8 6. Backup and Data Recovery
9 7. Interoperability with other Disk Encryption Tools
10 8. Issues with Specific Versions of cryptsetup
11 9. References and Further Reading
20 This is the FAQ (Frequently Asked Questions) for cryptsetup. It
21 covers Linux disk encryption with plain dm-crypt (one passphrase,
22 no management, no metadata on disk) and LUKS (multiple user keys
23 with one master key, anti-forensic features, metadata block at
24 start of device, ...). The latest version of this FAQ should
25 usually be available at
26 http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions
31 ATTENTION: If you are going to read just one thing, make it the
32 section on Backup and Data Recovery. By far the most questions on
33 the cryptsetup mailing list are from people that managed to damage
34 the start of their LUKS partitions, i.e. the LUKS header. In
35 most cases, there is nothing that can be done to help these poor
36 souls recover their data. Make sure you understand the problem and
37 limitations imposed by the LUKS security model BEFORE you face
38 such a disaster! In particular, make sure you have a current header
39 backup before doing any potentially dangerous operations.
41 SSDs/FLASH DRIVES: SSDs and Flash are different. Currently it is
42 unclear how to get LUKS or plain dm-crypt to run on them with the
43 full set of security features intact. This may or may not be a
44 problem, depending on the attacher model. See Section 5.19.
46 BACKUP: Yes, encrypted disks die, just as normal ones do. A full
47 backup is mandatory, see Section "6. Backup and Data Recovery" on
48 options for doing encrypted backup.
50 CLONING/IMAGING: If you clone or image a LUKS container, you make a
51 copy of the LUKS header and the master key will stay the same!
52 That means that if you distribute an image to several machines, the
53 same master key will be used on all of them, regardless of whether
54 you change the passphrases. Do NOT do this! If you do, a root-user
55 on any of the machines with a mapped (decrypted) container or a
56 passphrase on that machine can decrypt all other copies, breaking
57 security. See also Item 6.15.
59 DISTRIBUTION INSTALLERS: Some distribution installers offer to
60 create LUKS containers in a way that can be mistaken as activation
61 of an existing container. Creating a new LUKS container on top of
62 an existing one leads to permanent, complete and irreversible data
63 loss. It is strongly recommended to only use distribution
64 installers after a complete backup of all LUKS containers has been
67 UBUNTU INSTALLER: In particular the Ubuntu installer seems to be
68 quite willing to kill LUKS containers in several different ways.
69 Those responsible at Ubuntu seem not to care very much (it is very
70 easy to recognize a LUKS container), so treat the process of
71 installing Ubuntu as a severe hazard to any LUKS container you may
74 NO WARNING ON NON-INTERACTIVE FORMAT: If you feed cryptsetup from
75 STDIN (e.g. via GnuPG) on LUKS format, it does not give you the
76 warning that you are about to format (and e.g. will lose any
77 pre-existing LUKS container on the target), as it assumes it is
78 used from a script. In this scenario, the responsibility for
79 warning the user and possibly checking for an existing LUKS header
80 is shifted to the script. This is a more general form of the
83 LUKS PASSPHRASE IS NOT THE MASTER KEY: The LUKS passphrase is not
84 used in deriving the master key. It is used in decrypting a master
85 key that is randomly selected on header creation. This means that
86 if you create a new LUKS header on top of an old one with
87 exactly the same parameters and exactly the same passphrase as the
88 old one, it will still have a different master key and your data
89 will be permanently lost.
91 PASSPHRASE CHARACTER SET: Some people have had difficulties with
92 this when upgrading distributions. It is highly advisable to only
93 use the 95 printable characters from the first 128 characters of
94 the ASCII table, as they will always have the same binary
95 representation. Other characters may have different encoding
96 depending on system configuration and your passphrase will not
97 work with a different encoding. A table of the standardized first
98 128 ASCII characters can, e.g. be found on
99 http://en.wikipedia.org/wiki/ASCII
102 * 1.3 System specific warnings
104 - Ubuntu as of 4/2011: It seems the installer offers to create
105 LUKS partitions in a way that several people mistook for an offer
106 to activate their existing LUKS partition. The installer gives no
107 or an inadequate warning and will destroy your old LUKS header,
108 causing permanent data loss. See also the section on Backup and
111 This issue has been acknowledged by the Ubuntu dev team, see here:
112 http://launchpad.net/bugs/420080
114 Update 4/2013: I am still unsure whether this has been fixed by
115 now, best be careful. They also seem to have added even more LUKS
116 killer functionality to the Ubuntu installer. I can only strongly
117 recommended to not install Ubuntu on a system with existing LUKS
118 containers without complete backups.
121 * 1.4 My LUKS-device is broken! Help!
123 First: Do not panic! In many cases the data is still recoverable.
124 Do not do anything hasty! Steps:
126 - Take some deep breaths. Maybe add some relaxing music. This may
127 sound funny, but I am completely serious. Often, critical damage is
128 done only after the initial problem.
130 - Do not reboot. The keys mays still be in the kernel if the device
133 - Make sure others do not reboot the system.
135 - Do not write to your disk without a clear understanding why this
136 will not make matters worse. Do a sector-level backup before any
137 writes. Often you do not need to write at all to get enough access
138 to make a backup of the data.
142 - Read section 6 of this FAQ.
144 - Ask on the mailing-list if you need more help.
147 * 1.5 Who wrote this?
149 Current FAQ maintainer is Arno Wagner <arno@wagner.name>. If you
150 want to send me encrypted email, my current PGP key is DSA key
151 CB5D9718, fingerprint 12D6 C03B 1B30 33BB 13CF B774 E35C 5FA1 CB5D
154 Other contributors are listed at the end. If you want to contribute,
155 send your article, including a descriptive headline, to the
156 maintainer, or the dm-crypt mailing list with something like "FAQ
157 ..." in the subject. You can also send more raw information and
158 have me write the section. Please note that by contributing to this
159 FAQ, you accept the license described below.
161 This work is under the "Attribution-Share Alike 3.0 Unported"
162 license, which means distribution is unlimited, you may create
163 derived works, but attributions to original authors and this
164 license statement must be retained and the derived work must be
165 under the same license. See
166 http://creativecommons.org/licenses/by-sa/3.0/ for more details of
169 Side note: I did text license research some time ago and I think
170 this license is best suited for the purpose at hand and creates the
174 * 1.5 Where is the project website?
176 There is the project website at http://code.google.com/p/cryptsetup/
177 Please do not post questions there, nobody will read them. Use
178 the mailing-list instead.
181 * 1.6 Is there a mailing-list?
183 Instructions on how to subscribe to the mailing-list are at on the
184 project website. People are generally helpful and friendly on the
187 The question of how to unsubscribe from the list does crop up
188 sometimes. For this you need your list management URL, which is
189 sent to you initially and once at the start of each month. Go to
190 the URL mentioned in the email and select "unsubscribe". This page
191 also allows you to request a password reminder.
193 Alternatively, you can send an Email to dm-crypt-request@saout.de
194 with just the word "help" in the subject or message body. Make sure
195 to send it from your list address.
197 The mailing list archive is here:
198 http://dir.gmane.org/gmane.linux.kernel.device-mapper.dm-crypt
201 * 1.7 Unsubscribe from the mailing-list
203 Send mail to dm-crypt-unsubscribe@saout.de from the subscribed
204 account. You will get an email with instructions.
206 Basically, you just have to respond to it unmodified to get
207 unsubscribed. The listserver admin functions are not very fast. It
208 can take 15 minutes or longer for a reply to arrive (I suspect
209 greylisting is in use), so be patient.
211 Also note that nobody on the list can unsubscribe you, sending
212 demands to be unsubscribed to the list just annoys people that are
213 entirely blameless for you being subscribed.
215 If you are subscribed, a subscription confirmation email was sent
216 to your email account and it had to be answered before the
217 subscription went active. The confirmation emails from the
218 listserver have subjects like these (with other numbers):
220 Subject: confirm 9964cf10.....
222 and are sent from dm-crypt-request@saout.de. You should check
223 whether you have anything like it in your sent email folder. If
224 you find nothing and are sure you did not confirm, then you should
225 look into a possible compromise of your email account.
231 * 2.1 LUKS Container Setup mini-HOWTO
233 This item tries to give you a very brief list of all the steps you
234 should go though when creating a new LUKS encrypted container, i.e.
235 encrypted disk, partition or loop-file.
237 01) All data will be lost, if there is data on the target, make a
240 02) Make very sure you have the right target disk, partition or
243 03) If the target was in use previously, it is a good idea to
244 wipe it before creating the LUKS container in order to remove any
245 trace of old file systems and data. For example, some users have
246 managed to run e2fsck on a partition containing a LUKS container,
247 possibly because of residual ext2 superblocks from an earlier use.
248 This can do arbitrary damage up to complete and permanent loss of
249 all data in the LUKS container.
251 To just quickly wipe file systems (old data may remain), use
253 wipefs -a <target device>
255 To wipe file system and data, use something like
257 cat /dev/zero > <target device>
259 This can take a while. To get a progress indicator, you can use
260 the tool dd_rescue (->google) instead or use my stream meter "wcs"
261 (source here: http://www.tansi.org/tools/index.html) in the
264 cat /dev/zero | wcs > <target device>
266 Be very sure you have the right target, all data will be lost!
268 Note that automatic wiping is on the TODO list for cryptsetup, so
269 at some time in the future this will become unnecessary.
271 04) Create the LUKS container:
272 cryptsetup luksFormat <target device>
274 Just follow the on-screen instructions.
276 Note: Passphrase iteration is determined by cryptsetup depending on
277 CPU power. On a slow device, this may be lower than you want. I
278 recently benchmarked this on a Raspberry Pi and it came out at
279 about 1/15 of the iteration count for a typical PC. If security is
280 paramount, you may want to increase the time spent in iteration, at
281 the cost of a slower unlock later. For the Raspberry Pi, using
283 cryptsetup luksFormat -i 15000 <target device>
285 gives you an iteration count and security level equal to an average
286 PC for passphrase iteration and master-key iteration. If in doubt,
287 check the iteration counts with
289 cryptsetup luksDump <target device>
291 and adjust the iteration count accordingly by creating the container
292 again with a different iteration time (the number after '-i' is the
293 iteration time in milicesonds) until your requirements are met.
295 05) Map the container. Here it will be mapped to /dev/mapper/c1:
296 cryptsetup luksOpen <target device> c1
298 06) (Optionally) wipe the container (make sure you have the right target!):
299 cat /dev/zero > /dev/mapper/c1
301 Note that this creates a small information leak, as an attacker can
302 determine whether a 512 byte block is zero if the attacker has
303 access to the encrypted container multiple times. Typically a
304 competent attacker that has access multiple times can install a
305 passphrase sniffer anyways, so this leakage is not very
306 significant. For getting a progress indicator, see step 03.
308 Note that at some time in the future, cryptsetup will do this for
309 you, but currently it is a TODO list item.
311 07) Create a file system in the mapped container, for example an
312 ext3 file system (any other file system is possible):
314 mke2fs -j /dev/mapper/c1
316 08) Mount your encrypted file system, here on /mnt:
317 mount /dev/mapper/c1 /mnt
319 Done. You can now use the encrypted file system to store data. Be
320 sure to read though the rest of the FAQ, these are just the very
321 basics. In particular, there are a number of mistakes that are
322 easy to make, but will compromise your security.
325 * 2.2 How do I set up encrypted swap?
327 As things that are confidential can end up in swap (keys,
328 passphrases, etc. are usually protected against being swapped to
329 disk, but other things may not be), it may be advisable to do
330 something about the issue. One option is to run without swap, which
331 generally works well in a desktop-context. It may cause problems
332 in a server-setting or under special circumstances. The solution to
333 that is to encrypt swap with a random key at boot-time.
335 NOTE: This is for Debian, and should work for Debian-derived
336 distributions. For others you may have to write your own startup
337 script or use other mechanisms.
339 01) Add the swap partition to /etc/crypttab. A line like the following
342 swap /dev/<partition> /dev/urandom swap,noearly
344 Warning: While Debian refuses to overwrite partitions with a
345 filesystem or RAID signature on it, if your disk IDs may change
346 (adding or removing disks, failure of disk during boot, etc.), you
347 may want to take additional precautions. This is not a concern if
348 you have only one disk. One possibility is to make sure the
349 partition number is not present on additional disks or also swap
350 there. Another is to encapsulate the swap partition (by making it a
351 1-disk RAID1 or by using LVM), so that it gets a persistent
352 identifier. Specifying it directly by UUID does not work,
353 unfortunately, as the UUID is part of the swap signature and that
354 is not visible from the outside due to the encryption and in
355 addition changes on each reboot with this setup.
357 Note: Use /dev/random if you are paranoid or in a potential
358 low-entropy situation (embedded system, etc.). This may cause the
359 operation to take a long time during boot. If you are in a "no
360 entropy" situation, you cannot encrypt swap securely. In this
361 situation you should find some entropy, also because nothing else
362 using crypto will be secure, like ssh, ssl or GnuPG.
364 Note: The "noearly" option makes sure things like LVM, RAID, etc.
365 are running. As swap is non-critical for boot, it is fine to start
368 02) Add the swap partition to /etc/fstab. A line like the following
371 /dev/mapper/swap none swap sw 0 0
373 That is it. Reboot or start it manually to activate encrypted swap.
374 Manual start would look like this:
376 /etc/init.d/crypdisks start
377 swapon /dev/mapper/swap
380 * 2.3 What is the difference between "plain" and LUKS format?
382 First, unless you happen to understand the cryptographic background
383 well, you should use LUKS. It does protect the user from a lot of
384 common mistakes. Plain dm-crypt is for experts.
386 Plain format is just that: It has no metadata on disk, reads all
387 parameters from the commandline (or the defaults), derives a
388 master-key from the passphrase and then uses that to de-/encrypt
389 the sectors of the device, with a direct 1:1 mapping between
390 encrypted and decrypted sectors.
392 Primary advantage is high resilience to damage, as one damaged
393 encrypted sector results in exactly one damaged decrypted sector.
394 Also, it is not readily apparent that there even is encrypted data
395 on the device, as an overwrite with crypto-grade randomness (e.g.
396 from /dev/urandom) looks exactly the same on disk.
398 Side-note: That has limited value against the authorities. In
399 civilized countries, they cannot force you to give up a crypto-key
400 anyways. In quite a few countries around the world, they can force
401 you to give up the keys (using imprisonment or worse to pressure
402 you, sometimes without due process), and in the worst case, they
403 only need a nebulous "suspicion" about the presence of encrypted
404 data. Sometimes this applies to everybody, sometimes only when you
405 are suspected of having "illicit data" (definition subject to
406 change) and sometimes specifically when crossing a border. Note
407 that this is going on in countries like the US and the UK, to
408 different degrees and sometimes with courts restricting what the
409 authorities can actually demand.
411 My advice is to either be ready to give up the keys or to not have
412 encrypted data when traveling to those countries, especially when
413 crossing the borders. The latter also means not having any
414 high-entropy (random) data areas on your disk, unless you can
415 explain them and demonstrate that explanation. Hence doing a
416 zero-wipe of all free space, including unused space, may be a good
419 Disadvantages are that you do not have all the nice features that
420 the LUKS metadata offers, like multiple passphrases that can be
421 changed, the cipher being stored in the metadata, anti-forensic
422 properties like key-slot diffusion and salts, etc..
424 LUKS format uses a metadata header and 8 key-slot areas that are
425 being placed at the beginning of the disk, see below under "What
426 does the LUKS on-disk format looks like?". The passphrases are used
427 to decrypt a single master key that is stored in the anti-forensic
430 Advantages are a higher usability, automatic configuration of
431 non-default crypto parameters, defenses against low-entropy
432 passphrases like salting and iterated PBKDF2 passphrase hashing,
433 the ability to change passphrases, and others.
435 Disadvantages are that it is readily obvious there is encrypted
436 data on disk (but see side note above) and that damage to the
437 header or key-slots usually results in permanent data-loss. See
438 below under "6. Backup and Data Recovery" on how to reduce that
439 risk. Also the sector numbers get shifted by the length of the
440 header and key-slots and there is a loss of that size in capacity
441 (1MB+4096B for defaults and 2MB for the most commonly used
442 non-default XTS mode).
445 * 2.4 Can I encrypt an already existing, non-empty partition to use
448 There is no converter, and it is not really needed. The way to do
449 this is to make a backup of the device in question, securely wipe
450 the device (as LUKS device initialization does not clear away old
451 data), do a luksFormat, optionally overwrite the encrypted device,
452 create a new filesystem and restore your backup on the now
453 encrypted device. Also refer to sections "Security Aspects" and
454 "Backup and Data Recovery".
456 For backup, plain GNU tar works well and backs up anything likely
457 to be in a filesystem.
460 * 2.5 How do I use LUKS with a loop-device?
462 This can be very handy for experiments. Setup is just the same as
463 with any block device. If you want, for example, to use a 100MiB
464 file as LUKS container, do something like this:
466 head -c 100M /dev/zero > luksfile # create empty file
467 losetup /dev/loop0 luksfile # map luksfile to /dev/loop0
468 cryptsetup luksFormat /dev/loop0 # create LUKS on loop device
470 Afterwards just use /dev/loop0 as a you would use a LUKS partition.
471 To unmap the file when done, use "losetup -d /dev/loop0".
474 * 2.6 When I add a new key-slot to LUKS, it asks for a passphrase but
475 then complains about there not being a key-slot with that
478 That is as intended. You are asked a passphrase of an existing
479 key-slot first, before you can enter the passphrase for the new
480 key-slot. Otherwise you could break the encryption by just adding a
481 new key-slot. This way, you have to know the passphrase of one of
482 the already configured key-slots in order to be able to configure a
486 * 2.7 Encryption on top of RAID or the other way round?
488 Unless you have special needs, place encryption between RAID and
489 filesystem, i.e. encryption on top of RAID. You can do it the other
490 way round, but you have to be aware that you then need to give the
491 passphrase for each individual disk and RAID autodetection will
492 not work anymore. Therefore it is better to encrypt the RAID
493 device, e.g. /dev/dm0 .
495 This means that the typical layering looks like this:
507 The big advantage is that you can manage the RAID container just
508 like any RAID container, it does not care that what is in it is
512 * 2.8 How do I read a dm-crypt key from file?
514 Use the --key-file option, like this:
516 cryptsetup create --key-file keyfile e1 /dev/loop0
518 This will read the binary key from file, i.e. no hashing or
519 transformation will be applied to the keyfile before its bits are
520 used as key. Extra bits (beyond the length of the key) at the end
521 are ignored. Note that if you read from STDIN, the data will still
522 be hashed, just as a key read interactively from the terminal. See
523 the man-page sections "NOTES ON PASSPHRASE PROCESSING..." for more
527 * 2.9 How do I read a LUKS slot key from file?
529 What you really do here is to read a passphrase from file, just as
530 you would with manual entry of a passphrase for a key-slot. You can
531 add a new passphrase to a free key-slot, set the passphrase of an
532 specific key-slot or put an already configured passphrase into a
533 file. In the last case make sure no trailing newline (0x0a) is
534 contained in the key file, or the passphrase will not work because
535 the whole file is used as input.
537 To add a new passphrase to a free key slot from file, use something
540 cryptsetup luksAddKey /dev/loop0 keyfile
542 To add a new passphrase to a specific key-slot, use something like
545 cryptsetup luksAddKey --key-slot 7 /dev/loop0 keyfile
547 To supply a key from file to any LUKS command, use the --key-file
548 option, e.g. like this:
550 cryptsetup luksOpen --key-file keyfile /dev/loop0 e1
553 * 2.10 How do I read the LUKS master key from file?
555 The question you should ask yourself first is why you would want to
556 do this. The only legitimate reason I can think of is if you want
557 to have two LUKS devices with the same master key. Even then, I
558 think it would be preferable to just use key-slots with the same
559 passphrase, or to use plain dm-crypt instead. If you really have a
560 good reason, please tell me. If I am convinced, I will add how to
564 * 2.11 What are the security requirements for a key read from file?
566 A file-stored key or passphrase has the same security requirements
567 as one entered interactively, however you can use random bytes and
568 thereby use bytes you cannot type on the keyboard. You can use any
569 file you like as key file, for example a plain text file with a
570 human readable passphrase. To generate a file with random bytes,
571 use something like this:
573 head -c 256 /dev/random > keyfile
576 * 2.12 If I map a journaled file system using dm-crypt/LUKS, does it
577 still provide its usual transactional guarantees?
579 Yes, it does, unless a very old kernel is used. The required flags
580 come from the filesystem layer and are processed and passed onwards
581 by dm-crypt. A bit more information on the process by which
582 transactional guarantees are implemented can be found here:
584 http://lwn.net/Articles/400541/
586 Please note that these "guarantees" are weaker than they appear to
587 be. One problem is that quite a few disks lie to the OS about
588 having flushed their buffers. Some other things can go wrong as
589 well. The filesystem developers are aware of these problems and
590 typically can make it work anyways. That said, dm-crypt/LUKS will
591 not make things worse.
593 One specific problem you can run into though is that you can get
594 short freezes and other slowdowns due to the encryption layer.
595 Encryption takes time and forced flushes will block for that time.
596 For example, I did run into frequent small freezes (1-2 sec) when
597 putting a vmware image on ext3 over dm-crypt. When I went back to
598 ext2, the problem went away. This seems to have gotten better with
599 kernel 2.6.36 and the reworking of filesystem flush locking
600 mechanism (less blocking of CPU activity during flushes). It
601 should improve further and eventually the problem should go away.
604 * 2.13 Can I use LUKS or cryptsetup with a more secure (external)
605 medium for key storage, e.g. TPM or a smartcard?
607 Yes, see the answers on using a file-supplied key. You do have to
608 write the glue-logic yourself though. Basically you can have
609 cryptsetup read the key from STDIN and write it there with your
610 own tool that in turn gets the key from the more secure key
613 For TPM support, you may want to have a look at tpm-luks at
614 https://github.com/shpedoikal/tpm-luks. Note that tpm-luks is not
615 related to the cryptsetup project.
618 * 2.14 Can I resize a dm-crypt or LUKS partition?
620 Yes, you can, as neither dm-crypt nor LUKS stores partition size.
621 Whether you should is a different question. Personally I recommend
622 backup, recreation of the encrypted partition with new size,
623 recreation of the filesystem and restore. This gets around the
624 tricky business of resizing the filesystem. Resizing a dm-crypt or
625 LUKS container does not resize the filesystem in it. The backup is
626 really non-optional here, as a lot can go wrong, resulting in
627 partial or complete data loss. Using something like gparted to
628 resize an encrypted partition is slow, but typically works. This
629 will not change the size of the filesystem hidden under the
632 You also need to be aware of size-based limitations. The one
633 currently relevant is that aes-xts-plain should not be used for
634 encrypted container sizes larger than 2TiB. Use aes-xts-plain64
638 * 2.15 How do I Benchmark the Ciphers, Hashes and Modes?
640 Since version 1.60 cryptsetup supports the "benchmark" command.
645 It will output first iterations/second for the key-derivation
646 function PBKDF2 parameterized with different hash-functions, and
647 then the raw encryption speed of ciphers with different modes and
648 key-sizes. You can get more than the default benchmarks, see the
649 man-page for the relevant parameters. Note that XTS mode takes two
650 keys, hence the listed key sizes are double that for other modes
651 and half of it is the cipher key, the other half is the XTS key.
654 * 2.16 How do I Verify I have an Authentic cryptsetup Source Package?
656 Current maintainer is Milan Broz and he signs the release packages
657 with his PGP key. The key he currently uses is the "RSA key ID
658 D93E98FC", fingerprint 2A29 1824 3FDE 4664 8D06 86F9 D9B0 577B
659 D93E 98FC. While I have every confidence this really is his key and
660 that he is who he claims to be, don't depend on it if your life is
661 at stake. For that matter, if your life is at stake, don't depend
662 on me being who I claim to be either.
664 That said, as cryptsetup is under good version control, a malicious
665 change should be noticed sooner or later, but it may take a while.
666 Also, the attacker model makes compromising the sources in a
667 non-obvious way pretty hard. Sure, you could put the master-key
668 somewhere on disk, but that is rather obvious as soon as somebody
669 looks as there would be data in an empty LUKS container in a place
670 it should not be. Doing this in a more nefarious way, for example
671 hiding the master-key in the salts, would need a look at the
672 sources to be discovered, but I think that somebody would find that
673 sooner or later as well.
675 That said, this discussion is really a lot more complicated and
676 longer as an FAQ can sustain. If in doubt, ask on the mailing list.
682 * 3.1 My dm-crypt/LUKS mapping does not work! What general steps are
683 there to investigate the problem?
685 If you get a specific error message, investigate what it claims
686 first. If not, you may want to check the following things.
688 - Check that "/dev", including "/dev/mapper/control" is there. If it
689 is missing, you may have a problem with the "/dev" tree itself or
690 you may have broken udev rules.
692 - Check that you have the device mapper and the crypt target in your
693 kernel. The output of "dmsetup targets" should list a "crypt"
694 target. If it is not there or the command fails, add device mapper
695 and crypt-target to the kernel.
697 - Check that the hash-functions and ciphers you want to use are in
698 the kernel. The output of "cat /proc/crypto" needs to list them.
701 * 3.2 My dm-crypt mapping suddenly stopped when upgrading cryptsetup.
703 The default cipher, hash or mode may have changed (the mode changed
704 from 1.0.x to 1.1.x). See under "Issues With Specific Versions of
708 * 3.3 When I call cryptsetup from cron/CGI, I get errors about
711 If you get errors about unknown parameters or the like that are not
712 present when cryptsetup is called from the shell, make sure you
713 have no older version of cryptsetup on your system that then gets
714 called by cron/CGI. For example some distributions install
715 cryptsetup into /usr/sbin, while a manual install could go to
716 /usr/local/sbin. As a debugging aid, call "cryptsetup --version"
717 from cron/CGI or the non-shell mechanism to be sure the right
721 * 3.4 Unlocking a LUKS device takes very long. Why?
723 The iteration time for a key-slot (see Section 5 for an explanation
724 what iteration does) is calculated when setting a passphrase. By
725 default it is 1 second on the machine where the passphrase is set.
726 If you set a passphrase on a fast machine and then unlock it on a
727 slow machine, the unlocking time can be much longer. Also take into
728 account that up to 8 key-slots have to be tried in order to find the
731 If this is problem, you can add another key-slot using the slow
732 machine with the same passphrase and then remove the old key-slot.
733 The new key-slot will have an iteration count adjusted to 1 second
734 on the slow machine. Use luksKeyAdd and then luksKillSlot or
737 However, this operation will not change volume key iteration count
738 (MK iterations in output of "cryptsetup luksDump"). In order to
739 change that, you will have to backup the data in the LUKS
740 container (i.e. your encrypted data), luksFormat on the slow
741 machine and restore the data. Note that in the original LUKS
742 specification this value was fixed to 10, but it is now derived
743 from the PBKDF2 benchmark as well and set to iterations in 0.125
744 sec or 1000, whichever is larger. Also note that MK iterations
745 are not very security relevant. But as each key-slot already takes
746 1 second, spending the additional 0.125 seconds really does not
750 * 3.5 "blkid" sees a LUKS UUID and an ext2/swap UUID on the same
751 device. What is wrong?
753 Some old versions of cryptsetup have a bug where the header does
754 not get completely wiped during LUKS format and an older ext2/swap
755 signature remains on the device. This confuses blkid.
757 Fix: Wipe the unused header areas by doing a backup and restore of
758 the header with cryptsetup 1.1.x:
760 cryptsetup luksHeaderBackup --header-backup-file <file> <device>
761 cryptsetup luksHeaderRestore --header-backup-file <file> <device>
764 * 3.6 cryptsetup segfaults on Gentoo amd64 hardened ...
766 There seems to be some interference between the hardening and and
767 the way cryptsetup benchmarks PBKDF2. The solution to this is
768 currently not quite clear for an encrypted root filesystem. For
769 other uses, you can apparently specify USE="dynamic" as compile
770 flag, see http://bugs.gentoo.org/show_bug.cgi?id=283470
776 * 4.1 I get the error "LUKS keyslot x is invalid." What does that
779 This means that the given keyslot has an offset that points
780 outside the valid keyslot area. Typically, the reason is a
781 corrupted LUKS header because something was written to the start of
782 the device the LUKS container is on. Refer to Section "Backup and
783 Data Recovery" and ask on the mailing list if you have trouble
784 diagnosing and (if still possible) repairing this.
787 * 4.2 I cannot unlock my LUKS container! What could be the problem?
789 First, make sure you have a correct passphrase. Then make sure you
790 have the correct key-map and correct keyboard. And then make sure
791 you have the correct character set and encoding, see also
792 "PASSPHRASE CHARACTER SET" under Section 1.2.
794 If you are sure you are entering the passphrase right, there is the
795 possibility that the respective key-slot has been damaged. There
796 is no way to recover a damaged key-slot, except from a header
797 backup (see Section 6). For security reasons, there is also no
798 checksum in the key-slots that could tell you whether a key-slot has
799 been damaged. The only checksum present allows recognition of a
800 correct passphrase, but that only works if the passphrase is
801 correct and the respective key-slot is intact.
803 In order to find out whether a key-slot is damaged one has to look
804 for "non-random looking" data in it. There is a tool that
805 automatizes this in the cryptsetup distribution from version 1.6.0
806 onwards. It is located in misc/keyslot_checker/. Instructions how
807 to use and how to interpret results are in the README file. Note
808 that this tool requires a libcryptsetup from cryptsetup 1.6.0 or
809 later (which means libcryptsetup.so.4.5.0 or later). If the tool
810 complains about missing functions in libcryptsetup, you likely
811 have an earlier version from your distribution still installed. You
812 can either point the symbolic link(s) from libcryptsetup.so.4 to
813 the new version manually, or you can uninstall the distribution
814 version of cryptsetup and re-install that from cryptsetup >= 1.6.0
818 * 4.3 Can a bad RAM module cause problems?
820 LUKS and dm-crypt can give the RAM quite a workout, especially when
821 combined with software RAID. In particular the combination RAID5 +
822 LUKS + XFS seems to uncover RAM problems that never caused obvious
823 problems before. Symptoms vary, but often the problem manifest
824 itself when copying large amounts of data, typically several times
825 larger than your main memory.
827 Side note: One thing you should always do on large data
828 copy/movements is to run a verify, for example with the "-d"
829 option of "tar" or by doing a set of MD5 checksums on the source
832 find . -type f -exec md5sum \{\} \; > checksum-file
834 and then a "md5sum -c checksum-file" on the other side. If you get
835 mismatches here, RAM is the primary suspect. A lesser suspect is
836 an overclocked CPU. I have found countless hardware problems in
837 verify runs after copying or making backups. Bit errors are much
838 more common than most people think.
840 Some RAM issues are even worse and corrupt structures in one of the
841 layers. This typically results in lockups, CPU state dumps in the
842 system logs, kernel panic or other things. It is quite possible to
843 have the problem with an encrypted device, but not with an
844 otherwise the same unencrypted device. The reason for that is that
845 encryption has an error amplification property: You flip one bit
846 in an encrypted data block, and the decrypted version has half of
847 its bits flipped. This is an important security property for modern
848 ciphers. With the usual modes in cryptsetup (CBC, ESSIV, XTS), you
849 get up to a completely changed 512 byte block per bit error. A
850 corrupt block causes a lot more havoc than the occasionally
851 flipped single bit and can result in various obscure errors.
853 Note, that a verify run on copying between encrypted or
854 unencrypted devices will reliably detect corruption, even when the
855 copying itself did not report any problems. If you find defect
856 RAM, assume all backups and copied data to be suspect, unless you
860 * 4.4 How do I test RAM?
862 First you should know that overclocking often makes memory
863 problems worse. So if you overclock (which I strongly recommend
864 against in a system holding data that has some worth), run the
865 tests with the overclocking active.
867 There are two good options. One is Memtest86+ and the other is
868 "memtester" by Charles Cazabon. Memtest86+ requires a reboot and
869 then takes over the machine, while memtester runs from a
870 root-shell. Both use different testing methods and I have found
871 problems fast with each one that the other needed long to find. I
872 recommend running the following procedure until the first error is
875 - Run Memtest86+ for one cycle
877 - Run memtester for one cycle (shut down as many other applications
880 - Run Memtest86+ for 24h or more
882 - Run memtester for 24h or more
884 If all that does not produce error messages, your RAM may be sound,
885 but I have had one weak bit that Memtest86+ needed around 60 hours
886 to find. If you can reproduce the original problem reliably, a good
887 additional test may be to remove half of the RAM (if you have more
888 than one module) and try whether the problem is still there and if
889 so, try with the other half. If you just have one module, get a
890 different one and try with that. If you do overclocking, reduce
891 the settings to the most conservative ones available and try with
898 * 5.1 How long is a secure passphrase ?
900 This is just the short answer. For more info and explanation of
901 some of the terms used in this item, read the rest of Section 5.
902 The actual recommendation is at the end of this item.
904 First, passphrase length is not really the right measure,
905 passphrase entropy is. For example, a random lowercase letter (a-z)
906 gives you 4.7 bit of entropy, one element of a-z0-9 gives you 5.2
907 bits of entropy, an element of a-zA-Z0-9 gives you 5.9 bits and
908 a-zA-Z0-9!@#$%^&:-+ gives you 6.2 bits. On the other hand, a random
909 English word only gives you 0.6...1.3 bits of entropy per
910 character. Using sentences that make sense gives lower entropy,
911 series of random words gives higher entropy. Do not use sentences
912 that can be tied to you or found on your computer. This type of
913 attack is done routinely today.
915 That said, it does not matter too much what scheme you use, but it
916 does matter how much entropy your passphrase contains, because an
917 attacker has to try on average
919 1/2 * 2^(bits of entropy in passphrase)
921 different passphrases to guess correctly.
923 Historically, estimations tended to use computing time estimates,
924 but more modern approaches try to estimate cost of guessing a
927 As an example, I will try to get an estimate from the numbers in
928 http://it.slashdot.org/story/12/12/05/0623215/new-25-gpu-monster-devours-strong-passwords-in-minutes
929 More references can be found a the end of this document. Note that
930 these are estimates from the defender side, so assuming something
931 is easier than it actually is is fine. An attacker may still have
932 vastly higher cost than estimated here.
934 LUKS uses SHA1 for hashing per default. The claim in the reference
935 is 63 billion tries/second for SHA1. We will leave aside the check
936 whether a try actually decrypts a key-slot. Now, the machine has 25
937 GPUs, which I will estimate at an overall lifetime cost of USD/EUR
938 1000 each, and an useful lifetime of 2 years. (This is on the low
939 side.) Disregarding downtime, the machine can then break
941 N = 63*10^9 * 3600 * 24 * 365 * 2 ~ 4*10^18
943 passphrases for EUR/USD 25k. That is one 62 bit passphrase hashed
944 once with SHA1 for EUR/USD 25k. Note that as this can be
945 parallelized, it can be done faster than 2 years with several of
948 For plain dm-crypt (no hash iteration) this is it. This gives (with
949 SHA1, plain dm-crypt default is ripemd160 which seems to be
950 slightly slower than SHA1):
952 Passphrase entropy Cost to break
961 For LUKS, you have to take into account hash iteration in PBKDF2.
962 For a current CPU, there are about 100k iterations (as can be
963 queried with ''cryptsetup luksDump''.
965 The table above then becomes:
967 Passphrase entropy Cost to break
978 To get reasonable security for the next 10 years, it is a good idea
979 to overestimate by a factor of at least 1000.
981 Then there is the question of how much the attacker is willing to
982 spend. That is up to your own security evaluation. For general use,
983 I will assume the attacker is willing to spend up to 1 million
984 EUR/USD. Then we get the following recommendations:
986 Plain dm-crypt: Use > 80 bit. That is e.g. 17 random chars from a-z
987 or a random English sentence of > 135 characters length.
989 LUKS: Use > 65 bit. That is e.g. 14 random chars from a-z or a
990 random English sentence of > 108 characters length.
992 If paranoid, add at least 20 bit. That is roughly four additional
993 characters for random passphrases and roughly 32 characters for a
994 random English sentence.
997 * 5.2 Is LUKS insecure? Everybody can see I have encrypted data!
999 In practice it does not really matter. In most civilized countries
1000 you can just refuse to hand over the keys, no harm done. In some
1001 countries they can force you to hand over the keys, if they suspect
1002 encryption. However the suspicion is enough, they do not have to
1003 prove anything. This is for practical reasons, as even the presence
1004 of a header (like the LUKS header) is not enough to prove that you
1005 have any keys. It might have been an experiment, for example. Or it
1006 was used as encrypted swap with a key from /dev/random. So they
1007 make you prove you do not have encrypted data. Of course that is
1008 just as impossible as the other way round.
1010 This means that if you have a large set of random-looking data,
1011 they can already lock you up. Hidden containers (encryption hidden
1012 within encryption), as possible with Truecrypt, do not help
1013 either. They will just assume the hidden container is there and
1014 unless you hand over the key, you will stay locked up. Don't have
1015 a hidden container? Though luck. Anybody could claim that.
1017 Still, if you are concerned about the LUKS header, use plain
1018 dm-crypt with a good passphrase. See also Section 2, "What is the
1019 difference between "plain" and LUKS format?"
1022 * 5.3 Should I initialize (overwrite) a new LUKS/dm-crypt partition?
1024 If you just create a filesystem on it, most of the old data will
1025 still be there. If the old data is sensitive, you should overwrite
1026 it before encrypting. In any case, not initializing will leave the
1027 old data there until the specific sector gets written. That may
1028 enable an attacker to determine how much and where on the
1029 partition data was written. If you think this is a risk, you can
1030 prevent this by overwriting the encrypted device (here assumed to
1031 be named "e1") with zeros like this:
1033 dd_rescue -w /dev/zero /dev/mapper/e1
1035 or alternatively with one of the following more standard commands:
1037 cat /dev/zero > /dev/mapper/e1
1038 dd if=/dev/zero of=/dev/mapper/e1
1041 * 5.4 How do I securely erase a LUKS (or other) partition?
1043 For LUKS, if you are in a desperate hurry, overwrite the LUKS
1044 header and key-slot area. This means overwriting the first
1045 (keyslots x stripes x keysize) + offset bytes. For the default
1046 parameters, this is the 1'052'672 bytes, i.e. 1MiB + 4096 of the
1047 LUKS partition. For 512 bit key length (e.g. for aes-xts-plain with
1048 512 bit key) this is 2MiB. (The different offset stems from
1049 differences in the sector alignment of the key-slots.) If in doubt,
1050 just be generous and overwrite the first 10MB or so, it will likely
1051 still be fast enough. A single overwrite with zeros should be
1052 enough. If you anticipate being in a desperate hurry, prepare the
1053 command beforehand. Example with /dev/sde1 as the LUKS partition
1054 and default parameters:
1056 head -c 1052672 /dev/zero > /dev/sde1; sync
1058 A LUKS header backup or full backup will still grant access to
1059 most or all data, so make sure that an attacker does not have
1060 access to backups or destroy them as well.
1062 If you have time, overwrite the whole LUKS partition with a single
1063 pass of zeros. This is enough for current HDDs. For SSDs or FLASH
1064 (USB sticks) you may want to overwrite the whole drive several
1065 times to be sure data is not retained by wear leveling. This is
1066 possibly still insecure as SSD technology is not fully understood
1067 in this regard. Still, due to the anti-forensic properties of the
1068 LUKS key-slots, a single overwrite of an SSD or FLASH drive could
1069 be enough. If in doubt, use physical destruction in addition. Here
1070 is a link to some current research results on erasing SSDs and
1072 http://www.usenix.org/events/fast11/tech/full_papers/Wei.pdf
1074 Keep in mind to also erase all backups.
1076 Example for a zero-overwrite erase of partition sde1 done with
1079 dd_rescue -w /dev/zero /dev/sde1
1082 * 5.5 How do I securely erase a backup of a LUKS partition or header?
1084 That depends on the medium it is stored on. For HDD and SSD, use
1085 overwrite with zeros. For an SSD or FLASH drive (USB stick), you
1086 may want to overwrite the complete SSD several times and use
1087 physical destruction in addition, see last item. For re-writable
1088 CD/DVD, a single overwrite should also be enough, due to the
1089 anti-forensic properties of the LUKS keyslots. For write-once
1090 media, use physical destruction. For low security requirements,
1091 just cut the CD/DVD into several parts. For high security needs,
1092 shred or burn the medium. If your backup is on magnetic tape, I
1093 advise physical destruction by shredding or burning, after
1094 overwriting . The problem with magnetic tape is that it has a
1095 higher dynamic range than HDDs and older data may well be
1096 recoverable after overwrites. Also write-head alignment issues can
1097 lead to data not actually being deleted at all during overwrites.
1100 * 5.6 What about backup? Does it compromise security?
1102 That depends. See item 6.7.
1105 * 5.7 Why is all my data permanently gone if I overwrite the LUKS
1108 Overwriting the LUKS header in part or in full is the most common
1109 reason why access to LUKS containers is lost permanently.
1110 Overwriting can be done in a number of fashions, like creating a
1111 new filesystem on the raw LUKS partition, making the raw partition
1112 part of a raid array and just writing to the raw partition.
1114 The LUKS header contains a 256 bit "salt" value and without that no
1115 decryption is possible. While the salt is not secret, it is
1116 key-grade material and cannot be reconstructed. This is a
1117 cryptographically strong "cannot". From observations on the
1118 cryptsetup mailing-list, people typically go though the usual
1119 stages of grief (Denial, Anger, Bargaining, Depression, Acceptance)
1120 when this happens to them. Observed times vary between 1 day and 2
1121 weeks to complete the cycle. Seeking help on the mailing-list is
1122 fine. Even if we usually cannot help with getting back your data,
1123 most people found the feedback comforting.
1125 If your header does not contain an intact salt, best go directly
1126 to the last stage ("Acceptance") and think about what to do now.
1127 There is one exception that I know of: If your LUKS container is
1128 still open, then it may be possible to extract the master key from
1129 the running system. See Item "How do I recover the master key from
1130 a mapped LUKS container?" in Section "Backup and Data Recovery".
1133 * 5.8 What is a "salt"?
1135 A salt is a random key-grade value added to the passphrase before
1136 it is processed. It is not kept secret. The reason for using salts
1137 is as follows: If an attacker wants to crack the password for a
1138 single LUKS container, then every possible passphrase has to be
1139 tried. Typically an attacker will not try every binary value, but
1140 will try words and sentences from a dictionary.
1142 If an attacker wants to attack several LUKS containers with the
1143 same dictionary, then a different approach makes sense: Compute the
1144 resulting slot-key for each dictionary element and store it on
1145 disk. Then the test for each entry is just the slow unlocking with
1146 the slot key (say 0.00001 sec) instead of calculating the slot-key
1147 first (1 sec). For a single attack, this does not help. But if you
1148 have more than one container to attack, this helps tremendously,
1149 also because you can prepare your table before you even have the
1150 container to attack! The calculation is also very simple to
1151 parallelize. You could, for example, use the night-time unused CPU
1152 power of your desktop PCs for this.
1154 This is where the salt comes in. If the salt is combined with the
1155 passphrase (in the simplest form, just appended to it), you
1156 suddenly need a separate table for each salt value. With a
1157 reasonably-sized salt value (256 bit, e.g.) this is quite
1161 * 5.9 Is LUKS secure with a low-entropy (bad) passphrase?
1163 Note: You should only use the 94 printable characters from 7 bit
1164 ASCII code to prevent your passphrase from failing when the
1165 character encoding changes, e.g. because of a system upgrade, see
1166 also the note at the very start of this FAQ under "WARNINGS".
1168 This needs a bit of theory. The quality of your passphrase is
1169 directly related to its entropy (information theoretic, not
1170 thermodynamic). The entropy says how many bits of "uncertainty" or
1171 "randomness" are in you passphrase. In other words, that is how
1172 difficult guessing the passphrase is.
1174 Example: A random English sentence has about 1 bit of entropy per
1175 character. A random lowercase (or uppercase) character has about
1178 Now, if n is the number of bits of entropy in your passphrase and t
1179 is the time it takes to process a passphrase in order to open the
1180 LUKS container, then an attacker has to spend at maximum
1182 attack_time_max = 2^n * t
1184 time for a successful attack and on average half that. There is no
1185 way getting around that relationship. However, there is one thing
1186 that does help, namely increasing t, the time it takes to use a
1187 passphrase, see next FAQ item.
1189 Still, if you want good security, a high-entropy passphrase is the
1190 only option. For example, a low-entropy passphrase can never be
1191 considered secure against a TLA-level (Three Letter Agency level,
1192 i.e. government-level) attacker, no matter what tricks are used in
1193 the key-derivation function. Use at least 64 bits for secret stuff.
1194 That is 64 characters of English text (but only if randomly chosen)
1195 or a combination of 12 truly random letters and digits.
1197 For passphrase generation, do not use lines from very well-known
1198 texts (religious texts, Harry potter, etc.) as they are to easy to
1199 guess. For example, the total Harry Potter has about 1'500'000
1200 words (my estimation). Trying every 64 character sequence starting
1201 and ending at a word boundary would take only something like 20
1202 days on a single CPU and is entirely feasible. To put that into
1203 perspective, using a number of Amazon EC2 High-CPU Extra Large
1204 instances (each gives about 8 real cores), this test costs
1205 currently about 50USD/EUR, but can be made to run arbitrarily fast.
1207 On the other hand, choosing 1.5 lines from, say, the Wheel of Time
1208 is in itself not more secure, but the book selection adds quite a
1209 bit of entropy. (Now that I have mentioned it here, don't use tWoT
1210 either!) If you add 2 or 3 typos or switch some words around, then
1211 this is good passphrase material.
1214 * 5.10 What is "iteration count" and why is decreasing it a bad idea?
1216 Iteration count is the number of PBKDF2 iterations a passphrase is
1217 put through before it is used to unlock a key-slot. Iterations are
1218 done with the explicit purpose to increase the time that it takes
1219 to unlock a key-slot. This provides some protection against use of
1220 low-entropy passphrases.
1222 The idea is that an attacker has to try all possible passphrases.
1223 Even if the attacker knows the passphrase is low-entropy (see last
1224 item), it is possible to make each individual try take longer. The
1225 way to do this is to repeatedly hash the passphrase for a certain
1226 time. The attacker then has to spend the same time (given the same
1227 computing power) as the user per try. With LUKS, the default is 1
1228 second of PBKDF2 hashing.
1230 Example 1: Lets assume we have a really bad passphrase (e.g. a
1231 girlfriends name) with 10 bits of entropy. With the same CPU, an
1232 attacker would need to spend around 500 seconds on average to
1233 break that passphrase. Without iteration, it would be more like
1234 0.0001 seconds on a modern CPU.
1236 Example 2: The user did a bit better and has 32 chars of English
1237 text. That would be about 32 bits of entropy. With 1 second
1238 iteration, that means an attacker on the same CPU needs around 136
1239 years. That is pretty impressive for such a weak passphrase.
1240 Without the iterations, it would be more like 50 days on a modern
1241 CPU, and possibly far less.
1243 In addition, the attacker can both parallelize and use special
1244 hardware like GPUs or FPGAs to speed up the attack. The attack can
1245 also happen quite some time after the luksFormat operation and CPUs
1246 can have become faster and cheaper. For that reason you want a
1247 bit of extra security. Anyways, in Example 1 your are screwed.
1248 In example 2, not necessarily. Even if the attack is faster, it
1249 still has a certain cost associated with it, say 10000 EUR/USD
1250 with iteration and 1 EUR/USD without iteration. The first can be
1251 prohibitively expensive, while the second is something you try
1252 even without solid proof that the decryption will yield something
1255 The numbers above are mostly made up, but show the idea. Of course
1256 the best thing is to have a high-entropy passphrase.
1258 Would a 100 sec iteration time be even better? Yes and no.
1259 Cryptographically it would be a lot better, namely 100 times better.
1260 However, usability is a very important factor for security
1261 technology and one that gets overlooked surprisingly often. For
1262 LUKS, if you have to wait 2 minutes to unlock the LUKS container,
1263 most people will not bother and use less secure storage instead. It
1264 is better to have less protection against low-entropy passphrases
1265 and people actually use LUKS, than having them do without
1266 encryption altogether.
1268 Now, what about decreasing the iteration time? This is generally a
1269 very bad idea, unless you know and can enforce that the users only
1270 use high-entropy passphrases. If you decrease the iteration time
1271 without ensuring that, then you put your users at increased risk,
1272 and considering how rarely LUKS containers are unlocked in a
1273 typical work-flow, you do so without a good reason. Don't do it.
1274 The iteration time is already low enough that users with entropy
1275 low passphrases are vulnerable. Lowering it even further increases
1276 this danger significantly.
1279 * 5.11 Some people say PBKDF2 is insecure?
1281 There is some discussion that a hash-function should have a "large
1282 memory" property, i.e. that it should require a lot of memory to be
1283 computed. This serves to prevent attacks using special programmable
1284 circuits, like FPGAs, and attacks using graphics cards. PBKDF2
1285 does not need a lot of memory and is vulnerable to these attacks.
1286 However, the publication usually referred in these discussions is
1287 not very convincing in proving that the presented hash really is
1288 "large memory" (that may change, email the FAQ maintainer when it
1289 does) and it is of limited usefulness anyways. Attackers that use
1290 clusters of normal PCs will not be affected at all by a "large
1291 memory" property. For example the US Secret Service is known to
1292 use the off-hour time of all the office PCs of the Treasury for
1293 password breaking. The Treasury has about 110'000 employees.
1294 Assuming every one has an office PC, that is significant computing
1295 power, all of it with plenty of memory for computing "large
1296 memory" hashes. Bot-net operators also have all the memory they
1297 want. The only protection against a resourceful attacker is a
1298 high-entropy passphrase, see items 5.9 and 5.10.
1301 * 5.12 What about iteration count with plain dm-crypt?
1303 Simple: There is none. There is also no salting. If you use plain
1304 dm-crypt, the only way to be secure is to use a high entropy
1305 passphrase. If in doubt, use LUKS instead.
1308 * 5.13 Is LUKS with default parameters less secure on a slow CPU?
1310 Unfortunately, yes. However the only aspect affected is the
1311 protection for low-entropy passphrase or master-key. All other
1312 security aspects are independent of CPU speed.
1314 The master key is less critical, as you really have to work at it
1315 to give it low entropy. One possibility is to supply the master key
1316 yourself. If that key is low-entropy, then you get what you
1317 deserve. The other known possibility is to use /dev/urandom for
1318 key generation in an entropy-starved situation (e.g. automatic
1319 installation on an embedded device without network and other entropy
1322 For the passphrase, don't use a low-entropy passphrase. If your
1323 passphrase is good, then a slow CPU will not matter. If you insist
1324 on a low-entropy passphrase on a slow CPU, use something like
1325 "--iter-time=10" or higher and wait a long time on each LUKS unlock
1326 and pray that the attacker does not find out in which way exactly
1327 your passphrase is low entropy. This also applies to low-entropy
1328 passphrases on fast CPUs. Technology can do only so much to
1329 compensate for problems in front of the keyboard.
1332 * 5.14 Why was the default aes-cbc-plain replaced with aes-cbc-essiv?
1334 Note: This item applies both to plain dm-crypt and to LUKS
1336 The problem is that cbc-plain has a fingerprint vulnerability, where
1337 a specially crafted file placed into the crypto-container can be
1338 recognized from the outside. The issue here is that for cbc-plain
1339 the initialization vector (IV) is the sector number. The IV gets
1340 XORed to the first data chunk of the sector to be encrypted. If you
1341 make sure that the first data block to be stored in a sector
1342 contains the sector number as well, the first data block to be
1343 encrypted is all zeros and always encrypted to the same ciphertext.
1344 This also works if the first data chunk just has a constant XOR
1345 with the sector number. By having several shifted patterns you can
1346 take care of the case of a non-power-of-two start sector number of
1349 This mechanism allows you to create a pattern of sectors that have
1350 the same first ciphertext block and signal one bit per sector to the
1351 outside, allowing you to e.g. mark media files that way for
1352 recognition without decryption. For large files this is a
1353 practical attack. For small ones, you do not have enough blocks to
1354 signal and take care of different file starting offsets.
1356 In order to prevent this attack, the default was changed to
1357 cbc-essiv. ESSIV uses a keyed hash of the sector number, with the
1358 encryption key as key. This makes the IV unpredictable without
1359 knowing the encryption key and the watermarking attack fails.
1362 * 5.15 Are there any problems with "plain" IV? What is "plain64"?
1364 First, "plain" and "plain64" are both not secure to use with CBC,
1365 see previous FAQ item.
1367 However there are modes, like XTS, that are secure with "plain" IV.
1368 The next limit is that "plain" is 64 bit, with the upper 32 bit set
1369 to zero. This means that on volumes larger than 2TiB, the IV
1370 repeats, creating a vulnerability that potentially leaks some
1371 data. To avoid this, use "plain64", which uses the full sector
1372 number up to 64 bit. Note that "plain64" requires a kernel >=
1373 2.6.33. Also note that "plain64" is backwards compatible for
1374 volume sizes <= 2TiB, but not for those > 2TiB. Finally, "plain64"
1375 does not cause any performance penalty compared to "plain".
1378 * 5.16 What about XTS mode?
1380 XTS mode is potentially even more secure than cbc-essiv (but only if
1381 cbc-essiv is insecure in your scenario). It is a NIST standard and
1382 used, e.g. in Truecrypt. At the moment, if you want to use it, you
1383 have to specify it manually as "aes-xts-plain", i.e.
1385 cryptsetup -c aes-xts-plain luksFormat <device>
1387 For volumes >2TiB and kernels >= 2.6.33 use "plain64" (see FAQ
1388 item on "plain" and "plain64"):
1390 cryptsetup -c aes-xts-plain64 luksFormat <device>
1392 There is a potential security issue with XTS mode and large blocks.
1393 LUKS and dm-crypt always use 512B blocks and the issue does not
1397 * 5.17 Is LUKS FIPS-140-2 certified?
1399 No. But that is more a problem of FIPS-140-2 than of LUKS. From a
1400 technical point-of-view, LUKS with the right parameters would be
1401 FIPS-140-2 compliant, but in order to make it certified, somebody
1402 has to pay real money for that. And then, whenever cryptsetup is
1403 changed or extended, the certification lapses and has to be
1406 From the aspect of actual security, LUKS with default parameters
1407 should be as good as most things that are FIPS-140-2 certified,
1408 although you may want to make sure to use /dev/random (by
1409 specifying --use-random on luksFormat) as randomness source for
1410 the master key to avoid being potentially insecure in an
1411 entropy-starved situation.
1414 * 5.18 What about Plausible Deniability?
1416 First let me attempt a definition for the case of encrypted
1417 filesystems: Plausible deniability is when you hide encrypted data
1418 inside an encrypted container and it is not possible to prove it is
1419 there. The idea is compelling and on first glance it seems
1420 possible to do it. And from a cryptographic point of view, it
1421 actually is possible.
1423 So, does it work in practice? No, unfortunately. The reasoning used
1424 by its proponents is fundamentally flawed in several ways and the
1425 cryptographic properties fail fatally when colliding with the real
1428 First, why should "I do not have a hidden partition" be any more
1429 plausible than "I forgot my crypto key" or "I wiped that partition
1430 with random data, nothing in there"? I do not see any reason.
1432 Second, there are two types of situations: Either they cannot force
1433 you to give them the key (then you simply do not) or the can. In
1434 the second case, they can always do bad things to you, because they
1435 cannot prove that you have the key in the first place! This means
1436 they do not have to prove you have the key, or that this random
1437 looking data on your disk is actually encrypted data. So the
1438 situation will allow them to waterboard/lock-up/deport you
1439 anyways, regardless of how "plausible" your deniability is. Do not
1440 have a hidden partition you could show to them, but there are
1441 indications you may? Too bad for you. Unfortunately "plausible
1442 deniability" also means you cannot prove there is no hidden data.
1444 Third, hidden partitions are not that hidden. There are basically
1445 just two possibilities: a) Make a large crypto container, but put a
1446 smaller filesystem in there and put the hidden partition into the
1447 free space. Unfortunately this is glaringly obvious and can be
1448 detected in an automated fashion. This means that the initial
1449 suspicion to put you under duress in order to make you reveal you
1450 hidden data is given. b) Make a filesystem that spans the whole
1451 encrypted partition, and put the hidden partition into space not
1452 currently used by that filesystem. Unfortunately that is also
1453 glaringly obvious, as you then cannot write to the filesystem
1454 without a high risk of destroying data in the hidden container.
1455 Have not written anything to the encrypted filesystem in a while?
1456 Too bad, they have the suspicion they need to do unpleasant things
1459 To be fair, if you prepare option b) carefully and directly before
1460 going into danger, it may work. But then, the mere presence of
1461 encrypted data may already be enough to get you into trouble in
1462 those places were they can demand encryption keys.
1464 Here is an additional reference for some problems with plausible
1465 deniability: http://www.schneier.com/paper-truecrypt-dfs.pdf I
1466 strongly suggest you read it.
1468 So, no, I will not provide any instructions on how to do it with
1469 plain dm-crypt or LUKS. If you insist on shooting yourself in the
1470 foot, you can figure out how to do it yourself.
1473 * 5.19 What about SSDs, Flash and Hybrid Drives?
1475 The problem is that you cannot reliably erase parts of these
1476 devices, mainly due to wear-leveling and possibly defect
1479 Basically, when overwriting a sector (of 512B), what the device
1480 does is to move an internal sector (may be 128kB or even larger) to
1481 some pool of discarded, not-yet erased unused sectors, take a
1482 fresh empty sector from the empty-sector pool and copy the old
1483 sector over with the changes to the small part you wrote. This is
1484 done in some fashion so that larger writes do not cause a lot of
1485 small internal updates.
1487 The thing is that the mappings between outside-addressable sectors
1488 and inside sectors is arbitrary (and the vendors are not talking).
1489 Also the discarded sectors are not necessarily erased immediately.
1490 They may linger a long time.
1492 For plain dm-crypt, the consequences are that older encrypted data
1493 may be lying around in some internal pools of the device. Thus may
1494 or may not be a problem and depends on the application. Remember
1495 the same can happen with a filesystem if consecutive writes to the
1496 same area of a file can go to different sectors.
1498 However, for LUKS, the worst case is that key-slots and LUKS
1499 header may end up in these internal pools. This means that password
1500 management functionality is compromised (the old passwords may
1501 still be around, potentially for a very long time) and that fast
1502 erase by overwriting the header and key-slot area is insecure.
1504 Also keep in mind that the discarded/used pool may be large. For
1505 example, a 240GB SSD has about 16GB of spare area in the chips that
1506 it is free to do with as it likes. You would need to make each
1507 individual key-slot larger than that to allow reliable overwriting.
1508 And that assumes the disk thinks all other space is in use.
1509 Reading the internal pools using forensic tools is not that hard,
1510 but may involve some soldering.
1514 If you trust the device vendor (you probably should not...) you can
1515 try an ATA "secure erase" command for SSDs. That does not work for
1516 USB keys though and may or may not be secure for a hybrid drive. If
1517 it finishes on an SSD after a few seconds, it was possibly faked.
1518 UNfortunately, for hybrid drives that indicator does not work, as
1519 the drive may well take the time to dully erase the magnetic part,
1520 but only mark the SSD/Flash part as erased while data is still in
1523 If you can do without password management and are fine with doing
1524 physical destruction for permanently deleting data (always after
1525 one or several full overwrites!), you can use plain dm-crypt or
1528 If you want or need the original LUKS security features to work,
1529 you can use a detached LUKS header and put that on a conventional,
1530 magnetic disk. That leaves potentially old encrypted data in the
1531 pools on the disk, but otherwise you get LUKS with the same
1532 security as on a magnetic disk.
1534 If you are concerned about your laptop being stolen, you are likely
1535 fine using LUKS on an SSD or hybrid drive. An attacker would need
1536 to have access to an old passphrase (and the key-slot for this old
1537 passphrase would actually need to still be somewhere in the SSD)
1538 for your data to be at risk. So unless you pasted your old
1539 passphrase all over the Internet or the attacker has knowledge of
1540 it from some other source and does a targeted laptop theft to get
1541 at your data, you should be fine.
1544 6. Backup and Data Recovery
1547 * 6.1 Why do I need Backup?
1549 First, disks die. The rate for well-treated (!) disk is about 5%
1550 per year, which is high enough to worry about. There is some
1551 indication that this may be even worse for some SSDs. This applies
1552 both to LUKS and plain dm-crypt partitions.
1554 Second, for LUKS, if anything damages the LUKS header or the
1555 key-stripe area then decrypting the LUKS device can become
1556 impossible. This is a frequent occurrence. For example an
1557 accidental format as FAT or some software overwriting the first
1558 sector where it suspects a partition boot sector typically makes a
1559 LUKS partition permanently inaccessible. See more below on LUKS
1562 So, data-backup in some form is non-optional. For LUKS, you may
1563 also want to store a header backup in some secure location. This
1564 only needs an update if you change passphrases.
1567 * 6.2 How do I backup a LUKS header?
1569 While you could just copy the appropriate number of bytes from the
1570 start of the LUKS partition, the best way is to use command option
1571 "luksHeaderBackup" of cryptsetup. This protects also against
1572 errors when non-standard parameters have been used in LUKS
1573 partition creation. Example:
1576 cryptsetup luksHeaderBackup --header-backup-file <file> <device>
1578 To restore, use the inverse command, i.e.
1580 cryptsetup luksHeaderRestore --header-backup-file <file> <device>
1583 * 6.3 How do I test a LUKS header?
1587 cryptsetup -v isLuks <device>
1589 on the device. Without the "-v" it just signals its result via
1590 exit-status. You can also use the more general test
1594 which will also detect other types and give some more info. Omit
1595 "-p" for old versions of blkid that do not support it.
1598 * 6.4 How do I backup a LUKS or dm-crypt partition?
1600 There are two options, a sector-image and a plain file or
1601 filesystem backup of the contents of the partition. The sector
1602 image is already encrypted, but cannot be compressed and contains
1603 all empty space. The filesystem backup can be compressed, can
1604 contain only part of the encrypted device, but needs to be
1605 encrypted separately if so desired.
1607 A sector-image will contain the whole partition in encrypted form,
1608 for LUKS the LUKS header, the keys-slots and the data area. It can
1609 be done under Linux e.g. with dd_rescue (for a direct image copy)
1610 and with "cat" or "dd". Example:
1612 cat /dev/sda10 > sda10.img
1613 dd_rescue /dev/sda10 sda10.img
1615 You can also use any other backup software that is capable of making
1616 a sector image of a partition. Note that compression is
1617 ineffective for encrypted data, hence it does not make sense to
1620 For a filesystem backup, you decrypt and mount the encrypted
1621 partition and back it up as you would a normal filesystem. In this
1622 case the backup is not encrypted, unless your encryption method
1623 does that. For example you can encrypt a backup with "tar" as
1626 tar cjf - <path> | gpg --cipher-algo AES -c - > backup.tbz2.gpg
1628 And verify the backup like this if you are at "path":
1630 cat backup.tbz2.gpg | gpg - | tar djf -
1632 Note: Always verify backups, especially encrypted ones.
1634 In both cases GnuPG will ask you interactively for your symmetric
1635 key. The verify will only output errors. Use "tar dvjf -" to get
1636 all comparison results. To make sure no data is written to disk
1637 unencrypted, turn off swap if it is not encrypted before doing the
1640 You can of course use different or no compression and you can use
1641 an asymmetric key if you have one and have a backup of the secret
1642 key that belongs to it.
1644 A second option for a filesystem-level backup that can be used when
1645 the backup is also on local disk (e.g. an external USB drive) is
1646 to use a LUKS container there and copy the files to be backed up
1647 between both mounted containers. Also see next item.
1650 * 6.5 Do I need a backup of the full partition? Would the header and
1651 key-slots not be enough?
1653 Backup protects you against two things: Disk loss or corruption
1654 and user error. By far the most questions on the dm-crypt mailing
1655 list about how to recover a damaged LUKS partition are related
1656 to user error. For example, if you create a new filesystem on a
1657 LUKS partition, chances are good that all data is lost
1660 For this case, a header+key-slot backup would often be enough. But
1661 keep in mind that a well-treated (!) HDD has roughly a failure
1662 risk of 5% per year. It is highly advisable to have a complete
1663 backup to protect against this case.
1666 * *6.6 What do I need to backup if I use "decrypt_derived"?
1668 This is a script in Debian, intended for mounting /tmp or swap with
1669 a key derived from the master key of an already decrypted device.
1670 If you use this for an device with data that should be persistent,
1671 you need to make sure you either do not lose access to that master
1672 key or have a backup of the data. If you derive from a LUKS
1673 device, a header backup of that device would cover backing up the
1674 master key. Keep in mind that this does not protect against disk
1677 Note: If you recreate the LUKS header of the device you derive from
1678 (using luksFormat), the master key changes even if you use the same
1679 passphrase(s) and you will not be able to decrypt the derived
1680 device with the new LUKS header.
1683 * 6.7 Does a backup compromise security?
1685 Depends on how you do it. However if you do not have one, you are
1686 going to eventually lose your encrypted data.
1688 There are risks introduced by backups. For example if you
1689 change/disable a key-slot in LUKS, a binary backup of the partition
1690 will still have the old key-slot. To deal with this, you have to
1691 be able to change the key-slot on the backup as well, securely
1692 erase the backup or do a filesystem-level backup instead of a binary
1695 If you use dm-crypt, backup is simpler: As there is no key
1696 management, the main risk is that you cannot wipe the backup when
1697 wiping the original. However wiping the original for dm-crypt
1698 should consist of forgetting the passphrase and that you can do
1699 without actual access to the backup.
1701 In both cases, there is an additional (usually small) risk with
1702 binary backups: An attacker can see how many sectors and which
1703 ones have been changed since the backup. To prevent this, use a
1704 filesystem level backup method that encrypts the whole backup in
1705 one go, e.g. as described above with tar and GnuPG.
1707 My personal advice is to use one USB disk (low value data) or
1708 three disks (high value data) in rotating order for backups, and
1709 either use independent LUKS partitions on them, or use encrypted
1710 backup with tar and GnuPG.
1712 If you do network-backup or tape-backup, I strongly recommend to
1713 go the filesystem backup path with independent encryption, as you
1714 typically cannot reliably delete data in these scenarios,
1715 especially in a cloud setting. (Well, you can burn the tape if it
1716 is under your control...)
1719 * 6.8 What happens if I overwrite the start of a LUKS partition or
1720 damage the LUKS header or key-slots?
1722 There are two critical components for decryption: The salt values
1723 in the header itself and the key-slots. If the salt values are
1724 overwritten or changed, nothing (in the cryptographically strong
1725 sense) can be done to access the data, unless there is a backup
1726 of the LUKS header. If a key-slot is damaged, the data can still
1727 be read with a different key-slot, if there is a remaining
1728 undamaged and used key-slot. Note that in order to make a key-slot
1729 unrecoverable in a cryptographically strong sense, changing about
1730 4-6 bits in random locations of its 128kiB size is quite enough.
1733 * 6.9 What happens if I (quick) format a LUKS partition?
1735 I have not tried the different ways to do this, but very likely you
1736 will have written a new boot-sector, which in turn overwrites the
1737 LUKS header, including the salts, making your data permanently
1738 irretrievable, unless you have a LUKS header backup. You may also
1739 damage the key-slots in part or in full. See also last item.
1742 * 6.10 How do I recover the master key from a mapped LUKS container?
1744 This is typically only needed if you managed to damage your LUKS
1745 header, but the container is still mapped, i.e. "luksOpen"ed. It
1746 also helps if you have a mapped container that you forgot or do not
1747 know a passphrase for (e.g. on a long running server.)
1749 WARNING: Things go wrong, do a full backup before trying this!
1751 WARNING: This exposes the master key of the LUKS container. Note
1752 that both ways to recreate a LUKS header with the old master key
1753 described below will write the master key to disk. Unless you are
1754 sure you have securely erased it afterwards, e.g. by writing it to
1755 an encrypted partition, RAM disk or by erasing the filesystem you
1756 wrote it to by a complete overwrite, you should change the master
1757 key afterwards. Changing the master key requires a full data
1758 backup, luksFormat and then restore of the backup.
1760 First, there is a script by Milan that automates the whole
1761 process, except generating a new LUKS header with the old master
1762 key (it prints the command for that though):
1764 http://code.google.com/p/cryptsetup/source/browse/misc/luks-header-from-active
1766 You can also do this manually. Here is how:
1768 - Get the master key from the device mapper. This is done by the
1769 following command. Substitute c5 for whatever you mapped to:
1771 # dmsetup table --target crypt --showkey /dev/mapper/c5
1773 0 200704 crypt aes-cbc-essiv:sha256
1774 a1704d9715f73a1bb4db581dcacadaf405e700d591e93e2eaade13ba653d0d09
1777 The result is actually one line, wrapped here for clarity. The long
1778 hex string is the master key.
1780 - Convert the master key to a binary file representation. You can
1781 do this manually, e.g. with hexedit. You can also use the tool
1782 "xxd" from vim like this:
1784 echo "a1704d9....53d0d09" | xxd -r -p > <master-key-file>
1786 - Do a luksFormat to create a new LUKS header.
1788 NOTE: If your header is intact and you just forgot the
1789 passphrase, you can just set a new passphrase, see next
1792 Unmap the device before you do that (luksClose). Then do
1794 cryptsetup luksFormat --master-key-file=<master-key-file> <luks device>
1796 Note that if the container was created with other than the default
1797 settings of the cryptsetup version you are using, you need to give
1798 additional parameters specifying the deviations. If in doubt, try
1799 the script by Milan. It does recover the other parameters as well.
1801 Side note: This is the way the decrypt_derived script gets at the
1802 master key. It just omits the conversion and hashes the master key
1805 - If the header is intact and you just forgot the passphrase, just
1806 set a new passphrase like this:
1808 cryptsetup luksAddKey --master-key-file=<master-key-file> <luks device>
1810 You may want to disable the old one afterwards.
1813 * 6.11 What does the on-disk structure of dm-crypt look like?
1815 There is none. dm-crypt takes a block device and gives encrypted
1816 access to each of its blocks with a key derived from the passphrase
1817 given. If you use a cipher different than the default, you have to
1818 specify that as a parameter to cryptsetup too. If you want to
1819 change the password, you basically have to create a second
1820 encrypted device with the new passphrase and copy your data over.
1821 On the plus side, if you accidentally overwrite any part of a
1822 dm-crypt device, the damage will be limited to the are you
1826 * 6.12 What does the on-disk structure of LUKS look like?
1828 A LUKS partition consists of a header, followed by 8 key-slot
1829 descriptors, followed by 8 key slots, followed by the encrypted
1832 Header and key-slot descriptors fill the first 592 bytes. The
1833 key-slot size depends on the creation parameters, namely on the
1834 number of anti-forensic stripes, key material offset and master
1837 With the default parameters, each key-slot is a bit less than
1838 128kiB in size. Due to sector alignment of the key-slot start,
1839 that means the key block 0 is at offset 0x1000-0x20400, key
1840 block 1 at offset 0x21000-0x40400, and key block 7 at offset
1841 0xc1000-0xe0400. The space to the next full sector address is
1842 padded with zeros. Never used key-slots are filled with what the
1843 disk originally contained there, a key-slot removed with
1844 "luksRemoveKey" or "luksKillSlot" gets filled with 0xff. Due to
1845 2MiB default alignment, start of the data area for cryptsetup 1.3
1846 and later is at 2MiB, i.e. at 0x200000. For older versions, it is
1847 at 0x101000, i.e. at 1'052'672 bytes, i.e. at 1MiB + 4096 bytes
1848 from the start of the partition. Incidentally, "luksHeaderBackup"
1849 for a LUKS container created with default parameters dumps exactly
1850 the first 2MiB (or 1'052'672 bytes for headers created with
1851 cryptsetup versions < 1.3) to file and "luksHeaderRestore" restores
1854 For non-default parameters, you have to figure out placement
1855 yourself. "luksDump" helps. See also next item. For the most common
1856 non-default settings, namely aes-xts-plain with 512 bit key, the
1857 offsets are: 1st keyslot 0x1000-0x3f800, 2nd keyslot
1858 0x40000-0x7e000, 3rd keyslot 0x7e000-0xbd800, ..., and start of
1859 bulk data at 0x200000.
1861 The exact specification of the format is here:
1862 http://code.google.com/p/cryptsetup/wiki/Specification
1865 * 6.13 What is the smallest possible LUKS container?
1867 Note: From cryptsetup 1.3 onwards, alignment is set to 1MB. With
1868 modern Linux partitioning tools that also align to 1MB, this will
1869 result in alignment to 2k sectors and typical Flash/SSD sectors,
1870 which is highly desirable for a number of reasons. Changing the
1871 alignment is not recommended.
1873 That said, with default parameters, the data area starts at
1874 exactly 2MB offset (at 0x101000 for cryptsetup versions before
1875 1.3). The smallest data area you can have is one sector of 512
1876 bytes. Data areas of 0 bytes can be created, but fail on mapping.
1878 While you cannot put a filesystem into something this small, it may
1879 still be used to contain, for example, key. Note that with current
1880 formatting tools, a partition for a container this size will be
1881 3MiB anyways. If you put the LUKS container into a file (via
1882 losetup and a loopback device), the file needs to be 2097664 bytes
1883 in size, i.e. 2MiB + 512B.
1885 There two ways to influence the start of the data area are key-size
1888 For alignment, you can go down to 1 on the parameter. This will
1889 still leave you with a data-area starting at 0x101000, i.e.
1890 1MiB+4096B (default parameters) as alignment will be rounded up to
1891 the next multiple of 8 (i.e. 4096 bytes) If in doubt, do a dry-run
1892 on a larger file and dump the LUKS header to get actual
1895 For key-size, you can use 128 bit (e.g. AES-128 with CBC), 256 bit
1896 (e.g. AES-256 with CBC) or 512 bit (e.g. AES-256 with XTS mode).
1897 You can do 64 bit (e.g. blowfish-64 with CBC), but anything below
1898 128 bit has to be considered insecure today.
1900 Example 1 - AES 128 bit with CBC:
1902 cryptsetup luksFormat -s 128 --align-payload=8 <device>
1904 This results in a data offset of 0x81000, i.e. 516KiB or 528384
1905 bytes. Add one 512 byte sector and the smallest LUKS container size
1906 with these parameters is 516KiB + 512B or 528896 bytes.
1908 Example 2 - Blowfish 64 bit with CBC (WARNING: insecure):
1910 cryptsetup luksFormat -c blowfish -s 64 --align-payload=8 /dev/loop0
1912 This results in a data offset of 0x41000, i.e. 260kiB or 266240
1913 bytes, with a minimal LUKS container size of 260kiB + 512B or
1917 * 6.14 I think this is overly complicated. Is there an alternative?
1919 Not really. Encryption comes at a price. You can use plain
1920 dm-crypt to simplify things a bit. It does not allow multiple
1921 passphrases, but on the plus side, it has zero on disk description
1922 and if you overwrite some part of a plain dm-crypt partition,
1923 exactly the overwritten parts are lost (rounded up to sector
1927 * 6.15 Can I clone a LUKS container?
1929 You can, but it breaks security, because the cloned container has
1930 the same header and hence the same master key. You cannot change
1931 the master key on a LUKS container, even if you change the
1932 passphrase(s), the master key stays the same. That means whoever
1933 has access to one of the clones can decrypt them all, completely
1934 bypassing the passphrases.
1936 The right way to do this is to first luksFormat the target
1937 container, then to clone the contents of the source container, with
1938 both containers mapped, i.e. decrypted. You can clone the decrypted
1939 contents of a LUKS container in binary mode, although you may run
1940 into secondary issues with GUIDs in filesystems, partition tables,
1941 RAID-components and the like. These are just the normal problems
1942 binary cloning causes.
1944 Note that if you need to ship (e.g.) cloned LUKS containers with a
1945 default passphrase, that is fine as long as each container was
1946 individually created (and hence has its own master key). In this
1947 case, changing the default passphrase will make it secure again.
1950 7. Interoperability with other Disk Encryption Tools
1953 * 7.1 What is this section about?
1955 Cryptsetup for plain dm-crypt can be used to access a number of
1956 on-disk formats created by tools like loop-aes patched into
1957 losetup. This sometimes works and sometimes does not. This
1958 section collects insights into what works, what does not and where
1959 more information is required.
1961 Additional information may be found in the mailing-list archives,
1962 mentioned at the start of this FAQ document. If you have a
1963 solution working that is not yet documented here and think a wider
1964 audience may be interested, please email the FAQ maintainer.
1967 * 7.2 loop-aes: General observations.
1969 One problem is that there are different versions of losetup around.
1970 loop-aes is a patch for losetup. Possible problems and deviations
1971 from cryptsetup option syntax include:
1973 - Offsets specified in bytes (cryptsetup: 512 byte sectors)
1975 - The need to specify an IV offset
1977 - Encryption mode needs specifying (e.g. "-c twofish-cbc-plain")
1979 - Key size needs specifying (e.g. "-s 128" for 128 bit keys)
1981 - Passphrase hash algorithm needs specifying
1983 Also note that because plain dm-crypt and loop-aes format does not
1984 have metadata, and while the loopAES extension for cryptsetup tries
1985 autodetection (see command loopaesOpen), it may not always work.
1986 If you still have the old set-up, using a verbosity option (-v)
1987 on mapping with the old tool or having a look into the system logs
1988 after setup could give you the information you need. Below, there
1989 are also some things that worked for somebody.
1992 * 7.3 loop-aes patched into losetup on Debian 5.x, kernel 2.6.32
1994 In this case, the main problem seems to be that this variant of
1995 losetup takes the offset (-o option) in bytes, while cryptsetup
1996 takes it in sectors of 512 bytes each. Example: The losetup command
1998 losetup -e twofish -o 2560 /dev/loop0 /dev/sdb1
1999 mount /dev/loop0 mount-point
2003 cryptsetup create -c twofish -o 5 --skip 5 e1 /dev/sdb1
2004 mount /dev/mapper/e1 mount-point
2007 * 7.4 loop-aes with 160 bit key
2009 This seems to be sometimes used with twofish and blowfish and
2010 represents a 160 bit ripemed160 hash output padded to 196 bit key
2011 length. It seems the corresponding options for cryptsetup are
2013 --cipher twofish-cbc-null -s 192 -h ripemd160:20
2016 * 7.5 loop-aes v1 format OpenSUSE
2018 Apparently this is done by older OpenSUSE distros and stopped
2019 working from OpenSUSE 12.1 to 12.2. One user had success with the
2022 cryptsetup create <target> <device> -c aes -s 128 -h sha256
2025 * 7.6 Kernel encrypted loop device (cryptoloop)
2027 There are a number of different losetup implementations for using
2028 encrypted loop devices so getting this to work may need a bit of
2031 NOTE: Do NOT use this for new containers! Some of the existing
2032 implementations are insecure and future support is uncertain.
2034 Example for a compatible mapping:
2036 losetup -e twofish -N /dev/loop0 /image.img
2040 cryptsetup create image_plain /image.img -c twofish-cbc-plain -H plain
2042 with the mapping being done to /dev/mapper/image_plain instead of
2047 Cipher, mode and pasword hash (or no hash):
2049 -e cipher [-N] => -c cipher-cbc-plain -H plain [-s 256]
2050 -e cipher => -c cipher-cbc-plain -H ripemd160 [-s 256]
2052 Key size and offsets (losetup: bytes, cryptsetuop: sectors of 512
2056 -o 2560 => -o 5 -p 5 # 2560/512 = 5
2058 There is no replacement for --pass-fd, it has to be emulated using
2059 keyfiles, see the cryptsetup man-page.
2062 8. Issues with Specific Versions of cryptsetup
2065 * 8.1 When using the create command for plain dm-crypt with
2066 cryptsetup 1.1.x, the mapping is incompatible and my data is not
2069 With cryptsetup 1.1.x, the distro maintainer can define different
2070 default encryption modes for LUKS and plain devices. You can check
2071 these compiled-in defaults using "cryptsetup --help". Moreover, the
2072 plain device default changed because the old IV mode was
2073 vulnerable to a watermarking attack.
2075 If you are using a plain device and you need a compatible mode, just
2076 specify cipher, key size and hash algorithm explicitly. For
2077 compatibility with cryptsetup 1.0.x defaults, simple use the
2080 cryptsetup create -c aes-cbc-plain -s 256 -h ripemd160 <name> <dev>
2082 LUKS stores cipher and mode in the metadata on disk, avoiding this
2086 * 8.2 cryptsetup on SLED 10 has problems...
2088 SLED 10 is missing an essential kernel patch for dm-crypt, which
2089 is broken in its kernel as a result. There may be a very old
2090 version of cryptsetup (1.0.x) provided by SLED, which should also
2091 not be used anymore as well. My advice would be to drop SLED 10.
2094 9. References and Further Reading
2097 * Purpose of this Section
2099 The purpose of this section is to collect references to all
2100 materials that do not fit the FAQ but are relevant in some fashion.
2101 This can be core topics like the LUKS spec or disk encryption, but
2102 it can also be more tangential, like secure storage management or
2103 cryptography used in LUKS. It should still have relevance to
2104 cryptsetup and its applications.
2106 If you wan to see something added here, send email to the
2107 maintainer (or the cryptsetup mailing list) giving an URL, a
2108 description (1-3 lines preferred) and a section to put it in. You
2109 can also propose new sections.
2111 At this time I would like to limit the references to things that
2112 are available on the web.
2117 - LUKS on-disk format spec:
2118 http://code.google.com/p/cryptsetup/wiki/Specification
2123 - Some code examples are in the source package under docs/examples
2126 * Brute-forcing passphrases
2129 http://news.electricalchemy.net/2009/10/password-cracking-in-cloud-part-5.html
2132 http://it.slashdot.org/story/12/12/05/0623215/new-25-gpu-monster-devours-strong-passwords-in-minutes
2138 * SSD and Flash Disk Related
2144 * Attacks Against Disk Encryption
2147 * Risk Management as Relevant for Disk Encryption
2155 A. Contributors In no particular order: