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 NO WARNING ON NON-INTERACTIVE FORMAT: If you feed cryptsetup from
68 STDIN (e.g. via GnuPG) on LUKS format, it does not give you the
69 warning that you are about to format (and e.g. will lose any
70 pre-existing LUKS container on the target), as it assumes it is
71 used from a script. In this scenario, the responsibility for
72 warning the user and possibly checking for an existing LUKS header
73 is shifted to the script. This is a more general form of the
76 LUKS PASSPHRASE IS NOT THE MASTER KEY: The LUKS passphrase is not
77 used in deriving the master key. It is used in decrypting a master
78 key that is randomly selected on header creation. This means that
79 if you create a new LUKS header on top of an old one with
80 exactly the same parameters and exactly the same passphrase as the
81 old one, it will still have a different master key and your data
82 will be permanently lost.
84 PASSPHRASE CHARACTER SET: Some people have had difficulties with
85 this when upgrading distributions. It is highly advisable to only
86 use the 95 printable characters from the first 128 characters of
87 the ASCII table, as they will always have the same binary
88 representation. Other characters may have different encoding
89 depending on system configuration and your passphrase will not
90 work with a different encoding. A table of the standardized first
91 128 ASCII characters can, e.g. be found on
92 http://en.wikipedia.org/wiki/ASCII
95 * 1.3 System specific warnings
97 - Ubuntu as of 4/2011: It seems the installer offers to create
98 LUKS partitions in a way that several people mistook for an offer
99 to activate their existing LUKS partition. The installer gives no
100 or an inadequate warning and will destroy your old LUKS header,
101 causing permanent data loss. See also the section on Backup and
104 This issue has been acknowledged by the Ubuntu dev team, see here:
105 http://launchpad.net/bugs/420080
107 Update 7/2012: I am unsure whether this has been fixed by now, best
111 * 1.4 My LUKS-device is broken! Help!
113 First: Do not panic! In many cases the data is still recoverable.
114 Do not do anything hasty! Steps:
116 - Take some deep breaths. Maybe add some relaxing music. This may
117 sound funny, but I am completely serious. Often, critical damage is
118 done only after the initial problem.
120 - Do not reboot. The keys mays still be in the kernel if the device
123 - Make sure others do not reboot the system.
125 - Do not write to your disk without a clear understanding why this
126 will not make matters worse. Do a sector-level backup before any
127 writes. Often you do not need to write at all to get enough access
128 to make a backup of the data.
132 - Read section 6 of this FAQ.
134 - Ask on the mailing-list if you need more help.
137 * 1.5 Who wrote this?
139 Current FAQ maintainer is Arno Wagner <arno@wagner.name>. If you
140 want to send me encrypted email, my current PGP key is DSA key
141 CB5D9718, fingerprint 12D6 C03B 1B30 33BB 13CF B774 E35C 5FA1 CB5D
144 Other contributors are listed at the end. If you want to contribute,
145 send your article, including a descriptive headline, to the
146 maintainer, or the dm-crypt mailing list with something like "FAQ
147 ..." in the subject. You can also send more raw information and
148 have me write the section. Please note that by contributing to this
149 FAQ, you accept the license described below.
151 This work is under the "Attribution-Share Alike 3.0 Unported"
152 license, which means distribution is unlimited, you may create
153 derived works, but attributions to original authors and this
154 license statement must be retained and the derived work must be
155 under the same license. See
156 http://creativecommons.org/licenses/by-sa/3.0/ for more details of
159 Side note: I did text license research some time ago and I think
160 this license is best suited for the purpose at hand and creates the
164 * 1.5 Where is the project website?
166 There is the project website at http://code.google.com/p/cryptsetup/
167 Please do not post questions there, nobody will read them. Use
168 the mailing-list instead.
171 * 1.6 Is there a mailing-list?
173 Instructions on how to subscribe to the mailing-list are at on the
174 project website. People are generally helpful and friendly on the
177 The question of how to unsubscribe from the list does crop up
178 sometimes. For this you need your list management URL, which is
179 sent to you initially and once at the start of each month. Go to
180 the URL mentioned in the email and select "unsubscribe". This page
181 also allows you to request a password reminder.
183 Alternatively, you can send an Email to dm-crypt-request@saout.de
184 with just the word "help" in the subject or message body. Make sure
185 to send it from your list address.
187 The mailing list archive is here:
188 http://dir.gmane.org/gmane.linux.kernel.device-mapper.dm-crypt
191 * 1.7 Unsubscribe from the mailing-list
193 Send mail to dm-crypt-unsubscribe@saout.de from the subscribed
194 account. You will get an email with instructions.
196 Basically, you just have to respond to it unmodified to get
197 unsubscribed. The listserver admin functions are not very fast. It
198 can take 15 minutes or longer for a reply to arrive (I suspect
199 greylisting is in use), so be patient.
201 Also note that nobody on the list can unsubscribe you, sending
202 demands to be unsubscribed to the list just annoys people that are
203 entirely blameless for you being subscribed.
205 If you are subscribed, a subscription confirmation email was sent
206 to your email account and it had to be answered before the
207 subscription went active. The confirmation emails from the
208 listserver have subjects like these (with other numbers):
210 Subject: confirm 9964cf10.....
212 and are sent from dm-crypt-request@saout.de. You should check
213 whether you have anything like it in your sent email folder. If
214 you find nothing and are sure you did not confirm, then you should
215 look into a possible compromise of your email account.
221 * 2.1 LUKS Container Setup mini-HOWTO
223 This item tries to give you a very brief list of all the steps you
224 should go though when creating a new LUKS encrypted container, i.e.
225 encrypted disk, partition or loop-file.
227 01) All data will be lost, if there is data on the target, make a
230 02) Make very sure you have the right target disk, partition or
233 03) If the target was in use previously, it is a good idea to
234 wipe it before creating the LUKS container in order to remove any
235 trace of old file systems and data. For example, some users have
236 managed to run e2fsck on a partition containing a LUKS container,
237 possibly because of residual ext2 superblocks from an earlier use.
238 This can do arbitrary damage up to complete and permanent loss of
239 all data in the LUKS container.
241 To just quickly wipe file systems (old data may remain), use
243 wipefs -a <target device>
245 To wipe file system and data, use something like
247 cat /dev/zero > <target device>
249 This can take a while. To get a progress indicator, you can use
250 the tool dd_rescue (->google) instead or use my stream meter "wcs"
251 (source here: http://www.tansi.org/tools/index.html) in the
254 cat /dev/zero | wcs > <target device>
256 Be very sure you have the right target, all data will be lost!
258 Note that automatic wiping is on the TODO list for cryptsetup, so
259 at some time in the future this will become unnecessary.
261 04) Create the LUKS container:
262 cryptsetup luksFormat <target device>
264 Just follow the on-screen instructions.
266 05) Map the container. Here it will be mapped to /dev/mapper/c1:
267 cryptsetup luksOpen <target device> c1
269 06) (Optionally) wipe the container (make sure you have the right target!):
270 cat /dev/zero > /dev/mapper/c1
272 Note that this creates a small information leak, as an attacker can
273 determine whether a 512 byte block is zero if the attacker has
274 access to the encrypted container multiple times. Typically a
275 competent attacker that has access multiple times can install a
276 passphrase sniffer anyways, so this leakage is not very
277 significant. For getting a progress indicator, see step 03.
279 Note that at some time in the future, cryptsetup will do this for
280 you, but currently it is a TODO list item.
282 07) Create a file system in the mapped container, for example an
283 ext3 file system (any other file system is possible):
285 mke2fs -j /dev/mapper/c1
287 08) Mount your encrypted file system, here on /mnt:
288 mount /dev/mapper/c1 /mnt
290 Done. You can now use the encrypted file system to store data. Be
291 sure to read though the rest of the FAQ, these are just the very
292 basics. In particular, there are a number of mistakes that are
293 easy to make, but will compromise your security.
296 * 2.2 What is the difference between "plain" and LUKS format?
298 First, unless you happen to understand the cryptographic background
299 well, you should use LUKS. It does protect the user from a lot of
300 common mistakes. Plain dm-crypt is for experts.
302 Plain format is just that: It has no metadata on disk, reads all
303 parameters from the commandline (or the defaults), derives a
304 master-key from the passphrase and then uses that to de-/encrypt
305 the sectors of the device, with a direct 1:1 mapping between
306 encrypted and decrypted sectors.
308 Primary advantage is high resilience to damage, as one damaged
309 encrypted sector results in exactly one damaged decrypted sector.
310 Also, it is not readily apparent that there even is encrypted data
311 on the device, as an overwrite with crypto-grade randomness (e.g.
312 from /dev/urandom) looks exactly the same on disk.
314 Side-note: That has limited value against the authorities. In
315 civilized countries, they cannot force you to give up a crypto-key
316 anyways. In quite a few countries around the world, they can force
317 you to give up the keys (using imprisonment or worse to pressure
318 you, sometimes without due process), and in the worst case, they
319 only need a nebulous "suspicion" about the presence of encrypted
320 data. Sometimes this applies to everybody, sometimes only when you
321 are suspected of having "illicit data" (definition subject to
322 change) and sometimes specifically when crossing a border. Note
323 that this is going on in countries like the US and the UK, to
324 different degrees and sometimes with courts restricting what the
325 authorities can actually demand.
327 My advice is to either be ready to give up the keys or to not have
328 encrypted data when traveling to those countries, especially when
329 crossing the borders. The latter also means not having any
330 high-entropy (random) data areas on your disk, unless you can
331 explain them and demonstrate that explanation. Hence doing a
332 zero-wipe of all free space, including unused space, may be a good
335 Disadvantages are that you do not have all the nice features that
336 the LUKS metadata offers, like multiple passphrases that can be
337 changed, the cipher being stored in the metadata, anti-forensic
338 properties like key-slot diffusion and salts, etc..
