First we should unmount the encrypted container:

umount /mnt/secret
cryptsetup luksClose secret
losetup -d /dev/loop0 # free the loopdevice.

After this we need to expand our container file with the size of the data we want to add:

dd if=/dev/urandom bs=1M count=1024 | cat - >> /bigsecret

Be careful to really use TWO >, or you will override your current container!

You could use /dev/zero instead of /dev/urandom to significantly speed up the process, but with /dev/zero your encrypted filesystems will not be as secure. (A better option to create random data quicker than /dev/urandom is frandom [1], available from the AUR). A faster (almost instant) method than dd is truncate , but its use has the same security implications as using /dev/zero. The size passed to truncate is the final size to make the file, so don't use a value less than that of the current file or you will lose data. e.g. to increase a 20G file by 10G: truncate -s 30G filename.

Now we have to map the container to the loop device:

losetup /dev/loop0 /bigsecret
cryptsetup open --type luks /dev/loop0 secret

After this we will resize the encrypted part of the container to the maximum size of the container file:

cryptsetup resize secret

Finally, we can resize the filesystem. Here is an example for ext2/3/4:

e2fsck -f /dev/mapper/secret # Just doing a filesystem check, because it's a bad idea to resize a broken fs
resize2fs /dev/mapper/secret

You can now mount your container again:

mount /dev/mapper/secret /mnt/secret

Encrypting a LVM setup

It is easy to use encryption with LVM. This section describes specific aspects of setting up an encrypted LVM setup and has a mini-howto for users with experiences of using both, cryptsetup and LVM. If you are looking for verbose installation instructions of such a setup, you might want to read Encrypted_LVM first.

If you do not know how to set up LVM, then read Installing with Software RAID or LVM. Using LVM is particularly helpful when a system with multiple partitions is planned. While there are a number of alternatives for unlocking multiple partitions with the same passphrase/key, the default Arch mkinitcpio hooks do not implement non-standard ways. A combination of the encrypt and lvm2 hooks, however, enables to setup a system with numerous logical volumes as partitions while using one passphrase/key to unlock them.

LVM on LUKS

The straight-forward method is to set up LVM on top of the encrypted partition instead of the other way round. Technically the LVM is setup inside one big encrypted blockdevice. Hence, the LVM is not transparent until the blockdevice is unlocked and the underlying volume structure is scanned and mounted during boot.

The most important thing in setting LVM on top of encryption is to configure the initramfs for running both the encrypt hook and the lvm2 hook (and those two before the filesystems hook). In the past, it was necessary to ensure the correct ordering of these hooks in /etc/mkinitcpio.conf but the order no longer matters with the current implementation of lvm2.

LUKS on LVM

To use encryption on top of LVM, the LVM volumes are set up first and then used as the base for the encrypted partitions. This way around a mixture of encrypted and non-encrypted volumes/partitions is possible as well.

For encrypted partitions inside an LVM, the LVM-hook has to run first, before the respective encrypted logical volumes can be unlocked. So for this add the encrypt hook in /etc/mkinitcpio.conf after the lvm2 hook, if you chose to set up encrypted partitions on top of LVM.

Both these options are described in more detail in Encrypted_LVM. The following represents the historic guide from before the verbose page.

LVM and dm-crypt manually (short version)

The following short example creates a LUKS on LVM setup and mixes in the use of a key-file for the /home partition and temporary crypt volumes for /tmp and /swap. The latter is considered desirable from a security perspective, because no potentially sensitive temporary data survives the reboot, when the encryption is re-initialised. If you are experienced with LVM, you will be able to ignore/replace LVM- and other specifics according to your plan. For cryptsetup options, please see above.

Partitioning scheme

/dev/sda1 -> /boot
/dev/sda2 -> LVM

Setting up the encrypted system

cryptsetup -d /dev/random -c aes-xts-plain -s 512 create lvm /dev/sda2
dd if=/dev/urandom of=/dev/mapper/lvm
cryptsetup remove lvm
lvm pvcreate /dev/sda2
lvm vgcreate lvm /dev/sda2
lvm lvcreate -L 10G -n root lvm
lvm lvcreate -L 500M -n swap lvm
lvm lvcreate -L 500M -n tmp lvm
lvm lvcreate -l 100%FREE -n home lvm
cryptsetup luksFormat -c aes-xts-plain -s 512 /dev/lvm/root
cryptsetup open --type luks /dev/lvm/root root
mkreiserfs /dev/mapper/root
mount /dev/mapper/root /mnt
dd if=/dev/zero of=/dev/sda1 bs=1M
mkreiserfs /dev/sda1
mkdir /mnt/boot
mount /dev/sda1 /mnt/boot
mkdir -p -m 700 /mnt/etc/luks-keys
dd if=/dev/random of=/mnt/etc/luks-keys/home bs=1 count=256

Install Arch Linux

Now after setup of the encrypted LVM partitioning, it would be time to install: Arch Install Scripts.

Configuration

/etc/mkinitcpio.conf

Put lvm2 and encrypt (in that order) before filesystems in the HOOKS array. Again, note that this is setting encryption on top of LVM.

If you want install the system on a usb stick, you need to put usb just after udev.

Boot options for LUKS-LVM

For the above example, change the kernel options for the root-device auto-configured in the bootloader installation from root=/dev/hda3 to

cryptdevice=/dev/lvm/root:root root=/dev/mapper/root

More general, the kernel command line for LUKS <-> LVM is constructed like this:

root=/dev/mapper/<volume-group>-<logical-volume> cryptdevice=/dev/<luks-part>:<volume-group>

For example:

root=/dev/mapper/vg-arch cryptdevice=/dev/sda4:vg

Or like this:

cryptdevice=/dev/<volume-group>/<logical-volume>:root root=/dev/mapper/root

If you want install the system on a usb stick, you need to add lvmdelay=/dev/mapper/lvm-root

Filesystem mounts system
/etc/fstab
 /dev/mapper/root        /       reiserfs        defaults        0       1
 /dev/sda1               /boot   reiserfs        defaults        0       2
 /dev/mapper/tmp         /tmp    tmpfs           defaults        0       0
 /dev/mapper/swap        none    swap            sw              0       0
/etc/crypttab
 swap	/dev/lvm/swap	SWAP		-c aes-xts-plain -h whirlpool -s 512
 tmp	/dev/lvm/tmp	/dev/urandom	-c aes-xts-plain -s 512
Encrypting /home after reboot
cryptsetup luksFormat -c aes-xts-plain -s 512 /dev/lvm/home /etc/luks-keys/home
cryptsetup open --type luks -d /etc/luks-keys/home /dev/lvm/home home
mkreiserfs /dev/mapper/home
mount /dev/mapper/home /home
Filesystem mounts home
/etc/crypttab
 home	/dev/lvm/home   /etc/luks-keys/home
/etc/fstab
 /dev/mapper/home        /home   reiserfs        defaults        0       0

Specialties

Using GPG or OpenSSL Encrypted Keyfiles

The following forum posts give instructions to use two factor authentication, gpg or openssl encrypted keyfiles, instead of a plaintext keyfile described earlier in this wiki article System Encryption using LUKS with GPG encrypted keys:

Note that:

  • You can follow the above instructions with only two primary partitions one boot partition

(required because of LVM), and one primary LVM partition. Within the LVM partition you can have as many partitions as you need, but most importantly it should contain at least root, swap, and home logical volume partitions. This has the added benefit of having only one keyfile for all your partitions, and having the ability to hibernate your computer (suspend to disk) where the swap partition is encrypted. If you decide to do so your hooks in /etc/mkinitcpio.conf should look like HOOKS=" ... usb usbinput (etwo or ssldec) encrypt(if using openssl) lvm2 resume ... " and you should add "resume=/dev/mapper/<VolumeGroupName>-<LVNameOfSwap>" to your kernel parameters.

  • If you need to temporarily store the unecrypted keyfile somewhere, do not store them on an unencrypted disk. Even better make sure to store them to RAM such as /dev/shm.
  • If you want to use a GPG encrypted keyfile, you need to use a statically compiled GnuPG version 1.4 or you could edit the hooks and use this AUR package gnupg1
  • It is possible that an update to OpenSSL could break the custom ssldec mentioned in the second forum post.

Remote unlocking of the root (or other) partition

If you want to be able to reboot a fully LUKS-encrypted system remotely, or start it with a Wake-on-LAN service, you will need a way to enter a passphrase for the root partition/volume at startup. This can be achieved by running the net hook along with an SSH server in initrd. Install the dropbear_initrd_encryptAUR package from the AUR and follow the post-installation instructions. Replace the encrypt hook with dropbear encryptssh in /etc/mkinitcpio.conf. Put the net hook early in the HOOKS array if your DHCP server takes a long time to lease IP addresses.

If you would simply like a nice solution to mount other encrypted partitions (such as /home)remotely, you may want to look at this forum thread.

Modifying the encrypt hook for a non-root partition

Maybe you have a requirement for using the encrypt hook on a non-root partition. Arch does not support this out of the box, however, you can easily change the cryptdev and cryptname values in /lib/initcpio/hooks/encrypt (the first one to your /dev/sd* partition, the second to the name you want to attribute). That should be enough.

The big advantage is you can have everything automated, while setting up /etc/crypttab with an external key file (i.e. the keyfile is not on any internal hard drive partition) can be a pain - you need to make sure the USB/FireWire/... device gets mounted before the encrypted partition, which means you have to change the order of /etc/fstab (at least).

Of course, if the cryptsetup package gets upgraded, you will have to change this script again. Unlike /etc/crypttab, only one partition is supported, but with some further hacking one should be able to have multiple partitions unlocked.

If you want to do this on a software RAID partition, there is one more thing you need to do. Just setting the /dev/mdX device in /lib/initcpio/hooks/encrypt is not enough; the encrypt hook will fail to find the key for some reason, and not prompt for a passphrase either. It looks like the RAID devices are not brought up until after the encrypt hook is run. You can solve this by putting the RAID array in /boot/grub/menu.lst, like

kernel /boot/vmlinuz-linux md=1,/dev/hda5,/dev/hdb5

If you set up your root partition as a RAID, you will notice the similarities with that setup ;-). GRUB can handle multiple array definitions just fine:

kernel /boot/vmlinuz-linux root=/dev/md0 ro md=0,/dev/sda1,/dev/sdb1 md=1,/dev/sda5,/dev/sdb5,/dev/sdc5

Securing the unencrypted boot partition

Referring to an article from the ct-magazine (Issue 3/12, page 146, 01.16.2012 http://www.heise.de/ct/inhalt/2012/03/6/) the following script checks files under /boot for changes of SHA-1 hash, inode and occupied blocks on the hard drive. It also checks the MBR. The script cannot prevent certain type of attacks, but a lot are made harder. No configuration of the script itself is stored in unencrypted /boot. With a locked/powered-off crypted system this makes it infeasible for an attacker to recognize that an automatic checksum comparison of the partition is done upon boot.

The script with installation instructions is available here: ftp://ftp.heise.de/pub/ct/listings/1203-146.zip (Author: Juergen Schmidt, ju at heisec.de; License: GPLv2). There is also an AUR package: chkbootAUR

After installation:

  • For classical sysvinit: add /usr/local/bin/chkboot.sh & to your /etc/rc.local
  • For systemd: add a service file and enable the service: systemd. The service file might look like:
[Unit]
Description=Check that boot is what we want
Requires=basic.target
After=basic.target

[Service]
Type=oneshot
ExecStart=/usr/local/bin/chkboot.sh

[Install]
WantedBy=multi-user.target

There is a small caveat for systemd: At the time of writing the original chkboot.sh script provided contains an empty space at the beginning of #!/bin/bash which has to be removed for the service to start successfully.

As /usr/local/bin/chkboot_user.sh need to be excuted after login, add it to the autostart (e.g. under KDE -> System Settings -> Startup and Shutdown -> Autostart; Gnome3: gnome-session-properties).

With Arch Linux changes to /boot are pretty frequent, for example by new kernels rolling-in. Therefore it may be helpful to use the scripts with every full system update. One way to do so:

#!/bin/bash
#
# Note: Insert your <user>  and execute it with sudo for pacman & chkboot to work automagically
#
echo "Pacman update [1] Quickcheck before updating" & 
sudo -u <user> /usr/local/bin/chkboot_user.sh		# insert your logged on <user> 
/usr/local/bin/chkboot.sh
sync							# sync disks with any results 
sudo -u <user> /usr/local/bin/chkboot_user.sh		# insert your logged on <user> 
echo "Pacman update [2] Syncing repos for pacman" 
pacman -Syu
/usr/local/bin/chkboot.sh
sync	
sudo -u <user> /usr/local/bin/chkboot_user.sh		# insert your logged on <user>
echo "Pacman update [3] All done, let's roll on ..."

Alternatively to above scripts, a hash check can be set up with AIDE which can be customized via a very flexible configuration file.

While one of these methods should serve the purpose for most users, they do not address all security problems associated with the unencrypted /boot. One approach which endeavours to provide a fully authenticated boot chain was published with POTTS as an academic thesis to implement the STARK authentication framework.

The POTTS proof-of-concept uses Arch Linux as a base distribution and implements a system boot chain with

  • POTTS - a bootmenu for a one-time authentication message prompt
  • TrustedGrub - a grub-legacy implementation which authenticates the kernel and initramfs against TPM chip registers
  • TRESOR - a kernel patch which implements AES but keeps the master-key not in RAM but in CPU registers during runtime.

As part of the thesis installation instructions based on Arch Linux (iso 2013-01) have been published. If you want to try it, be aware these tools are not in standard repositories and the solution will be time consuming to maintain.

Automount user homes on login

See Pam mount.

Resources

  • cryptsetup FAQ - The main and foremost help resource, directly from the developers.
  • FreeOTFE - Supports unlocking LUKS encrypted volumes in Microsoft Windows.