340 LUKS format uses a metadata header and 8 key-slot areas that are
341 being placed at the beginning of the disk, see below under "What
342 does the LUKS on-disk format looks like?". The passphrases are used
343 to decrypt a single master key that is stored in the anti-forensic
346 Advantages are a higher usability, automatic configuration of
347 non-default crypto parameters, defenses against low-entropy
348 passphrases like salting and iterated PBKDF2 passphrase hashing,
349 the ability to change passphrases, and others.
351 Disadvantages are that it is readily obvious there is encrypted
352 data on disk (but see side note above) and that damage to the
353 header or key-slots usually results in permanent data-loss. See
354 below under "6. Backup and Data Recovery" on how to reduce that
355 risk. Also the sector numbers get shifted by the length of the
356 header and key-slots and there is a loss of that size in capacity
357 (1MB+4096B for defaults and 2MB for the most commonly used
358 non-default XTS mode).
361 * 2.3 Can I encrypt an already existing, non-empty partition to use
364 There is no converter, and it is not really needed. The way to do
365 this is to make a backup of the device in question, securely wipe
366 the device (as LUKS device initialization does not clear away old
367 data), do a luksFormat, optionally overwrite the encrypted device,
368 create a new filesystem and restore your backup on the now
369 encrypted device. Also refer to sections "Security Aspects" and
370 "Backup and Data Recovery".
372 For backup, plain GNU tar works well and backs up anything likely
373 to be in a filesystem.
376 * 2.4 How do I use LUKS with a loop-device?
378 This can be very handy for experiments. Setup is just the same as
379 with any block device. If you want, for example, to use a 100MiB
380 file as LUKS container, do something like this:
382 head -c 100M /dev/zero > luksfile # create empty file
383 losetup /dev/loop0 luksfile # map luksfile to /dev/loop0
384 cryptsetup luksFormat /dev/loop0 # create LUKS on loop device
386 Afterwards just use /dev/loop0 as a you would use a LUKS partition.
387 To unmap the file when done, use "losetup -d /dev/loop0".
390 * 2.5 When I add a new key-slot to LUKS, it asks for a passphrase but
391 then complains about there not being a key-slot with that
394 That is as intended. You are asked a passphrase of an existing
395 key-slot first, before you can enter the passphrase for the new
396 key-slot. Otherwise you could break the encryption by just adding a
397 new key-slot. This way, you have to know the passphrase of one of
398 the already configured key-slots in order to be able to configure a
402 * 2.6 Encryption on top of RAID or the other way round?
404 Unless you have special needs, place encryption between RAID and
405 filesystem, i.e. encryption on top of RAID. You can do it the other
406 way round, but you have to be aware that you then need to give the
407 passphrase for each individual disk and RAID autodetection will
408 not work anymore. Therefore it is better to encrypt the RAID
409 device, e.g. /dev/dm0 .
411 This means that the typical layering looks like this:
423 The big advantage is that you can manage the RAID container just
424 like any RAID container, it does not care that what is in it is
428 * 2.7 How do I read a dm-crypt key from file?
430 Note that the file will still be hashed first, just like keyboard
431 input. Use the --key-file option, like this:
433 cryptsetup create --key-file keyfile e1 /dev/loop0
436 * 2.8 How do I read a LUKS slot key from file?
438 What you really do here is to read a passphrase from file, just as
439 you would with manual entry of a passphrase for a key-slot. You can
440 add a new passphrase to a free key-slot, set the passphrase of an
441 specific key-slot or put an already configured passphrase into a
442 file. In the last case make sure no trailing newline (0x0a) is
443 contained in the key file, or the passphrase will not work because
444 the whole file is used as input.
446 To add a new passphrase to a free key slot from file, use something
449 cryptsetup luksAddKey /dev/loop0 keyfile
451 To add a new passphrase to a specific key-slot, use something like
454 cryptsetup luksAddKey --key-slot 7 /dev/loop0 keyfile
456 To supply a key from file to any LUKS command, use the --key-file
457 option, e.g. like this:
459 cryptsetup luksOpen --key-file keyfile /dev/loop0 e1
462 * 2.9 How do I read the LUKS master key from file?
464 The question you should ask yourself first is why you would want to
465 do this. The only legitimate reason I can think of is if you want
466 to have two LUKS devices with the same master key. Even then, I
467 think it would be preferable to just use key-slots with the same
468 passphrase, or to use plain dm-crypt instead. If you really have a
469 good reason, please tell me. If I am convinced, I will add how to
473 * 2.10 What are the security requirements for a key read from file?
475 A file-stored key or passphrase has the same security requirements
476 as one entered interactively, however you can use random bytes and
477 thereby use bytes you cannot type on the keyboard. You can use any
478 file you like as key file, for example a plain text file with a
479 human readable passphrase. To generate a file with random bytes,
480 use something like this:
482 head -c 256 /dev/random > keyfile
485 * 2.11 If I map a journaled file system using dm-crypt/LUKS, does it
486 still provide its usual transactional guarantees?
488 Yes, it does, unless a very old kernel is used. The required flags
489 come from the filesystem layer and are processed and passed onwards
490 by dm-crypt. A bit more information on the process by which
491 transactional guarantees are implemented can be found here:
493 http://lwn.net/Articles/400541/
495 Please note that these "guarantees" are weaker than they appear to
496 be. One problem is that quite a few disks lie to the OS about
497 having flushed their buffers. Some other things can go wrong as
498 well. The filesystem developers are aware of these problems and
499 typically can make it work anyways. That said, dm-crypt/LUKS will
500 not make things worse.
502 One specific problem you can run into though is that you can get
503 short freezes and other slowdowns due to the encryption layer.
504 Encryption takes time and forced flushes will block for that time.
505 For example, I did run into frequent small freezes (1-2 sec) when
506 putting a vmware image on ext3 over dm-crypt. When I went back to
507 ext2, the problem went away. This seems to have gotten better with
508 kernel 2.6.36 and the reworking of filesystem flush locking
509 mechanism (less blocking of CPU activity during flushes). It
510 should improve further and eventually the problem should go away.
513 * 2.12 Can I use LUKS or cryptsetup with a more secure (external)
514 medium for key storage, e.g. TPM or a smartcard?
516 Yes, see the answers on using a file-supplied key. You do have to
517 write the glue-logic yourself though. Basically you can have
518 cryptsetup read the key from STDIN and write it there with your
519 own tool that in turn gets the key from the more secure key
522 For TPM support, you may want to have a look at tpm-luks at
523 https://github.com/shpedoikal/tpm-luks. Note that tpm-luks is not
524 related to the cryptsetup project.
527 * 2.13 Can I resize a dm-crypt or LUKS partition?
529 Yes, you can, as neither dm-crypt nor LUKS stores partition size.
530 Whether you should is a different question. Personally I recommend
531 backup, recreation of the encrypted partition with new size,
532 recreation of the filesystem and restore. This gets around the
533 tricky business of resizing the filesystem. Resizing a dm-crypt or
534 LUKS container does not resize the filesystem in it. The backup is
535 really non-optional here, as a lot can go wrong, resulting in
536 partial or complete data loss. Using something like gparted to
537 resize an encrypted partition is slow, but typically works. This
538 will not change the size of the filesystem hidden under the
541 You also need to be aware of size-based limitations. The one
542 currently relevant is that aes-xts-plain should not be used for
543 encrypted container sizes larger than 2TiB. Use aes-xts-plain64
547 * 2.14 How do I Benchmark the Ciphers, Hashes and Modes?
549 Since version 1.60 cryptsetup supports the "benchmark" command.
554 It will output first iterations/second for the key-derivation
555 function PBKDF2 parameterized with different hash-functions, and
556 then the raw encryption speed of ciphers with different modes and
557 key-sizes. You can get more than the default benchmarks, see the
558 man-page for the relevant parameters. Note that XTS mode takes two
559 keys, hence the listed key sizes are double that for other modes
560 and half of it is the cipher key, the other half is the XTS key.
563 * 2.15 How do I Verify I have an Authentic cryptsetup Source Package?
565 Current maintainer is Milan Broz and he signs the release packages
566 with his PGP key. The key he currently uses is the "RSA key ID
567 D93E98FC", fingerprint 2A29 1824 3FDE 4664 8D06 86F9 D9B0 577B
568 D93E 98FC. While I have every confidence this really is his key and
569 that he is who he claims to be, don't depend on it if your life is
570 at stake. For that matter, if your life is at stake, don't depend
571 on me being who I claim to be either.
573 That said, as cryptsetup is under good version control, a malicious
574 change should be noticed sooner or later, but it may take a while.
575 Also, the attacker model makes compromising the sources in a
576 non-obvious way pretty hard. Sure, you could put the master-key
577 somewhere on disk, but that is rather obvious as soon as somebody
578 looks as there would be data in an empty LUKS container in a place
579 it should not be. Doing this in a more nefarious way, for example
580 hiding the master-key in the salts, would need a look at the
581 sources to be discovered, but I think that somebody would find that
582 sooner or later as well.
584 That said, this discussion is really a lot more complicated and
585 longer as an FAQ can sustain. If in doubt, ask on the mailing list.
591 * 3.1 My dm-crypt/LUKS mapping does not work! What general steps are
592 there to investigate the problem?
594 If you get a specific error message, investigate what it claims
595 first. If not, you may want to check the following things.
597 - Check that "/dev", including "/dev/mapper/control" is there. If it
598 is missing, you may have a problem with the "/dev" tree itself or
599 you may have broken udev rules.
601 - Check that you have the device mapper and the crypt target in your
602 kernel. The output of "dmsetup targets" should list a "crypt"
603 target. If it is not there or the command fails, add device mapper
604 and crypt-target to the kernel.
606 - Check that the hash-functions and ciphers you want to use are in
607 the kernel. The output of "cat /proc/crypto" needs to list them.
610 * 3.2 My dm-crypt mapping suddenly stopped when upgrading cryptsetup.
612 The default cipher, hash or mode may have changed (the mode changed
613 from 1.0.x to 1.1.x). See under "Issues With Specific Versions of
617 * 3.3 When I call cryptsetup from cron/CGI, I get errors about
620 If you get errors about unknown parameters or the like that are not
621 present when cryptsetup is called from the shell, make sure you
622 have no older version of cryptsetup on your system that then gets
623 called by cron/CGI. For example some distributions install
624 cryptsetup into /usr/sbin, while a manual install could go to
625 /usr/local/sbin. As a debugging aid, call "cryptsetup --version"
626 from cron/CGI or the non-shell mechanism to be sure the right
630 * 3.4 Unlocking a LUKS device takes very long. Why?
632 The iteration time for a key-slot (see Section 5 for an explanation
633 what iteration does) is calculated when setting a passphrase. By
634 default it is 1 second on the machine where the passphrase is set.
635 If you set a passphrase on a fast machine and then unlock it on a
636 slow machine, the unlocking time can be much longer. Also take into
637 account that up to 8 key-slots have to be tried in order to find the
640 If this is problem, you can add another key-slot using the slow
641 machine with the same passphrase and then remove the old key-slot.
642 The new key-slot will have an iteration count adjusted to 1 second
643 on the slow machine. Use luksKeyAdd and then luksKillSlot or
646 However, this operation will not change volume key iteration count
647 (MK iterations in output of "cryptsetup luksDump"). In order to
648 change that, you will have to backup the data in the LUKS
649 container (i.e. your encrypted data), luksFormat on the slow
650 machine and restore the data. Note that in the original LUKS
651 specification this value was fixed to 10, but it is now derived
652 from the PBKDF2 benchmark as well and set to iterations in 0.125
653 sec or 1000, whichever is larger. Also note that MK iterations
654 are not very security relevant. But as each key-slot already takes
655 1 second, spending the additional 0.125 seconds really does not
659 * 3.5 "blkid" sees a LUKS UUID and an ext2/swap UUID on the same
660 device. What is wrong?
662 Some old versions of cryptsetup have a bug where the header does
663 not get completely wiped during LUKS format and an older ext2/swap
664 signature remains on the device. This confuses blkid.
666 Fix: Wipe the unused header areas by doing a backup and restore of
667 the header with cryptsetup 1.1.x:
669 cryptsetup luksHeaderBackup --header-backup-file <file> <device>
670 cryptsetup luksHeaderRestore --header-backup-file <file> <device>
673 * 3.6 cryptsetup segfaults on Gentoo amd64 hardened ...
675 There seems to be some interference between the hardening and and
676 the way cryptsetup benchmarks PBKDF2. The solution to this is
677 currently not quite clear for an encrypted root filesystem. For
678 other uses, you can apparently specify USE="dynamic" as compile
679 flag, see http://bugs.gentoo.org/show_bug.cgi?id=283470
685 * 4.1 I get the error "LUKS keyslot x is invalid." What does that
688 This means that the given keyslot has an offset that points
689 outside the valid keyslot area. Typically, the reason is a
690 corrupted LUKS header because something was written to the start of
691 the device the LUKS container is on. Refer to Section "Backup and
692 Data Recovery" and ask on the mailing list if you have trouble
693 diagnosing and (if still possible) repairing this.
696 * 4.2 I cannot unlock my LUKS container! What could be the problem?
698 First, make sure you have a correct passphrase. Then make sure you
699 have the correct key-map and correct keyboard. And then make sure
700 you have the correct character set and encoding, see also
701 "PASSPHRASE CHARACTER SET" under Section 1.2.
703 If you are sure you are entering the passphrase right, there is the
704 possibility that the respective key-slot has been damaged. There
705 is no way to recover a damaged key-slot, except from a header
706 backup (see Section 6). For security reasons, there is also no
707 checksum in the key-slots that could tell you whether a key-slot has
708 been damaged. The only checksum present allows recognition of a
709 correct passphrase, but that only works if the passphrase is
710 correct and the respective key-slot is intact.
712 In order to find out whether a key-slot is damaged one has to look
713 for "non-random looking" data in it. There is a tool that
714 automatizes this in the cryptsetup distribution from version 1.6.0
715 onwards. It is located in misc/keyslot_checker/. Instructions how
716 to use and how to interpret results are in the README file. Note
717 that this tool requires a libcryptsetup from cryptsetup 1.6.0 or
718 later (which means libcryptsetup.so.4.5.0 or later). If the tool
719 complains about missing functions in libcryptsetup, you likely
720 have an earlier version from your distribution still installed. You
721 can either point the symbolic link(s) from libcryptsetup.so.4 to
722 the new version manually, or you can uninstall the distribution
723 version of cryptsetup and re-install that from cryptsetup >= 1.6.0
727 * 4.3 Can a bad RAM module cause problems?
729 LUKS and dm-crypt can give the RAM quite a workout, especially when
730 combined with software RAID. In particular the combination RAID5 +
731 LUKS + XFS seems to uncover RAM problems that never caused obvious
732 problems before. Symptoms vary, but often the problem manifest
733 itself when copying large amounts of data, typically several times
734 larger than your main memory.
736 Side note: One thing you should always do on large data
737 copy/movements is to run a verify, for example with the "-d"
738 option of "tar" or by doing a set of MD5 checksums on the source
741 find . -type f -exec md5sum \{\} \; > checksum-file
743 and then a "md5sum -c checksum-file" on the other side. If you get
744 mismatches here, RAM is the primary suspect. A lesser suspect is
745 an overclocked CPU. I have found countless hardware problems in
746 verify runs after copying or making backups. Bit errors are much
747 more common than most people think.
749 Some RAM issues are even worse and corrupt structures in one of the
750 layers. This typically results in lockups, CPU state dumps in the
751 system logs, kernel panic or other things. It is quite possible to
752 have the problem with an encrypted device, but not with an
753 otherwise the same unencrypted device. The reason for that is that
754 encryption has an error amplification property: You flip one bit
755 in an encrypted data block, and the decrypted version has half of
756 its bits flipped. This is an important security property for modern
757 ciphers. With the usual modes in cryptsetup (CBC, ESSIV, XTS), you
758 get up to a completely changed 512 byte block per bit error. A
759 corrupt block causes a lot more havoc than the occasionally
760 flipped single bit and can result in various obscure errors.
762 Note, that a verify run on copying between encrypted or
763 unencrypted devices will reliably detect corruption, even when the
764 copying itself did not report any problems. If you find defect
765 RAM, assume all backups and copied data to be suspect, unless you
769 * 4.4 How do I test RAM?
771 First you should know that overclocking often makes memory
772 problems worse. So if you overclock (which I strongly recommend
773 against in a system holding data that has some worth), run the
774 tests with the overclocking active.
776 There are two good options. One is Memtest86+ and the other is
777 "memtester" by Charles Cazabon. Memtest86+ requires a reboot and
778 then takes over the machine, while memtester runs from a
779 root-shell. Both use different testing methods and I have found
780 problems fast with each one that the other needed long to find. I
781 recommend running the following procedure until the first error is
784 - Run Memtest86+ for one cycle
786 - Run memtester for one cycle (shut down as many other applications
789 - Run Memtest86+ for 24h or more
791 - Run memtester for 24h or more
793 If all that does not produce error messages, your RAM may be sound,
794 but I have had one weak bit that Memtest86+ needed around 60 hours
795 to find. If you can reproduce the original problem reliably, a good
796 additional test may be to remove half of the RAM (if you have more
797 than one module) and try whether the problem is still there and if
798 so, try with the other half. If you just have one module, get a
799 different one and try with that. If you do overclocking, reduce
800 the settings to the most conservative ones available and try with
807 * 5.1 How long is a secure passphrase ?
809 This is just the short answer. For more info and explanation of
810 some of the terms used in this item, read the rest of Section 5.
811 The actual recommendation is at the end of this item.
813 First, passphrase length is not really the right measure,
814 passphrase entropy is. For example, a random lowercase letter (a-z)
815 gives you 4.7 bit of entropy, one element of a-z0-9 gives you 5.2
816 bits of entropy, an element of a-zA-Z0-9 gives you 5.9 bits and
817 a-zA-Z0-9!@#$%^&:-+ gives you 6.2 bits. On the other hand, a random
818 English word only gives you 0.6...1.3 bits of entropy per
819 character. Using sentences that make sense gives lower entropy,
820 series of random words gives higher entropy. Do not use sentences
821 that can be tied to you or found on your computer. This type of
822 attack is done routinely today.
824 That said, it does not matter too much what scheme you use, but it
825 does matter how much entropy your passphrase contains, because an
826 attacker has to try on average
828 1/2 * 2^(bits of entropy in passphrase)
830 different passphrases to guess correctly.
832 Historically, estimations tended to use computing time estimates,
833 but more modern approaches try to estimate cost of guessing a
836 As an example, I will try to get an estimate from the numbers in
837 http://it.slashdot.org/story/12/12/05/0623215/new-25-gpu-monster-devours-strong-passwords-in-minutes
838 More references can be found a the end of this document. Note that
839 these are estimates from the defender side, so assuming something
840 is easier than it actually is is fine. An attacker may still have
841 vastly higher cost than estimated here.
843 LUKS uses SHA1 for hasing per default. The claim in the reference is
844 63 billion tries/second for SHA1. We will leave aside the check
845 whether a try actually decrypts a key-slot. Now, the machine has 25
846 GPUs, which I will estimate at an overall lifetime cost of USD/EUR
847 1000 each, and an useful lifetime of 2 years. (This is on the low
848 side.) Disregarding downtime, the machine can then break
850 N = 63*10^9 * 3600 * 24 * 365 * 2 ~ 4*10^18
852 passphrases for EUR/USD 25k. That is one 62 bit passphrase hashed
853 once with SHA1 for EUR/USD 25k. Note that as this can be
854 parallelized, it can be done faster than 2 years with several of
857 For plain dm-crypt (no hash iteration) this is it. This gives (with
858 SHA1, plain dm-crypt default is ripemd160 which seems to be
859 slightly slower than SHA1):
861 Passphrase entropy Cost to break
870 For LUKS, you have to take into account hash iteration in PBKDF2.
871 For a current CPU, there are about 100k iterations (as can be
872 queried with ''cryptsetup luksDump''.
874 The table above then becomes:
876 Passphrase entropy Cost to break
887 To get reasonable security for the next 10 years, it is a good idea
888 to overestimate by a factor of at least 1000.
890 Then there is the question of how much the attacker is willing to
891 spend. That is up to your own security evaluation. For general use,
892 I will assume the attacker is willing to spend up to 1 million
893 EUR/USD. Then we get the following recommendations:
895 Plain dm-crypt: Use > 80 bit. That is e.g. 17 random chars from a-z
896 or a random English sentence of > 135 characters length.
898 LUKS: Use > 65 bit. That is e.g. 14 random chars from a-z or a
899 random English sentence of > 108 characters length.
901 If paranoid, add at least 20 bit. That is roughly four additional
902 characters for random passphrases and roughly 32 characters for a
903 random English sentence.
906 * 5.2 Is LUKS insecure? Everybody can see I have encrypted data!
908 In practice it does not really matter. In most civilized countries
909 you can just refuse to hand over the keys, no harm done. In some
910 countries they can force you to hand over the keys, if they suspect
911 encryption. However the suspicion is enough, they do not have to
912 prove anything. This is for practical reasons, as even the presence
913 of a header (like the LUKS header) is not enough to prove that you
914 have any keys. It might have been an experiment, for example. Or it
915 was used as encrypted swap with a key from /dev/random. So they
916 make you prove you do not have encrypted data. Of course that is
917 just as impossible as the other way round.
919 This means that if you have a large set of random-looking data,
920 they can already lock you up. Hidden containers (encryption hidden
921 within encryption), as possible with Truecrypt, do not help
922 either. They will just assume the hidden container is there and
923 unless you hand over the key, you will stay locked up. Don't have
924 a hidden container? Though luck. Anybody could claim that.
926 Still, if you are concerned about the LUKS header, use plain
927 dm-crypt with a good passphrase. See also Section 2, "What is the
928 difference between "plain" and LUKS format?"
931 * 5.3 Should I initialize (overwrite) a new LUKS/dm-crypt partition?
933 If you just create a filesystem on it, most of the old data will
934 still be there. If the old data is sensitive, you should overwrite
935 it before encrypting. In any case, not initializing will leave the
936 old data there until the specific sector gets written. That may
937 enable an attacker to determine how much and where on the
938 partition data was written. If you think this is a risk, you can
939 prevent this by overwriting the encrypted device (here assumed to
940 be named "e1") with zeros like this:
942 dd_rescue -w /dev/zero /dev/mapper/e1
944 or alternatively with one of the following more standard commands:
946 cat /dev/zero > /dev/mapper/e1
947 dd if=/dev/zero of=/dev/mapper/e1
950 * 5.4 How do I securely erase a LUKS (or other) partition?
952 For LUKS, if you are in a desperate hurry, overwrite the LUKS
953 header and key-slot area. This means overwriting the first
954 (keyslots x stripes x keysize) + offset bytes. For the default
955 parameters, this is the 1'052'672 bytes, i.e. 1MiB + 4096 of the
956 LUKS partition. For 512 bit key length (e.g. for aes-xts-plain with
957 512 bit key) this is 2MiB. (The different offset stems from
958 differences in the sector alignment of the key-slots.) If in doubt,
959 just be generous and overwrite the first 10MB or so, it will likely
960 still be fast enough. A single overwrite with zeros should be
961 enough. If you anticipate being in a desperate hurry, prepare the
962 command beforehand. Example with /dev/sde1 as the LUKS partition
963 and default parameters:
965 head -c 1052672 /dev/zero > /dev/sde1; sync
967 A LUKS header backup or full backup will still grant access to
968 most or all data, so make sure that an attacker does not have
969 access to backups or destroy them as well.
971 If you have time, overwrite the whole LUKS partition with a single
972 pass of zeros. This is enough for current HDDs. For SSDs or FLASH
973 (USB sticks) you may want to overwrite the whole drive several
974 times to be sure data is not retained by wear leveling. This is
975 possibly still insecure as SSD technology is not fully understood
976 in this regard. Still, due to the anti-forensic properties of the
977 LUKS key-slots, a single overwrite of an SSD or FLASH drive could
978 be enough. If in doubt, use physical destruction in addition. Here
979 is a link to some current research results on erasing SSDs and
981 http://www.usenix.org/events/fast11/tech/full_papers/Wei.pdf
983 Keep in mind to also erase all backups.
985 Example for a zero-overwrite erase of partition sde1 done with
988 dd_rescue -w /dev/zero /dev/sde1
991 * 5.5 How do I securely erase a backup of a LUKS partition or header?
993 That depends on the medium it is stored on. For HDD and SSD, use
994 overwrite with zeros. For an SSD or FLASH drive (USB stick), you
995 may want to overwrite the complete SSD several times and use
996 physical destruction in addition, see last item. For re-writable
997 CD/DVD, a single overwrite should also be enough, due to the
998 anti-forensic properties of the LUKS keyslots. For write-once
999 media, use physical destruction. For low security requirements,
1000 just cut the CD/DVD into several parts. For high security needs,
1001 shred or burn the medium. If your backup is on magnetic tape, I
1002 advise physical destruction by shredding or burning, after
1003 overwriting . The problem with magnetic tape is that it has a
1004 higher dynamic range than HDDs and older data may well be
1005 recoverable after overwrites. Also write-head alignment issues can
1006 lead to data not actually being deleted at all during overwrites.
1009 * 5.6 What about backup? Does it compromise security?
1011 That depends. See item 6.7.
1014 * 5.7 Why is all my data permanently gone if I overwrite the LUKS
1017 Overwriting the LUKS header in part or in full is the most common
1018 reason why access to LUKS containers is lost permanently.
1019 Overwriting can be done in a number of fashions, like creating a
1020 new filesystem on the raw LUKS partition, making the raw partition
1021 part of a raid array and just writing to the raw partition.
1023 The LUKS header contains a 256 bit "salt" value and without that no
1024 decryption is possible. While the salt is not secret, it is
1025 key-grade material and cannot be reconstructed. This is a
1026 cryptographically strong "cannot". From observations on the
1027 cryptsetup mailing-list, people typically go though the usual
1028 stages of grief (Denial, Anger, Bargaining, Depression, Acceptance)
1029 when this happens to them. Observed times vary between 1 day and 2
1030 weeks to complete the cycle. Seeking help on the mailing-list is
1031 fine. Even if we usually cannot help with getting back your data,
1032 most people found the feedback comforting.
1034 If your header does not contain an intact salt, best go directly
1035 to the last stage ("Acceptance") and think about what to do now.
1036 There is one exception that I know of: If your LUKS container is
1037 still open, then it may be possible to extract the master key from
1038 the running system. See Item "How do I recover the master key from
1039 a mapped LUKS container?" in Section "Backup and Data Recovery".
1042 * 5.8 What is a "salt"?
1044 A salt is a random key-grade value added to the passphrase before
1045 it is processed. It is not kept secret. The reason for using salts
1046 is as follows: If an attacker wants to crack the password for a
1047 single LUKS container, then every possible passphrase has to be
1048 tried. Typically an attacker will not try every binary value, but
1049 will try words and sentences from a dictionary.
1051 If an attacker wants to attack several LUKS containers with the
1052 same dictionary, then a different approach makes sense: Compute the
1053 resulting slot-key for each dictionary element and store it on
1054 disk. Then the test for each entry is just the slow unlocking with
1055 the slot key (say 0.00001 sec) instead of calculating the slot-key
1056 first (1 sec). For a single attack, this does not help. But if you
1057 have more than one container to attack, this helps tremendously,
1058 also because you can prepare your table before you even have the
1059 container to attack! The calculation is also very simple to
1060 parallelize. You could, for example, use the night-time unused CPU
1061 power of your desktop PCs for this.
1063 This is where the salt comes in. If the salt is combined with the
1064 passphrase (in the simplest form, just appended to it), you
1065 suddenly need a separate table for each salt value. With a
1066 reasonably-sized salt value (256 bit, e.g.) this is quite
1070 * 5.9 Is LUKS secure with a low-entropy (bad) passphrase?
1072 Note: You should only use the 94 printable characters from 7 bit
1073 ASCII code to prevent your passphrase from failing when the
1074 character encoding changes, e.g. because of a system upgrade, see
1075 also the note at the very start of this FAQ under "WARNINGS".
1077 This needs a bit of theory. The quality of your passphrase is
1078 directly related to its entropy (information theoretic, not
1079 thermodynamic). The entropy says how many bits of "uncertainty" or
1080 "randomness" are in you passphrase. In other words, that is how
1081 difficult guessing the passphrase is.
1083 Example: A random English sentence has about 1 bit of entropy per
1084 character. A random lowercase (or uppercase) character has about
1087 Now, if n is the number of bits of entropy in your passphrase and t
1088 is the time it takes to process a passphrase in order to open the
1089 LUKS container, then an attacker has to spend at maximum
1091 attack_time_max = 2^n * t
1093 time for a successful attack and on average half that. There is no
1094 way getting around that relationship. However, there is one thing
1095 that does help, namely increasing t, the time it takes to use a
1096 passphrase, see next FAQ item.
1098 Still, if you want good security, a high-entropy passphrase is the
1099 only option. For example, a low-entropy passphrase can never be
1100 considered secure against a TLA-level (Three Letter Agency level,
1101 i.e. government-level) attacker, no matter what tricks are used in
1102 the key-derivation function. Use at least 64 bits for secret stuff.
1103 That is 64 characters of English text (but only if randomly chosen)
1104 or a combination of 12 truly random letters and digits.
1106 For passphrase generation, do not use lines from very well-known
1107 texts (religious texts, Harry potter, etc.) as they are to easy to
1108 guess. For example, the total Harry Potter has about 1'500'000
1109 words (my estimation). Trying every 64 character sequence starting
1110 and ending at a word boundary would take only something like 20
1111 days on a single CPU and is entirely feasible. To put that into
1112 perspective, using a number of Amazon EC2 High-CPU Extra Large
1113 instances (each gives about 8 real cores), this test costs
1114 currently about 50USD/EUR, but can be made to run arbitrarily fast.
1116 On the other hand, choosing 1.5 lines from, say, the Wheel of Time
1117 is in itself not more secure, but the book selection adds quite a
1118 bit of entropy. (Now that I have mentioned it here, don't use tWoT
1119 either!) If you add 2 or 3 typos or switch some words around, then
1120 this is good passphrase material.
1123 * 5.10 What is "iteration count" and why is decreasing it a bad idea?
1125 Iteration count is the number of PBKDF2 iterations a passphrase is
1126 put through before it is used to unlock a key-slot. Iterations are
1127 done with the explicit purpose to increase the time that it takes
1128 to unlock a key-slot. This provides some protection against use of
1129 low-entropy passphrases.
1131 The idea is that an attacker has to try all possible passphrases.
1132 Even if the attacker knows the passphrase is low-entropy (see last
1133 item), it is possible to make each individual try take longer. The
1134 way to do this is to repeatedly hash the passphrase for a certain
1135 time. The attacker then has to spend the same time (given the same
1136 computing power) as the user per try. With LUKS, the default is 1
1137 second of PBKDF2 hashing.
1139 Example 1: Lets assume we have a really bad passphrase (e.g. a
1140 girlfriends name) with 10 bits of entropy. With the same CPU, an
1141 attacker would need to spend around 500 seconds on average to
1142 break that passphrase. Without iteration, it would be more like
1143 0.0001 seconds on a modern CPU.
1145 Example 2: The user did a bit better and has 32 chars of English
1146 text. That would be about 32 bits of entropy. With 1 second
1147 iteration, that means an attacker on the same CPU needs around 136
1148 years. That is pretty impressive for such a weak passphrase.
1149 Without the iterations, it would be more like 50 days on a modern
1150 CPU, and possibly far less.
1152 In addition, the attacker can both parallelize and use special
1153 hardware like GPUs or FPGAs to speed up the attack. The attack can
1154 also happen quite some time after the luksFormat operation and CPUs
1155 can have become faster and cheaper. For that reason you want a
1156 bit of extra security. Anyways, in Example 1 your are screwed.
1157 In example 2, not necessarily. Even if the attack is faster, it
1158 still has a certain cost associated with it, say 10000 EUR/USD
1159 with iteration and 1 EUR/USD without iteration. The first can be
1160 prohibitively expensive, while the second is something you try
1161 even without solid proof that the decryption will yield something
1164 The numbers above are mostly made up, but show the idea. Of course
1165 the best thing is to have a high-entropy passphrase.
1167 Would a 100 sec iteration time be even better? Yes and no.
1168 Cryptographically it would be a lot better, namely 100 times better.
1169 However, usability is a very important factor for security
1170 technology and one that gets overlooked surprisingly often. For
1171 LUKS, if you have to wait 2 minutes to unlock the LUKS container,
1172 most people will not bother and use less secure storage instead. It
1173 is better to have less protection against low-entropy passphrases
1174 and people actually use LUKS, than having them do without
1175 encryption altogether.
1177 Now, what about decreasing the iteration time? This is generally a
1178 very bad idea, unless you know and can enforce that the users only
1179 use high-entropy passphrases. If you decrease the iteration time
1180 without ensuring that, then you put your users at increased risk,
1181 and considering how rarely LUKS containers are unlocked in a
1182 typical work-flow, you do so without a good reason. Don't do it.
1183 The iteration time is already low enough that users with entropy
1184 low passphrases are vulnerable. Lowering it even further increases
1185 this danger significantly.
1188 * 5.11 Some people say PBKDF2 is insecure?
1190 There is some discussion that a hash-function should have a "large
1191 memory" property, i.e. that it should require a lot of memory to be
1192 computed. This serves to prevent attacks using special programmable
1193 circuits, like FPGAs, and attacks using graphics cards. PBKDF2
1194 does not need a lot of memory and is vulnerable to these attacks.
1195 However, the publication usually referred in these discussions is
1196 not very convincing in proving that the presented hash really is
1197 "large memory" (that may change, email the FAQ maintainer when it
1198 does) and it is of limited usefulness anyways. Attackers that use
1199 clusters of normal PCs will not be affected at all by a "large
1200 memory" property. For example the US Secret Service is known to
1201 use the off-hour time of all the office PCs of the Treasury for
1202 password breaking. The Treasury has about 110'000 employees.
1203 Assuming every one has an office PC, that is significant computing
1204 power, all of it with plenty of memory for computing "large
1205 memory" hashes. Bot-net operators also have all the memory they
1206 want. The only protection against a resourceful attacker is a
1207 high-entropy passphrase, see items 5.9 and 5.10.
1210 * 5.12 What about iteration count with plain dm-crypt?
1212 Simple: There is none. There is also no salting. If you use plain
1213 dm-crypt, the only way to be secure is to use a high entropy
1214 passphrase. If in doubt, use LUKS instead.
1217 * 5.13 Is LUKS with default parameters less secure on a slow CPU?
1219 Unfortunately, yes. However the only aspect affected is the
1220 protection for low-entropy passphrase or master-key. All other
1221 security aspects are independent of CPU speed.
1223 The master key is less critical, as you really have to work at it
1224 to give it low entropy. One possibility is to supply the master key
1225 yourself. If that key is low-entropy, then you get what you
1226 deserve. The other known possibility is to use /dev/urandom for
1227 key generation in an entropy-starved situation (e.g. automatic
1228 installation on an embedded device without network and other entropy
1231 For the passphrase, don't use a low-entropy passphrase. If your
1232 passphrase is good, then a slow CPU will not matter. If you insist
1233 on a low-entropy passphrase on a slow CPU, use something like
1234 "--iter-time=10" or higher and wait a long time on each LUKS unlock
1235 and pray that the attacker does not find out in which way exactly
1236 your passphrase is low entropy. This also applies to low-entropy
1237 passphrases on fast CPUs. Technology can do only so much to
1238 compensate for problems in front of the keyboard.
1241 * 5.14 Why was the default aes-cbc-plain replaced with aes-cbc-essiv?
1243 Note: This item applies both to plain dm-crypt and to LUKS
1245 The problem is that cbc-plain has a fingerprint vulnerability, where
1246 a specially crafted file placed into the crypto-container can be
1247 recognized from the outside. The issue here is that for cbc-plain
1248 the initialization vector (IV) is the sector number. The IV gets
1249 XORed to the first data chunk of the sector to be encrypted. If you
1250 make sure that the first data block to be stored in a sector
1251 contains the sector number as well, the first data block to be
1252 encrypted is all zeros and always encrypted to the same ciphertext.
1253 This also works if the first data chunk just has a constant XOR
1254 with the sector number. By having several shifted patterns you can
1255 take care of the case of a non-power-of-two start sector number of
1258 This mechanism allows you to create a pattern of sectors that have
1259 the same first ciphertext block and signal one bit per sector to the
1260 outside, allowing you to e.g. mark media files that way for
1261 recognition without decryption. For large files this is a
1262 practical attack. For small ones, you do not have enough blocks to
1263 signal and take care of different file starting offsets.
1265 In order to prevent this attack, the default was changed to
1266 cbc-essiv. ESSIV uses a keyed hash of the sector number, with the
1267 encryption key as key. This makes the IV unpredictable without
1268 knowing the encryption key and the watermarking attack fails.
1271 * 5.15 Are there any problems with "plain" IV? What is "plain64"?
1273 First, "plain" and "plain64" are both not secure to use with CBC,
1274 see previous FAQ item.
1276 However there are modes, like XTS, that are secure with "plain" IV.
1277 The next limit is that "plain" is 64 bit, with the upper 32 bit set
1278 to zero. This means that on volumes larger than 2TiB, the IV
1279 repeats, creating a vulnerability that potentially leaks some
1280 data. To avoid this, use "plain64", which uses the full sector
1281 number up to 64 bit. Note that "plain64" requires a kernel >=
1282 2.6.33. Also note that "plain64" is backwards compatible for
1283 volume sizes <= 2TiB, but not for those > 2TiB. Finally, "plain64"
1284 does not cause any performance penalty compared to "plain".
1287 * 5.16 What about XTS mode?
1289 XTS mode is potentially even more secure than cbc-essiv (but only if
1290 cbc-essiv is insecure in your scenario). It is a NIST standard and
1291 used, e.g. in Truecrypt. At the moment, if you want to use it, you
1292 have to specify it manually as "aes-xts-plain", i.e.
1294 cryptsetup -c aes-xts-plain luksFormat <device>
1296 For volumes >2TiB and kernels >= 2.6.33 use "plain64" (see FAQ
1297 item on "plain" and "plain64"):
1299 cryptsetup -c aes-xts-plain64 luksFormat <device>
1301 There is a potential security issue with XTS mode and large blocks.
1302 LUKS and dm-crypt always use 512B blocks and the issue does not
1306 * 5.17 Is LUKS FIPS-140-2 certified?
1308 No. But that is more a problem of FIPS-140-2 than of LUKS. From a
1309 technical point-of-view, LUKS with the right parameters would be
1310 FIPS-140-2 compliant, but in order to make it certified, somebody
1311 has to pay real money for that. And then, whenever cryptsetup is
1312 changed or extended, the certification lapses and has to be
1315 From the aspect of actual security, LUKS with default parameters
1316 should be as good as most things that are FIPS-140-2 certified,
1317 although you may want to make sure to use /dev/random (by
1318 specifying --use-random on luksFormat) as randomness source for
1319 the master key to avoid being potentially insecure in an
1320 entropy-starved situation.
1323 * 5.18 What about Plausible Deniability?
1325 First let me attempt a definition for the case of encrypted
1326 filesystems: Plausible deniability is when you hide encrypted data
1327 inside an encrypted container and it is not possible to prove it is
1328 there. The idea is compelling and on first glance it seems
1329 possible to do it. And from a cryptographic point of view, it
1330 actually is possible.
1332 So, does it work in practice? No, unfortunately. The reasoning used
1333 by its proponents is fundamentally flawed in several ways and the
1334 cryptographic properties fail fatally when colliding with the real
1337 First, why should "I do not have a hidden partition" be any more
1338 plausible than "I forgot my crypto key" or "I wiped that partition
1339 with random data, nothing in there"? I do not see any reason.
1341 Second, there are two types of situations: Either they cannot force
1342 you to give them the key (then you simply do not) or the can. In
1343 the second case, they can always do bad things to you, because they
1344 cannot prove that you have the key in the first place! This means
1345 they do not have to prove you have the key, or that this random
1346 looking data on your disk is actually encrypted data. So the
1347 situation will allow them to waterboard/lock-up/deport you
1348 anyways, regardless of how "plausible" your deniability is. Do not
1349 have a hidden partition you could show to them, but there are
1350 indications you may? Too bad for you. Unfortunately "plausible
1351 deniability" also means you cannot prove there is no hidden data.
1353 Third, hidden partitions are not that hidden. There are basically
1354 just two possibilities: a) Make a large crypto container, but put a
1355 smaller filesystem in there and put the hidden partition into the
1356 free space. Unfortunately this is glaringly obvious and can be
1357 detected in an automated fashion. This means that the initial
1358 suspicion to put you under duress in order to make you reveal you
1359 hidden data is given. b) Make a filesystem that spans the whole
1360 encrypted partition, and put the hidden partition into space not
1361 currently used by that filesystem. Unfortunately that is also
1362 glaringly obvious, as you then cannot write to the filesystem
1363 without a high risk of destroying data in the hidden container.
1364 Have not written anything to the encrypted filesystem in a while?
1365 Too bad, they have the suspicion they need to do unpleasant things
1368 To be fair, if you prepare option b) carefully and directly before
1369 going into danger, it may work. But then, the mere presence of
1370 encrypted data may already be enough to get you into trouble in
1371 those places were they can demand encryption keys.
1373 Here is an additional reference for some problems with plausible
1374 deniability: http://www.schneier.com/paper-truecrypt-dfs.pdf I
1375 strongly suggest you read it.
1377 So, no, I will not provide any instructions on how to do it with
1378 plain dm-crypt or LUKS. If you insist on shooting yourself in the
1379 foot, you can figure out how to do it yourself.
1382 * 5.19 What about SSDs, Flash and Hybrid Drives?
1384 The problem is that you cannot reliably erase parts of these
1385 devices, mainly due to wear-leveling and possibly defect
1388 Basically, when overwriting a sector (of 512B), what the device
1389 does is to move an internal sector (may be 128kB or even larger) to
1390 some pool of discarded, not-yet erased unused sectors, take a
1391 fresh empty sector from the empty-sector pool and copy the old
1392 sector over with the changes to the small part you wrote. This is
1393 done in some fashion so that larger writes do not cause a lot of
1394 small internal updates.
1396 The thing is that the mappings between outside-addressable sectors
1397 and inside sectors is arbitrary (and the vendors are not talking).
1398 Also the discarded sectors are not necessarily erased immediately.
1399 They may linger a long time.
1401 For plain dm-crypt, the consequences are that older encrypted data
1402 may be lying around in some internal pools of the device. Thus may
1403 or may not be a problem and depends on the application. Remember
1404 the same can happen with a filesystem if consecutive writes to the
1405 same area of a file can go to different sectors.
1407 However, for LUKS, the worst case is that key-slots and LUKS
1408 header may end up in these internal pools. This means that password
1409 management functionality is compromised (the old passwords may
1410 still be around, potentially for a very long time) and that fast
1411 erase by overwriting the header and key-slot area is insecure.
1413 Also keep in mind that the discarded/used pool may be large. For
1414 example, a 240GB SSD has about 16GB of spare area in the chips that
1415 it is free to do with as it likes. You would need to make each
1416 individual key-slot larger than that to allow reliable overwriting.
1417 And that assumes the disk thinks all other space is in use.
1418 Reading the internal pools using forensic tools is not that hard,
1419 but may involve some soldering.
1423 If you trust the device vendor (you probably should not...) you can
1424 try an ATA "secure erase" command for SSDs. That does not work for
1425 USB keys though and may or may not be secure for a hybrid drive. If
1426 it finishes on an SSD after a few seconds, it was possibly faked.
1427 UNfortunately, for hybrid drives that indicator does not work, as
1428 the drive may well take the time to dully erase the magnetic part,
1429 but only mark the SSD/Flash part as erased while data is still in
1432 If you can do without password management and are fine with doing
1433 physical destruction for permanently deleting data (always after
1434 one or several full overwrites!), you can use plain dm-crypt or
1437 If you want or need the original LUKS security features to work,
1438 you can use a detached LUKS header and put that on a conventional,
1439 magnetic disk. That leaves potentially old encrypted data in the
1440 pools on the disk, but otherwise you get LUKS with the same
1441 security as on a magnetic disk.
1443 If you are concerned about your laptop being stolen, you are likely
1444 fine using LUKS on an SSD or hybrid drive. An attacker would need
1445 to have access to an old passphrase (and the key-slot for this old
1446 passphrase would actually need to still be somewhere in the SSD)
1447 for your data to be at risk. So unless you pasted your old
1448 passphrase all over the Internet or the attacker has knowledge of
1449 it from some other source and does a targeted laptop theft to get
1450 at your data, you should be fine.
1453 6. Backup and Data Recovery
1456 * 6.1 Why do I need Backup?
1458 First, disks die. The rate for well-treated (!) disk is about 5%
1459 per year, which is high enough to worry about. There is some
1460 indication that this may be even worse for some SSDs. This applies
1461 both to LUKS and plain dm-crypt partitions.
1463 Second, for LUKS, if anything damages the LUKS header or the
1464 key-stripe area then decrypting the LUKS device can become
1465 impossible. This is a frequent occurrence. For example an
1466 accidental format as FAT or some software overwriting the first
1467 sector where it suspects a partition boot sector typically makes a
1468 LUKS partition permanently inaccessible. See more below on LUKS
1471 So, data-backup in some form is non-optional. For LUKS, you may
1472 also want to store a header backup in some secure location. This
1473 only needs an update if you change passphrases.
1476 * 6.2 How do I backup a LUKS header?
1478 While you could just copy the appropriate number of bytes from the
1479 start of the LUKS partition, the best way is to use command option
1480 "luksHeaderBackup" of cryptsetup. This protects also against
1481 errors when non-standard parameters have been used in LUKS
1482 partition creation. Example:
1485 cryptsetup luksHeaderBackup --header-backup-file <file> <device>
1487 To restore, use the inverse command, i.e.
1489 cryptsetup luksHeaderRestore --header-backup-file <file> <device>
1492 * 6.3 How do I test a LUKS header?
1496 cryptsetup -v isLuks <device>
1498 on the device. Without the "-v" it just signals its result via
1499 exit-status. You can also use the more general test
1503 which will also detect other types and give some more info. Omit
1504 "-p" for old versions of blkid that do not support it.
1507 * 6.4 How do I backup a LUKS or dm-crypt partition?
1509 There are two options, a sector-image and a plain file or
1510 filesystem backup of the contents of the partition. The sector
1511 image is already encrypted, but cannot be compressed and contains
1512 all empty space. The filesystem backup can be compressed, can
1513 contain only part of the encrypted device, but needs to be
1514 encrypted separately if so desired.
1516 A sector-image will contain the whole partition in encrypted form,
1517 for LUKS the LUKS header, the keys-slots and the data area. It can
1518 be done under Linux e.g. with dd_rescue (for a direct image copy)
1519 and with "cat" or "dd". Example:
1521 cat /dev/sda10 > sda10.img
1522 dd_rescue /dev/sda10 sda10.img
1524 You can also use any other backup software that is capable of making
1525 a sector image of a partition. Note that compression is
1526 ineffective for encrypted data, hence it does not make sense to
1529 For a filesystem backup, you decrypt and mount the encrypted
1530 partition and back it up as you would a normal filesystem. In this
1531 case the backup is not encrypted, unless your encryption method
1532 does that. For example you can encrypt a backup with "tar" as
1535 tar cjf - <path> | gpg --cipher-algo AES -c - > backup.tbz2.gpg
1537 And verify the backup like this if you are at "path":
1539 cat backup.tbz2.gpg | gpg - | tar djf -
1541 Note: Always verify backups, especially encrypted ones.
1543 In both cases GnuPG will ask you interactively for your symmetric
1544 key. The verify will only output errors. Use "tar dvjf -" to get
1545 all comparison results. To make sure no data is written to disk
1546 unencrypted, turn off swap if it is not encrypted before doing the
1549 You can of course use different or no compression and you can use
1550 an asymmetric key if you have one and have a backup of the secret
1551 key that belongs to it.
1553 A second option for a filesystem-level backup that can be used when
1554 the backup is also on local disk (e.g. an external USB drive) is
1555 to use a LUKS container there and copy the files to be backed up
1556 between both mounted containers. Also see next item.
1559 * 6.5 Do I need a backup of the full partition? Would the header and
1560 key-slots not be enough?
1562 Backup protects you against two things: Disk loss or corruption
1563 and user error. By far the most questions on the dm-crypt mailing
1564 list about how to recover a damaged LUKS partition are related
1565 to user error. For example, if you create a new filesystem on a
1566 LUKS partition, chances are good that all data is lost
1569 For this case, a header+key-slot backup would often be enough. But
1570 keep in mind that a well-treated (!) HDD has roughly a failure
1571 risk of 5% per year. It is highly advisable to have a complete
1572 backup to protect against this case.
1575 * *6.6 What do I need to backup if I use "decrypt_derived"?
1577 This is a script in Debian, intended for mounting /tmp or swap with
1578 a key derived from the master key of an already decrypted device.
1579 If you use this for an device with data that should be persistent,
1580 you need to make sure you either do not lose access to that master
1581 key or have a backup of the data. If you derive from a LUKS
1582 device, a header backup of that device would cover backing up the
1583 master key. Keep in mind that this does not protect against disk
1586 Note: If you recreate the LUKS header of the device you derive from
1587 (using luksFormat), the master key changes even if you use the same
1588 passphrase(s) and you will not be able to decrypt the derived
1589 device with the new LUKS header.
1592 * 6.7 Does a backup compromise security?
1594 Depends on how you do it. However if you do not have one, you are
1595 going to eventually lose your encrypted data.
1597 There are risks introduced by backups. For example if you
1598 change/disable a key-slot in LUKS, a binary backup of the partition
1599 will still have the old key-slot. To deal with this, you have to
1600 be able to change the key-slot on the backup as well, securely
1601 erase the backup or do a filesystem-level backup instead of a binary
1604 If you use dm-crypt, backup is simpler: As there is no key
1605 management, the main risk is that you cannot wipe the backup when
1606 wiping the original. However wiping the original for dm-crypt
1607 should consist of forgetting the passphrase and that you can do
1608 without actual access to the backup.
1610 In both cases, there is an additional (usually small) risk with
1611 binary backups: An attacker can see how many sectors and which
1612 ones have been changed since the backup. To prevent this, use a
1613 filesystem level backup method that encrypts the whole backup in
1614 one go, e.g. as described above with tar and GnuPG.
1616 My personal advice is to use one USB disk (low value data) or
1617 three disks (high value data) in rotating order for backups, and
1618 either use independent LUKS partitions on them, or use encrypted
1619 backup with tar and GnuPG.
1621 If you do network-backup or tape-backup, I strongly recommend to
1622 go the filesystem backup path with independent encryption, as you
1623 typically cannot reliably delete data in these scenarios,
1624 especially in a cloud setting. (Well, you can burn the tape if it
1625 is under your control...)
1628 * 6.8 What happens if I overwrite the start of a LUKS partition or
1629 damage the LUKS header or key-slots?
1631 There are two critical components for decryption: The salt values
1632 in the header itself and the key-slots. If the salt values are
1633 overwritten or changed, nothing (in the cryptographically strong
1634 sense) can be done to access the data, unless there is a backup
1635 of the LUKS header. If a key-slot is damaged, the data can still
1636 be read with a different key-slot, if there is a remaining
1637 undamaged and used key-slot. Note that in order to make a key-slot
1638 unrecoverable in a cryptographically strong sense, changing about
1639 4-6 bits in random locations of its 128kiB size is quite enough.
1642 * 6.9 What happens if I (quick) format a LUKS partition?
1644 I have not tried the different ways to do this, but very likely you
1645 will have written a new boot-sector, which in turn overwrites the
1646 LUKS header, including the salts, making your data permanently
1647 irretrievable, unless you have a LUKS header backup. You may also
1648 damage the key-slots in part or in full. See also last item.
1651 * 6.10 How do I recover the master key from a mapped LUKS container?
1653 This is typically only needed if you managed to damage your LUKS
1654 header, but the container is still mapped, i.e. "luksOpen"ed. It
1655 also helps if you have a mapped container that you forgot or do not
1656 know a passphrase for (e.g. on a long running server.)
1658 WARNING: Things go wrong, do a full backup before trying this!
1660 WARNING: This exposes the master key of the LUKS container. Note
1661 that both ways to recreate a LUKS header with the old master key
1662 described below will write the master key to disk. Unless you are
1663 sure you have securely erased it afterwards, e.g. by writing it to
1664 an encrypted partition, RAM disk or by erasing the filesystem you
1665 wrote it to by a complete overwrite, you should change the master
1666 key afterwards. Changing the master key requires a full data
1667 backup, luksFormat and then restore of the backup.
1669 First, there is a script by Milan that automates the whole
1670 process, except generating a new LUKS header with the old master
1671 key (it prints the command for that though):
1673 http://code.google.com/p/cryptsetup/source/browse/misc/luks-header-from-active
1675 You can also do this manually. Here is how:
1677 - Get the master key from the device mapper. This is done by the
1678 following command. Substitute c5 for whatever you mapped to:
1680 # dmsetup table --target crypt --showkey /dev/mapper/c5
1682 0 200704 crypt aes-cbc-essiv:sha256
1683 a1704d9715f73a1bb4db581dcacadaf405e700d591e93e2eaade13ba653d0d09
1686 The result is actually one line, wrapped here for clarity. The long
1687 hex string is the master key.
1689 - Convert the master key to a binary file representation. You can
1690 do this manually, e.g. with hexedit. You can also use the tool
1691 "xxd" from vim like this:
1693 echo "a1704d9....53d0d09" | xxd -r -p > <master-key-file>
1695 - Do a luksFormat to create a new LUKS header.
1697 NOTE: If your header is intact and you just forgot the
1698 passphrase, you can just set a new passphrase, see next
1701 Unmap the device before you do that (luksClose). Then do
1703 cryptsetup luksFormat --master-key-file=<master-key-file> <luks device>
1705 Note that if the container was created with other than the default
1706 settings of the cryptsetup version you are using, you need to give
1707 additional parameters specifying the deviations. If in doubt, try
1708 the script by Milan. It does recover the other parameters as well.
1710 Side note: This is the way the decrypt_derived script gets at the
1711 master key. It just omits the conversion and hashes the master key
1714 - If the header is intact and you just forgot the passphrase, just
1715 set a new passphrase like this:
1717 cryptsetup luksAddKey --master-key-file=<master-key-file> <luks device>
1719 You may want to disable the old one afterwards.
1722 * 6.11 What does the on-disk structure of dm-crypt look like?
1724 There is none. dm-crypt takes a block device and gives encrypted
1725 access to each of its blocks with a key derived from the passphrase
1726 given. If you use a cipher different than the default, you have to
1727 specify that as a parameter to cryptsetup too. If you want to
1728 change the password, you basically have to create a second
1729 encrypted device with the new passphrase and copy your data over.
1730 On the plus side, if you accidentally overwrite any part of a
1731 dm-crypt device, the damage will be limited to the are you
1735 * 6.12 What does the on-disk structure of LUKS look like?
1737 A LUKS partition consists of a header, followed by 8 key-slot
1738 descriptors, followed by 8 key slots, followed by the encrypted
1741 Header and key-slot descriptors fill the first 592 bytes. The
1742 key-slot size depends on the creation parameters, namely on the
1743 number of anti-forensic stripes, key material offset and master
1746 With the default parameters, each key-slot is a bit less than
1747 128kiB in size. Due to sector alignment of the key-slot start,
1748 that means the key block 0 is at offset 0x1000-0x20400, key
1749 block 1 at offset 0x21000-0x40400, and key block 7 at offset
1750 0xc1000-0xe0400. The space to the next full sector address is
1751 padded with zeros. Never used key-slots are filled with what the
1752 disk originally contained there, a key-slot removed with
1753 "luksRemoveKey" or "luksKillSlot" gets filled with 0xff. Due to
1754 2MiB default alignment, start of the data area for cryptsetup 1.3
1755 and later is at 2MiB, i.e. at 0x200000. For older versions, it is
1756 at 0x101000, i.e. at 1'052'672 bytes, i.e. at 1MiB + 4096 bytes
1757 from the start of the partition. Incidentally, "luksHeaderBackup"
1758 for a LUKS container created with default parameters dumps exactly
1759 the first 2MiB (or 1'052'672 bytes for headers created with
1760 cryptsetup versions < 1.3) to file and "luksHeaderRestore" restores
1763 For non-default parameters, you have to figure out placement
1764 yourself. "luksDump" helps. See also next item. For the most common
1765 non-default settings, namely aes-xts-plain with 512 bit key, the
1766 offsets are: 1st keyslot 0x1000-0x3f800, 2nd keyslot
1767 0x40000-0x7e000, 3rd keyslot 0x7e000-0xbd800, ..., and start of
1768 bulk data at 0x200000.
1770 The exact specification of the format is here:
1771 http://code.google.com/p/cryptsetup/wiki/Specification
1774 * 6.13 What is the smallest possible LUKS container?
1776 Note: From cryptsetup 1.3 onwards, alignment is set to 1MB. With
1777 modern Linux partitioning tools that also align to 1MB, this will
1778 result in alignment to 2k sectors and typical Flash/SSD sectors,
1779 which is highly desirable for a number of reasons. Changing the
1780 alignment is not recommended.
1782 That said, with default parameters, the data area starts at
1783 exactly 2MB offset (at 0x101000 for cryptsetup versions before
1784 1.3). The smallest data area you can have is one sector of 512
1785 bytes. Data areas of 0 bytes can be created, but fail on mapping.
1787 While you cannot put a filesystem into something this small, it may
1788 still be used to contain, for example, key. Note that with current
1789 formatting tools, a partition for a container this size will be
1790 3MiB anyways. If you put the LUKS container into a file (via
1791 losetup and a loopback device), the file needs to be 2097664 bytes
1792 in size, i.e. 2MiB + 512B.
1794 There two ways to influence the start of the data area are key-size
1797 For alignment, you can go down to 1 on the parameter. This will
1798 still leave you with a data-area starting at 0x101000, i.e.
1799 1MiB+4096B (default parameters) as alignment will be rounded up to
1800 the next multiple of 8 (i.e. 4096 bytes) If in doubt, do a dry-run
1801 on a larger file and dump the LUKS header to get actual
1804 For key-size, you can use 128 bit (e.g. AES-128 with CBC), 256 bit
1805 (e.g. AES-256 with CBC) or 512 bit (e.g. AES-256 with XTS mode).
1806 You can do 64 bit (e.g. blowfish-64 with CBC), but anything below
1807 128 bit has to be considered insecure today.
1809 Example 1 - AES 128 bit with CBC:
1811 cryptsetup luksFormat -s 128 --align-payload=8 <device>
1813 This results in a data offset of 0x81000, i.e. 516KiB or 528384
1814 bytes. Add one 512 byte sector and the smallest LUKS container size
1815 with these parameters is 516KiB + 512B or 528896 bytes.
1817 Example 2 - Blowfish 64 bit with CBC (WARNING: insecure):
1819 cryptsetup luksFormat -c blowfish -s 64 --align-payload=8 /dev/loop0
1821 This results in a data offset of 0x41000, i.e. 260kiB or 266240
1822 bytes, with a minimal LUKS container size of 260kiB + 512B or
1826 * 6.14 I think this is overly complicated. Is there an alternative?
1828 Not really. Encryption comes at a price. You can use plain
1829 dm-crypt to simplify things a bit. It does not allow multiple
1830 passphrases, but on the plus side, it has zero on disk description
1831 and if you overwrite some part of a plain dm-crypt partition,
1832 exactly the overwritten parts are lost (rounded up to sector
1836 * 6.15 Can I clone a LUKS container?
1838 You can, but it breaks security, because the cloned container has
1839 the same header and hence the same master key. You cannot change
1840 the master key on a LUKS container, even if you change the
1841 passphrase(s), the master key stays the same. That means whoever
1842 has access to one of the clones can decrypt them all, completely
1843 bypassing the passphrases.
1845 The right way to do this is to first luksFormat the target
1846 container, then to clone the contents of the source container, with
1847 both containers mapped, i.e. decrypted. You can clone the decrypted
1848 contents of a LUKS container in binary mode, although you may run
1849 into secondary issues with GUIDs in filesystems, partition tables,
1850 RAID-components and the like. These are just the normal problems
1851 binary cloning causes.
1853 Note that if you need to ship (e.g.) cloned LUKS containers with a
1854 default passphrase, that is fine as long as each container was
1855 individually created (and hence has its own master key). In this
1856 case, changing the default passphrase will make it secure again.
1859 7. Interoperability with other Disk Encryption Tools
1862 * 7.1 What is this section about?
1864 Cryptsetup for plain dm-crypt can be used to access a number of
1865 on-disk formats created by tools like loop-aes patched into
1866 losetup. This sometimes works and sometimes does not. This
1867 section collects insights into what works, what does not and where
1868 more information is required.
1870 Additional information may be found in the mailing-list archives,
1871 mentioned at the start of this FAQ document. If you have a
1872 solution working that is not yet documented here and think a wider
1873 audience may be interested, please email the FAQ maintainer.
1876 * 7.2 loop-aes: General observations.
1878 One problem is that there are different versions of losetup around.
1879 loop-aes is a patch for losetup. Possible problems and deviations
1880 from cryptsetup option syntax include:
1882 - Offsets specified in bytes (cryptsetup: 512 byte sectors)
1884 - The need to specify an IV offset
1886 - Encryption mode needs specifying (e.g. "-c twofish-cbc-plain")
1888 - Key size needs specifying (e.g. "-s 128" for 128 bit keys)
1890 - Passphrase hash algorithm needs specifying
1892 Also note that because plain dm-crypt and loop-aes format does not
1893 have metadata, and while the loopAES extension for cryptsetup tries
1894 autodetection (see command loopaesOpen), it may not always work.
1895 If you still have the old set-up, using a verbosity option (-v)
1896 on mapping with the old tool or having a look into the system logs
1897 after setup could give you the information you need. Below, there
1898 are also some things that worked for somebody.
1901 * 7.3 loop-aes patched into losetup on Debian 5.x, kernel 2.6.32
1903 In this case, the main problem seems to be that this variant of
1904 losetup takes the offset (-o option) in bytes, while cryptsetup
1905 takes it in sectors of 512 bytes each. Example: The losetup command
1907 losetup -e twofish -o 2560 /dev/loop0 /dev/sdb1
1908 mount /dev/loop0 mount-point
1912 cryptsetup create -c twofish -o 5 --skip 5 e1 /dev/sdb1
1913 mount /dev/mapper/e1 mount-point
1916 * 7.4 loop-aes with 160 bit key
1918 This seems to be sometimes used with twofish and blowfish and
1919 represents a 160 bit ripemed160 hash output padded to 196 bit key
1920 length. It seems the corresponding options for cryptsetup are
1922 --cipher twofish-cbc-null -s 192 -h ripemd160:20
1925 * 7.5 loop-aes v1 format OpenSUSE
1927 Apparently this is done by older OpenSUSE distros and stopped
1928 working from OpenSUSE 12.1 to 12.2. One user had success with the
1931 cryptsetup create <target> <device> -c aes -s 128 -h sha256
1934 * 7.6 Kernel encrypted loop device (cryptoloop)
1936 There are a number of different losetup implementations for using
1937 encrypted loop devices so getting this to work may need a bit of
1940 NOTE: Do NOT use this for new containers! Some of the existing
1941 implementations are insecure and future support is uncertain.
1943 Example for a compatible mapping:
1945 losetup -e twofish -N /dev/loop0 /image.img
1949 cryptsetup create image_plain /image.img -c twofish-cbc-plain -H plain
1951 with the mapping being done to /dev/mapper/image_plain instead of
1956 Cipher, mode and pasword hash (or no hash):
1958 -e cipher [-N] => -c cipher-cbc-plain -H plain [-s 256]
1959 -e cipher => -c cipher-cbc-plain -H ripemd160 [-s 256]
1961 Key size and offsets (losetup: bytes, cryptsetuop: sectors of 512
1965 -o 2560 => -o 5 -p 5 # 2560/512 = 5
1967 There is no replacement for --pass-fd, it has to be emulated using
1968 keyfiles, see the cryptsetup man-page.
1971 8. Issues with Specific Versions of cryptsetup
1974 * 8.1 When using the create command for plain dm-crypt with
1975 cryptsetup 1.1.x, the mapping is incompatible and my data is not
1978 With cryptsetup 1.1.x, the distro maintainer can define different
1979 default encryption modes for LUKS and plain devices. You can check
1980 these compiled-in defaults using "cryptsetup --help". Moreover, the
1981 plain device default changed because the old IV mode was
1982 vulnerable to a watermarking attack.
1984 If you are using a plain device and you need a compatible mode, just
1985 specify cipher, key size and hash algorithm explicitly. For
1986 compatibility with cryptsetup 1.0.x defaults, simple use the
1989 cryptsetup create -c aes-cbc-plain -s 256 -h ripemd160 <name> <dev>
1991 LUKS stores cipher and mode in the metadata on disk, avoiding this
1995 * 8.2 cryptsetup on SLED 10 has problems...
1997 SLED 10 is missing an essential kernel patch for dm-crypt, which
1998 is broken in its kernel as a result. There may be a very old
1999 version of cryptsetup (1.0.x) provided by SLED, which should also
2000 not be used anymore as well. My advice would be to drop SLED 10.
2003 9. References and Further Reading
2006 * Purpose of this Section
2008 The purpose of this section is to collect references to all
2009 materials that do not fit the FAQ but are relevant in some fashion.
2010 This can be core topics like the LUKS spec or disk encryption, but
2011 it can also be more tangential, like secure storage management or
2012 cryptography used in LUKS. It should still have relevance to
2013 cryptsetup and its applications.
2015 If you wan to see something added here, send email to the
2016 maintainer (or the cryptsetup mailing list) giving an URL, a
2017 description (1-3 lines preferred) and a section to put it in. You
2018 can also propose new sections.
2020 At this time I would like to limit the references to things that
2021 are available on the web.
2026 - LUKS on-disk format spec:
2027 http://code.google.com/p/cryptsetup/wiki/Specification
2032 - Some code examples are in the source package under docs/examples
2035 * Brute-forcing passphrases
2038 http://news.electricalchemy.net/2009/10/password-cracking-in-cloud-part-5.html
2041 http://it.slashdot.org/story/12/12/05/0623215/new-25-gpu-monster-devours-strong-passwords-in-minutes
2047 * SSD and Flash Disk Related
2053 * Attacks Against Disk Encryption
2056 * Risk Management as Relevant for Disk Encryption
2064 A. Contributors In no particular order: