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 5.17.
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>. Other
140 contributors are listed at the end. If you want to contribute, send
141 your article, including a descriptive headline, to the maintainer,
142 or the dm-crypt mailing list with something like "FAQ ..." in the
143 subject. You can also send more raw information and have me write
144 the section. Please note that by contributing to this FAQ, you
145 accept the license described below.
147 This work is under the "Attribution-Share Alike 3.0 Unported"
148 license, which means distribution is unlimited, you may create
149 derived works, but attributions to original authors and this
150 license statement must be retained and the derived work must be
151 under the same license. See
152 http://creativecommons.org/licenses/by-sa/3.0/ for more details of
155 Side note: I did text license research some time ago and I think
156 this license is best suited for the purpose at hand and creates the
160 * 1.5 Where is the project website?
162 There is the project website at http://code.google.com/p/cryptsetup/
163 Please do not post questions there, nobody will read them. Use
164 the mailing-list instead.
167 * 1.6 Is there a mailing-list?
169 Instructions on how to subscribe to the mailing-list are at on the
170 project website. People are generally helpful and friendly on the
173 The question of how to unsubscribe from the list does crop up
174 sometimes. For this you need your list management URL, which is
175 sent to you initially and once at the start of each month. Go to
176 the URL mentioned in the email and select "unsubscribe". This page
177 also allows you to request a password reminder.
179 Alternatively, you can send an Email to dm-crypt-request@saout.de
180 with just the word "help" in the subject or message body. Make sure
181 to send it from your list address.
183 The mailing list archive is here:
184 http://dir.gmane.org/gmane.linux.kernel.device-mapper.dm-crypt
190 * 2.1 What is the difference between "plain" and LUKS format?
192 First, unless you happen to understand the cryptographic background
193 well, you should use LUKS. It does protect the user from a lot of
194 common mistakes. Plain dm-crypt is for experts.
196 Plain format is just that: It has no metadata on disk, reads all
197 parameters from the commandline (or the defaults), derives a
198 master-key from the passphrase and then uses that to de-/encrypt
199 the sectors of the device, with a direct 1:1 mapping between
200 encrypted and decrypted sectors.
202 Primary advantage is high resilience to damage, as one damaged
203 encrypted sector results in exactly one damaged decrypted sector.
204 Also, it is not readily apparent that there even is encrypted data
205 on the device, as an overwrite with crypto-grade randomness (e.g.
206 from /dev/urandom) looks exactly the same on disk.
208 Side-note: That has limited value against the authorities. In
209 civilized countries, they cannot force you to give up a crypto-key
210 anyways. In the US, the UK and dictatorships around the world,
211 they can force you to give up the keys (using imprisonment or worse
212 to pressure you), and in the worst case, they only need a
213 nebulous "suspicion" about the presence of encrypted data. My
214 advice is to either be ready to give up the keys or to not have
215 encrypted data when traveling to those countries, especially when
216 crossing the borders.
218 Disadvantages are that you do not have all the nice features that
219 the LUKS metadata offers, like multiple passphrases that can be
220 changed, the cipher being stored in the metadata, anti-forensic
221 properties like key-slot diffusion and salts, etc..
223 LUKS format uses a metadata header and 8 key-slot areas that are
224 being placed at the beginning of the disk, see below under "What
225 does the LUKS on-disk format looks like?". The passphrases are used
226 to decrypt a single master key that is stored in the anti-forensic
229 Advantages are a higher usability, automatic configuration of
230 non-default crypto parameters, defenses against low-entropy
231 passphrases like salting and iterated PBKDF2 passphrase hashing,
232 the ability to change passphrases, and others.
234 Disadvantages are that it is readily obvious there is encrypted
235 data on disk (but see side note above) and that damage to the
236 header or key-slots usually results in permanent data-loss. See
237 below under "6. Backup and Data Recovery" on how to reduce that
238 risk. Also the sector numbers get shifted by the length of the
239 header and key-slots and there is a loss of that size in capacity
240 (1MB+4096B for defaults and 2MB for the most commonly used
241 non-default XTS mode).
244 * 2.2 Can I encrypt an already existing, non-empty partition to use
247 There is no converter, and it is not really needed. The way to do
248 this is to make a backup of the device in question, securely wipe
249 the device (as LUKS device initialization does not clear away old
250 data), do a luksFormat, optionally overwrite the encrypted device,
251 create a new filesystem and restore your backup on the now
252 encrypted device. Also refer to sections "Security Aspects" and
253 "Backup and Data Recovery".
255 For backup, plain GNU tar works well and backs up anything likely
256 to be in a filesystem.
259 * 2.3 How do I use LUKS with a loop-device?
261 This can be very handy for experiments. Setup is just the same as
262 with any block device. If you want, for example, to use a 100MiB
263 file as LUKS container, do something like this:
265 head -c 100M /dev/zero > luksfile # create empty file
266 losetup /dev/loop0 luksfile # map luksfile to /dev/loop0
267 cryptsetup luksFormat /dev/loop0 # create LUKS on loop device
269 Afterwards just use /dev/loop0 as a you would use a LUKS partition.
270 To unmap the file when done, use "losetup -d /dev/loop0".
273 * 2.4 When I add a new key-slot to LUKS, it asks for a passphrase but
274 then complains about there not being a key-slot with that
277 That is as intended. You are asked a passphrase of an existing
278 key-slot first, before you can enter the passphrase for the new
279 key-slot. Otherwise you could break the encryption by just adding a
280 new key-slot. This way, you have to know the passphrase of one of
281 the already configured key-slots in order to be able to configure a
285 * 2.5 Encryption on top of RAID or the other way round?
287 Unless you have special needs, place encryption between RAID and
288 filesystem, i.e. encryption on top of RAID. You can do it the other
289 way round, but you have to be aware that you then need to give the
290 passphrase for each individual disk and RAID autodetection will
291 not work anymore. Therefore it is better to encrypt the RAID
292 device, e.g. /dev/dm0 .
295 * 2.6 How do I read a dm-crypt key from file?
297 Note that the file will still be hashed first, just like keyboard
298 input. Use the --key-file option, like this:
300 cryptsetup create --key-file keyfile e1 /dev/loop0
303 * 2.7 How do I read a LUKS slot key from file?
305 What you really do here is to read a passphrase from file, just as
306 you would with manual entry of a passphrase for a key-slot. You can
307 add a new passphrase to a free key-slot, set the passphrase of an
308 specific key-slot or put an already configured passphrase into a
309 file. In the last case make sure no trailing newline (0x0a) is
310 contained in the key file, or the passphrase will not work because
311 the whole file is used as input.
313 To add a new passphrase to a free key slot from file, use something
316 cryptsetup luksAddKey /dev/loop0 keyfile
318 To add a new passphrase to a specific key-slot, use something like
321 cryptsetup luksAddKey --key-slot 7 /dev/loop0 keyfile
323 To supply a key from file to any LUKS command, use the --key-file
324 option, e.g. like this:
326 cryptsetup luksOpen --key-file keyfile /dev/loop0 e1
329 * 2.8 How do I read the LUKS master key from file?
331 The question you should ask yourself first is why you would want to
332 do this. The only legitimate reason I can think of is if you want
333 to have two LUKS devices with the same master key. Even then, I
334 think it would be preferable to just use key-slots with the same
335 passphrase, or to use plain dm-crypt instead. If you really have a
336 good reason, please tell me. If I am convinced, I will add how to
340 * 2.9 What are the security requirements for a key read from file?
342 A file-stored key or passphrase has the same security requirements
343 as one entered interactively, however you can use random bytes and
344 thereby use bytes you cannot type on the keyboard. You can use any
345 file you like as key file, for example a plain text file with a
346 human readable passphrase. To generate a file with random bytes,
347 use something like this:
349 head -c 256 /dev/random > keyfile
352 * 2.10 If I map a journaled file system using dm-crypt/LUKS, does it
353 still provide its usual transactional guarantees?
355 Yes, it does, unless a very old kernel is used. The required flags
356 come from the filesystem layer and are processed and passed onwards
357 by dm-crypt. A bit more information on the process by which
358 transactional guarantees are implemented can be found here:
360 http://lwn.net/Articles/400541/
362 Please note that these "guarantees" are weaker than they appear to
363 be. One problem is that quite a few disks lie to the OS about
364 having flushed their buffers. Some other things can go wrong as
365 well. The filesystem developers are aware of these problems and
366 typically can make it work anyways. That said, dm-crypt/LUKS will
367 not make things worse.
369 One specific problem you can run into though is that you can get
370 short freezes and other slowdowns due to the encryption layer.
371 Encryption takes time and forced flushes will block for that time.
372 For example, I did run into frequent small freezes (1-2 sec) when
373 putting a vmware image on ext3 over dm-crypt. When I went back to
374 ext2, the problem went away. This seems to have gotten better with
375 kernel 2.6.36 and the reworking of filesystem flush locking
376 mechanism (less blocking of CPU activity during flushes). It
377 should improve further and eventually the problem should go away.
380 * 2.11 Can I use LUKS or cryptsetup with a more secure (external)
381 medium for key storage, e.g. TPM or a smartcard?
383 Yes, see the answers on using a file-supplied key. You do have to
384 write the glue-logic yourself though. Basically you can have
385 cryptsetup read the key from STDIN and write it there with your
386 own tool that in turn gets the key from the more secure key
390 * 2.12 Can I resize a dm-crypt or LUKS partition?
392 Yes, you can, as neither dm-crypt nor LUKS stores partition size.
393 Whether you should is a different question. Personally I recommend
394 backup, recreation of the encrypted partition with new size,
395 recreation of the filesystem and restore. This gets around the
396 tricky business of resizing the filesystem. Resizing a dm-crypt or
397 LUKS container does not resize the filesystem in it. The backup is
398 really non-optional here, as a lot can go wrong, resulting in
399 partial or complete data loss. Using something like gparted to
400 resize an encrypted partition is slow, but typically works. This
401 will not change the size of the filesystem hidden under the
404 You also need to be aware of size-based limitations. The one
405 currently relevant is that aes-xts-plain should not be used for
406 encrypted container sizes larger than 2TiB. Use aes-xts-plain64
413 * 3.1 My dm-crypt/LUKS mapping does not work! What general steps are
414 there to investigate the problem?
416 If you get a specific error message, investigate what it claims
417 first. If not, you may want to check the following things.
419 - Check that "/dev", including "/dev/mapper/control" is there. If it
420 is missing, you may have a problem with the "/dev" tree itself or
421 you may have broken udev rules.
423 - Check that you have the device mapper and the crypt target in your
424 kernel. The output of "dmsetup targets" should list a "crypt"
425 target. If it is not there or the command fails, add device mapper
426 and crypt-target to the kernel.
428 - Check that the hash-functions and ciphers you want to use are in
429 the kernel. The output of "cat /proc/crypto" needs to list them.
432 * 3.2 My dm-crypt mapping suddenly stopped when upgrading cryptsetup.
434 The default cipher, hash or mode may have changed (the mode changed
435 from 1.0.x to 1.1.x). See under "Issues With Specific Versions of
439 * 3.3 When I call cryptsetup from cron/CGI, I get errors about
442 If you get errors about unknown parameters or the like that are not
443 present when cryptsetup is called from the shell, make sure you
444 have no older version of cryptsetup on your system that then gets
445 called by cron/CGI. For example some distributions install
446 cryptsetup into /usr/sbin, while a manual install could go to
447 /usr/local/sbin. As a debugging aid, call "cryptsetup --version"
448 from cron/CGI or the non-shell mechanism to be sure the right
452 * 3.4 Unlocking a LUKS device takes very long. Why?
454 The iteration time for a key-slot (see Section 5 for an explanation
455 what iteration does) is calculated when setting a passphrase. By
456 default it is 1 second on the machine where the passphrase is set.
457 If you set a passphrase on a fast machine and then unlock it on a
458 slow machine, the unlocking time can be much longer. Also take into
459 account that up to 8 key-slots have to be tried in order to find the
462 If this is problem, you can add another key-slot using the slow
463 machine with the same passphrase and then remove the old key-slot.
464 The new key-slot will have an iteration count adjusted to 1 second
465 on the slow machine. Use luksKeyAdd and then luksKillSlot or
468 However, this operation will not change volume key iteration count
469 (MK iterations in output of "cryptsetup luksDump"). In order to
470 change that, you will have to backup the data in the LUKS
471 container (i.e. your encrypted data), luksFormat on the slow
472 machine and restore the data. Note that in the original LUKS
473 specification this value was fixed to 10, but it is now derived
474 from the PBKDF2 benchmark as well and set to iterations in 0.125
475 sec or 1000, whichever is larger. Also note that MK iterations
476 are not very security relevant. But as each key-slot already takes
477 1 second, spending the additional 0.125 seconds really does not
481 * 3.5 "blkid" sees a LUKS UUID and an ext2/swap UUID on the same
482 device. What is wrong?
484 Some old versions of cryptsetup have a bug where the header does
485 not get completely wiped during LUKS format and an older ext2/swap
486 signature remains on the device. This confuses blkid.
488 Fix: Wipe the unused header areas by doing a backup and restore of
489 the header with cryptsetup 1.1.x:
491 cryptsetup luksHeaderBackup --header-backup-file <file> <device>
492 cryptsetup luksHeaderRestore --header-backup-file <file> <device>
495 * 3.6 cryptsetup segfaults on Gentoo amd64 hardened ...
497 There seems to be some interference between the hardening and and
498 the way cryptsetup benchmarks PBKDF2. The solution to this is
499 currently not quite clear for an encrypted root filesystem. For
500 other uses, you can apparently specify USE="dynamic" as compile
501 flag, see http://bugs.gentoo.org/show_bug.cgi?id=283470
507 * 4.1 I get the error "LUKS keyslot x is invalid." What does that
510 This means that the given keyslot has an offset that points
511 outside the valid keyslot area. Typically, the reason is a
512 corrupted LUKS header because something was written to the start of
513 the device the LUKS container is on. Refer to Section "Backup and
514 Data Recovery" and ask on the mailing list if you have trouble
515 diagnosing and (if still possible) repairing this.
518 * 4.2 Can a bad RAM module cause problems?
520 LUKS and dm-crypt can give the RAM quite a workout, especially when
521 combined with software RAID. In particular the combination RAID5 +
522 LUKS + XFS seems to uncover RAM problems that never caused obvious
523 problems before. Symptoms vary, but often the problem manifest
524 itself when copying large amounts of data, typically several times
525 larger than your main memory.
527 Side note: One thing you should always do on large data
528 copy/movements is to run a verify, for example with the "-d"
529 option of "tar" or by doing a set of MD5 checksums on the source
532 find . -type f -exec md5sum \{\} \; > checksum-file
534 and then a "md5sum -c checksum-file" on the other side. If you get
535 mismatches here, RAM is the primary suspect. A lesser suspect is
536 an overclocked CPU. I have found countless hardware problems in
537 verify runs after copying or making backups. Bit errors are much
538 more common than most people think.
540 Some RAM issues are even worse and corrupt structures in one of the
541 layers. This typically results in lockups, CPU state dumps in the
542 system logs, kernel panic or other things. It is quite possible to
543 have the problem with an encrypted device, but not with an
544 otherwise the same unencrypted device. The reason for that is that
545 encryption has an error amplification property: You flip one bit
546 in an encrypted data block, and the decrypted version has half of
547 its bits flipped. This is an important security property for modern
548 ciphers. With the usual modes in cryptsetup (CBC, ESSIV, XTS), you
549 get up to a completely changed 512 byte block per bit error. A
550 corrupt block causes a lot more havoc than the occasionally
551 flipped single bit and can result in various obscure errors.
553 Note, that a verify run on copying between encrypted or
554 unencrypted devices will reliably detect corruption, even when the
555 copying itself did not report any problems. If you find defect
556 RAM, assume all backups and copied data to be suspect, unless you
560 * 4.3 How do I test RAM?
562 First you should know that overclocking often makes memory
563 problems worse. So if you overclock (which I strongly recommend
564 against in a system holding data that has some worth), run the
565 tests with the overclocking active.
567 There are two good options. One is Memtest86+ and the other is
568 "memtester" by Charles Cazabon. Memtest86+ requires a reboot and
569 then takes over the machine, while memtester runs from a
570 root-shell. Both use different testing methods and I have found
571 problems fast with each one that the other needed long to find. I
572 recommend running the following procedure until the first error is
575 - Run Memtest86+ for one cycle
577 - Run memtester for one cycle (shut down as many other applications
580 - Run Memtest86+ for 24h or more
582 - Run memtester for 24h or more
584 If all that does not produce error messages, your RAM may be sound,
585 but I have had one weak bit that Memtest86+ needed around 60 hours
586 to find. If you can reproduce the original problem reliably, a good
587 additional test may be to remove half of the RAM (if you have more
588 than one module) and try whether the problem is still there and if
589 so, try with the other half. If you just have one module, get a
590 different one and try with that. If you do overclocking, reduce
591 the settings to the most conservative ones available and try with
598 * 5.1 Is LUKS insecure? Everybody can see I have encrypted data!
600 In practice it does not really matter. In most civilized countries
601 you can just refuse to hand over the keys, no harm done. In some
602 countries they can force you to hand over the keys, if they suspect
603 encryption. However the suspicion is enough, they do not have to
604 prove anything. This is for practical reasons, as even the presence
605 of a header (like the LUKS header) is not enough to prove that you
606 have any keys. It might have been an experiment, for example. Or it
607 was used as encrypted swap with a key from /dev/random. So they
608 make you prove you do not have encrypted data. Of course that is
609 just as impossible as the other way round.
611 This means that if you have a large set of random-looking data,
612 they can already lock you up. Hidden containers (encryption hidden
613 within encryption), as possible with Truecrypt, do not help
614 either. They will just assume the hidden container is there and
615 unless you hand over the key, you will stay locked up. Don't have
616 a hidden container? Though luck. Anybody could claim that.
618 Still, if you are concerned about the LUKS header, use plain
619 dm-crypt with a good passphrase. See also Section 2, "What is the
620 difference between "plain" and LUKS format?"
623 * 5.2 Should I initialize (overwrite) a new LUKS/dm-crypt partition?
625 If you just create a filesystem on it, most of the old data will
626 still be there. If the old data is sensitive, you should overwrite
627 it before encrypting. In any case, not initializing will leave the
628 old data there until the specific sector gets written. That may
629 enable an attacker to determine how much and where on the
630 partition data was written. If you think this is a risk, you can
631 prevent this by overwriting the encrypted device (here assumed to
632 be named "e1") with zeros like this:
634 dd_rescue -w /dev/zero /dev/mapper/e1
636 or alternatively with one of the following more standard commands:
638 cat /dev/zero > /dev/mapper/e1
639 dd if=/dev/zero of=/dev/mapper/e1
642 * 5.3 How do I securely erase a LUKS (or other) partition?
644 For LUKS, if you are in a desperate hurry, overwrite the LUKS
645 header and key-slot area. This means overwriting the first
646 (keyslots x stripes x keysize) + offset bytes. For the default
647 parameters, this is the 1'052'672 bytes, i.e. 1MiB + 4096 of the
648 LUKS partition. For 512 bit key length (e.g. for aes-xts-plain with
649 512 bit key) this is 2MiB. (The different offset stems from
650 differences in the sector alignment of the key-slots.) If in doubt,
651 just be generous and overwrite the first 10MB or so, it will likely
652 still be fast enough. A single overwrite with zeros should be
653 enough. If you anticipate being in a desperate hurry, prepare the
654 command beforehand. Example with /dev/sde1 as the LUKS partition
655 and default parameters:
657 head -c 1052672 /dev/zero > /dev/sde1; sync
659 A LUKS header backup or full backup will still grant access to
660 most or all data, so make sure that an attacker does not have
661 access to backups or destroy them as well.
663 If you have time, overwrite the whole LUKS partition with a single
664 pass of zeros. This is enough for current HDDs. For SSDs or FLASH
665 (USB sticks) you may want to overwrite the whole drive several
666 times to be sure data is not retained by wear leveling. This is
667 possibly still insecure as SSD technology is not fully understood
668 in this regard. Still, due to the anti-forensic properties of the
669 LUKS key-slots, a single overwrite of an SSD or FLASH drive could
670 be enough. If in doubt, use physical destruction in addition. Here
671 is a link to some current research results on erasing SSDs and
673 http://www.usenix.org/events/fast11/tech/full_papers/Wei.pdf
675 Keep in mind to also erase all backups.
677 Example for a zero-overwrite erase of partition sde1 done with
680 dd_rescue -w /dev/zero /dev/sde1
683 * 5.4 How do I securely erase a backup of a LUKS partition or header?
685 That depends on the medium it is stored on. For HDD and SSD, use
686 overwrite with zeros. For an SSD or FLASH drive (USB stick), you
687 may want to overwrite the complete SSD several times and use
688 physical destruction in addition, see last item. For re-writable
689 CD/DVD, a single overwrite should also be enough, due to the
690 anti-forensic properties of the LUKS keyslots. For write-once
691 media, use physical destruction. For low security requirements,
692 just cut the CD/DVD into several parts. For high security needs,
693 shred or burn the medium. If your backup is on magnetic tape, I
694 advise physical destruction by shredding or burning, after
695 overwriting . The problem with magnetic tape is that it has a
696 higher dynamic range than HDDs and older data may well be
697 recoverable after overwrites. Also write-head alignment issues can
698 lead to data not actually being deleted at all during overwrites.
701 * 5.5 What about backup? Does it compromise security?
703 That depends. See item 6.7.
706 * 5.6 Why is all my data permanently gone if I overwrite the LUKS
709 Overwriting the LUKS header in part or in full is the most common
710 reason why access to LUKS containers is lost permanently.
711 Overwriting can be done in a number of fashions, like creating a
712 new filesystem on the raw LUKS partition, making the raw partition
713 part of a raid array and just writing to the raw partition.
715 The LUKS header contains a 256 bit "salt" value and without that no
716 decryption is possible. While the salt is not secret, it is
717 key-grade material and cannot be reconstructed. This is a
718 cryptographically strong "cannot". From observations on the
719 cryptsetup mailing-list, people typically go though the usual
720 stages of grief (Denial, Anger, Bargaining, Depression, Acceptance)
721 when this happens to them. Observed times vary between 1 day and 2
722 weeks to complete the cycle. Seeking help on the mailing-list is
723 fine. Even if we usually cannot help with getting back your data,
724 most people found the feedback comforting.
726 If your header does not contain an intact salt, best go directly
727 to the last stage ("Acceptance") and think about what to do now.
728 There is one exception that I know of: If your LUKS container is
729 still open, then it may be possible to extract the master key from
730 the running system. See Item "How do I recover the master key from
731 a mapped LUKS container?" in Section "Backup and Data Recovery".
734 * 5.7 What is a "salt"?
736 A salt is a random key-grade value added to the passphrase before
737 it is processed. It is not kept secret. The reason for using salts
738 is as follows: If an attacker wants to crack the password for a
739 single LUKS container, then every possible passphrase has to be
740 tried. Typically an attacker will not try every binary value, but
741 will try words and sentences from a dictionary.
743 If an attacker wants to attack several LUKS containers with the
744 same dictionary, then a different approach makes sense: Compute the
745 resulting slot-key for each dictionary element and store it on
746 disk. Then the test for each entry is just the slow unlocking with
747 the slot key (say 0.00001 sec) instead of calculating the slot-key
748 first (1 sec). For a single attack, this does not help. But if you
749 have more than one container to attack, this helps tremendously,
750 also because you can prepare your table before you even have the
751 container to attack! The calculation is also very simple to
752 parallelize. You could, for example, use the night-time unused CPU
753 power of your desktop PCs for this.
755 This is where the salt comes in. If the salt is combined with the
756 passphrase (in the simplest form, just appended to it), you
757 suddenly need a separate table for each salt value. With a
758 reasonably-sized salt value (256 bit, e.g.) this is quite
762 * 5.8 Is LUKS secure with a low-entropy (bad) passphrase?
764 Note: You should only use the 94 printable characters from 7 bit
765 ASCII code to prevent your passphrase from failing when the
766 character encoding changes, e.g. because of a system upgrade, see
767 also the note at the very start of this FAQ under "WARNINGS".
769 This needs a bit of theory. The quality of your passphrase is
770 directly related to its entropy (information theoretic, not
771 thermodynamic). The entropy says how many bits of "uncertainty" or
772 "randomness" are in you passphrase. In other words, that is how
773 difficult guessing the passphrase is.
775 Example: A random English sentence has about 1 bit of entropy per
776 character. A random lowercase (or uppercase) character has about
779 Now, if n is the number of bits of entropy in your passphrase and t
780 is the time it takes to process a passphrase in order to open the
781 LUKS container, then an attacker has to spend at maximum
783 attack_time_max = 2^n * t
785 time for a successful attack and on average half that. There is no
786 way getting around that relationship. However, there is one thing
787 that does help, namely increasing t, the time it takes to use a
788 passphrase, see next FAQ item.
790 Still, if you want good security, a high-entropy passphrase is the
791 only option. For example, a low-entropy passphrase can never be
792 considered secure against a TLA-level (Three Letter Agency level,
793 i.e. government-level) attacker, no matter what tricks are used in
794 the key-derivation function. Use at least 64 bits for secret stuff.
795 That is 64 characters of English text (but only if randomly chosen)
796 or a combination of 12 truly random letters and digits.
798 For passphrase generation, do not use lines from very well-known
799 texts (religious texts, Harry potter, etc.) as they are to easy to
800 guess. For example, the total Harry Potter has about 1'500'000
801 words (my estimation). Trying every 64 character sequence starting
802 and ending at a word boundary would take only something like 20
803 days on a single CPU and is entirely feasible. To put that into
804 perspective, using a number of Amazon EC2 High-CPU Extra Large
805 instances (each gives about 8 real cores), this test costs
806 currently about 50USD/EUR, but can be made to run arbitrarily fast.
808 On the other hand, choosing 1.5 lines from, say, the Wheel of Time
809 is in itself not more secure, but the book selection adds quite a
810 bit of entropy. (Now that I have mentioned it here, don't use tWoT
811 either!) If you add 2 or 3 typos or switch some words around, then
812 this is good passphrase material.
815 * 5.9 What is "iteration count" and why is decreasing it a bad idea?
817 Iteration count is the number of PBKDF2 iterations a passphrase is
818 put through before it is used to unlock a key-slot. Iterations are
819 done with the explicit purpose to increase the time that it takes
820 to unlock a key-slot. This provides some protection against use of
821 low-entropy passphrases.
823 The idea is that an attacker has to try all possible passphrases.
824 Even if the attacker knows the passphrase is low-entropy (see last
825 item), it is possible to make each individual try take longer. The
826 way to do this is to repeatedly hash the passphrase for a certain
827 time. The attacker then has to spend the same time (given the same
828 computing power) as the user per try. With LUKS, the default is 1
829 second of PBKDF2 hashing.
831 Example 1: Lets assume we have a really bad passphrase (e.g. a
832 girlfriends name) with 10 bits of entropy. With the same CPU, an
833 attacker would need to spend around 500 seconds on average to
834 break that passphrase. Without iteration, it would be more like
835 0.0001 seconds on a modern CPU.
837 Example 2: The user did a bit better and has 32 chars of English
838 text. That would be about 32 bits of entropy. With 1 second
839 iteration, that means an attacker on the same CPU needs around 136
840 years. That is pretty impressive for such a weak passphrase.
841 Without the iterations, it would be more like 50 days on a modern
842 CPU, and possibly far less.
844 In addition, the attacker can both parallelize and use special
845 hardware like GPUs or FPGAs to speed up the attack. The attack can
846 also happen quite some time after the luksFormat operation and CPUs
847 can have become faster and cheaper. For that reason you want a
848 bit of extra security. Anyways, in Example 1 your are screwed.
849 In example 2, not necessarily. Even if the attack is faster, it
850 still has a certain cost associated with it, say 10000 EUR/USD
851 with iteration and 1 EUR/USD without iteration. The first can be
852 prohibitively expensive, while the second is something you try
853 even without solid proof that the decryption will yield something
856 The numbers above are mostly made up, but show the idea. Of course
857 the best thing is to have a high-entropy passphrase.
859 Would a 100 sec iteration time be even better? Yes and no.
860 Cryptographically it would be a lot better, namely 100 times better.
861 However, usability is a very important factor for security
862 technology and one that gets overlooked surprisingly often. For
863 LUKS, if you have to wait 2 minutes to unlock the LUKS container,
864 most people will not bother and use less secure storage instead. It
865 is better to have less protection against low-entropy passphrases
866 and people actually use LUKS, than having them do without
867 encryption altogether.
869 Now, what about decreasing the iteration time? This is generally a
870 very bad idea, unless you know and can enforce that the users only
871 use high-entropy passphrases. If you decrease the iteration time
872 without ensuring that, then you put your users at increased risk,
873 and considering how rarely LUKS containers are unlocked in a
874 typical work-flow, you do so without a good reason. Don't do it.
875 The iteration time is already low enough that users with entropy
876 low passphrases are vulnerable. Lowering it even further increases
877 this danger significantly.
880 * 5.10 Some people say PBKDF2 is insecure?
882 There is some discussion that a hash-function should have a "large
883 memory" property, i.e. that it should require a lot of memory to be
884 computed. This serves to prevent attacks using special programmable
885 circuits, like FPGAs, and attacks using graphics cards. PBKDF2
886 does not need a lot of memory and is vulnerable to these attacks.
887 However, the publication usually referred in these discussions is
888 not very convincing in proving that the presented hash really is
889 "large memory" (that may change, email the FAQ maintainer when it
890 does) and it is of limited usefulness anyways. Attackers that use
891 clusters of normal PCs will not be affected at all by a "large
892 memory" property. For example the US Secret Service is known to
893 use the off-hour time of all the office PCs of the Treasury for
894 password breaking. The Treasury has about 110'000 employees.
895 Assuming every one has an office PC, that is significant computing
896 power, all of it with plenty of memory for computing "large
897 memory" hashes. Bot-net operators also have all the memory they
898 want. The only protection against a resourceful attacker is a
899 high-entropy passphrase, see items 5.8 and 5.9.
902 * 5.11 What about iteration count with plain dm-crypt?
904 Simple: There is none. There is also no salting. If you use plain
905 dm-crypt, the only way to be secure is to use a high entropy
906 passphrase. If in doubt, use LUKS instead.
909 * 5.12 Is LUKS with default parameters less secure on a slow CPU?
911 Unfortunately, yes. However the only aspect affected is the
912 protection for low-entropy passphrase or master-key. All other
913 security aspects are independent of CPU speed.
915 The master key is less critical, as you really have to work at it
916 to give it low entropy. One possibility is to supply the master key
917 yourself. If that key is low-entropy, then you get what you
918 deserve. The other known possibility is to use /dev/urandom for
919 key generation in an entropy-starved situation (e.g. automatic
920 installation on an embedded device without network and other entropy
923 For the passphrase, don't use a low-entropy passphrase. If your
924 passphrase is good, then a slow CPU will not matter. If you insist
925 on a low-entropy passphrase on a slow CPU, use something like
926 "--iter-time=10" or higher and wait a long time on each LUKS unlock
927 and pray that the attacker does not find out in which way exactly
928 your passphrase is low entropy. This also applies to low-entropy
929 passphrases on fast CPUs. Technology can do only so much to
930 compensate for problems in front of the keyboard.
933 * 5.13 Why was the default aes-cbc-plain replaced with aes-cbc-essiv?
935 Note: This item applies both to plain dm-crypt and to LUKS
937 The problem is that cbc-plain has a fingerprint vulnerability, where
938 a specially crafted file placed into the crypto-container can be
939 recognized from the outside. The issue here is that for cbc-plain
940 the initialization vector (IV) is the sector number. The IV gets
941 XORed to the first data chunk of the sector to be encrypted. If you
942 make sure that the first data block to be stored in a sector
943 contains the sector number as well, the first data block to be
944 encrypted is all zeros and always encrypted to the same ciphertext.
945 This also works if the first data chunk just has a constant XOR
946 with the sector number. By having several shifted patterns you can
947 take care of the case of a non-power-of-two start sector number of
950 This mechanism allows you to create a pattern of sectors that have
951 the same first ciphertext block and signal one bit per sector to the
952 outside, allowing you to e.g. mark media files that way for
953 recognition without decryption. For large files this is a
954 practical attack. For small ones, you do not have enough blocks to
955 signal and take care of different file starting offsets.
957 In order to prevent this attack, the default was changed to
958 cbc-essiv. ESSIV uses a keyed hash of the sector number, with the
959 encryption key as key. This makes the IV unpredictable without
960 knowing the encryption key and the watermarking attack fails.
963 * 5.14 Are there any problems with "plain" IV? What is "plain64"?
965 First, "plain" and "plain64" are both not secure to use with CBC,
966 see previous FAQ item.
968 However there are modes, like XTS, that are secure with "plain" IV.
969 The next limit is that "plain" is 64 bit, with the upper 32 bit set
970 to zero. This means that on volumes larger than 2TiB, the IV
971 repeats, creating a vulnerability that potentially leaks some
972 data. To avoid this, use "plain64", which uses the full sector
973 number up to 64 bit. Note that "plain64" requires a kernel >=
974 2.6.33. Also note that "plain64" is backwards compatible for
975 volume sizes <= 2TiB, but not for those > 2TiB. Finally, "plain64"
976 does not cause any performance penalty compared to "plain".
979 * 5.15 What about XTS mode?
981 XTS mode is potentially even more secure than cbc-essiv (but only if
982 cbc-essiv is insecure in your scenario). It is a NIST standard and
983 used, e.g. in Truecrypt. At the moment, if you want to use it, you
984 have to specify it manually as "aes-xts-plain", i.e.
986 cryptsetup -c aes-xts-plain luksFormat <device>
988 For volumes >2TiB and kernels >= 2.6.33 use "plain64" (see FAQ
989 item on "plain" and "plain64"):
991 cryptsetup -c aes-xts-plain64 luksFormat <device>
993 There is a potential security issue with XTS mode and large blocks.
994 LUKS and dm-crypt always use 512B blocks and the issue does not
998 * 5.16 Is LUKS FIPS-140-2 certified?
1000 No. But that is more a problem of FIPS-140-2 than of LUKS. From a
1001 technical point-of-view, LUKS with the right parameters would be
1002 FIPS-140-2 compliant, but in order to make it certified, somebody
1003 has to pay real money for that. And then, whenever cryptsetup is
1004 changed or extended, the certification lapses and has to be
1007 From the aspect of actual security, LUKS with default parameters
1008 should be as good as most things that are FIPS-140-2 certified,
1009 although you may want to make sure to use /dev/random (by
1010 specifying --use-random on luksFormat) as randomness source for
1011 the master key to avoid being potentially insecure in an
1012 entropy-starved situation.
1015 * 5.16 What about Plausible Deniability?
1017 First let me attempt a definition for the case of encrypted
1018 filesystems: Plausible deniability is when you hide encrypted data
1019 inside an encrypted container and it is not possible to prove it is
1020 there. The idea is compelling and on first glance it seems
1021 possible to do it. And from a cryptographic point of view, it
1022 actually is possible.
1024 So, does it work in practice? No, unfortunately. The reasoning used
1025 by its proponents is fundamentally flawed in several ways and the
1026 cryptographic properties fail fatally when colliding with the real
1029 First, why should "I do not have a hidden partition" be any more
1030 plausible than "I forgot my crypto key" or "I wiped that partition
1031 with random data, nothing in there"? I do not see any reason.
1033 Second, there are two types of situations: Either they cannot force
1034 you to give them the key (then you simply do not) or the can. In
1035 the second case, they can always do bad things to you, because they
1036 cannot prove that you have the key in the first place! This means
1037 they do not have to prove you have the key, or that this random
1038 looking data on your disk is actually encrypted data. So the
1039 situation will allow them to waterboard/lock-up/deport you
1040 anyways, regardless of how "plausible" your deniability is. Do not
1041 have a hidden partition you could show to them, but there are
1042 indications you may? Too bad for you. Unfortunately "plausible
1043 deniability" also means you cannot prove there is no hidden data.
1045 Third, hidden partitions are not that hidden. There are basically
1046 just two possibilities: a) Make a large crypto container, but put a
1047 smaller filesystem in there and put the hidden partition into the
1048 free space. Unfortunately this is glaringly obvious and can be
1049 detected in an automated fashion. This means that the initial
1050 suspicion to put you under duress in order to make you reveal you
1051 hidden data is given. b) Make a filesystem that spans the whole
1052 encrypted partition, and put the hidden partition into space not
1053 currently used by that filesystem. Unfortunately that is also
1054 glaringly obvious, as you then cannot write to the filesystem
1055 without a high risk of destroying data in the hidden container.
1056 Have not written anything to the encrypted filesystem in a while?
1057 Too bad, they have the suspicion they need to do unpleasant things
1060 To be fair, if you prepare option b) carefully and directly before
1061 going into danger, it may work. But then, the mere presence of
1062 encrypted data may already be enough to get you into trouble in
1063 those places were they can demand encryption keys.
1065 Here is an additional reference for some problems with plausible
1066 deniability: http://www.schneier.com/paper-truecrypt-dfs.pdf I
1067 strongly suggest you read it.
1069 So, no, I will not provide any instructions on how to do it with
1070 plain dm-crypt or LUKS. If you insist on shooting yourself in the
1071 foot, you can figure out how to do it yourself.
1074 * 5.17 What about SSDs or Flash Drives?
1076 The problem is that you cannot reliably erase parts of these
1077 devices, mainly due to wear-leveling and possibly defect
1080 Basically, when overwriting a sector (of 512B), what the device
1081 does is to move an internal sector (may be 128kB or even larger) to
1082 some pool of discarded, not-yet erased unused sectors, take a
1083 fresh empty sector from the empty-sector pool and copy the old
1084 sector over with the changes to the small part you wrote. This is
1085 done in some fashion so that larger writes do not cause a lot of
1086 small internal updates.
1088 The thing is that the mappings between outside-adressable sectors
1089 and inside sectors is arbitrary (and the vendors are not talking).
1090 Also the discarded sectors are not necessarily erased immediately.
1091 They may linger a long time.
1093 For plain dm-crypt, the consequences are that older encrypted data
1094 may be lying around in some internal pools of the device. Thus may
1095 or may not be a problem and depends on the application. Remember
1096 the same can happen with a filesystem if consecutive writes to the
1097 same area of a file can go to different sectors.
1099 However, for LUKS, the worst case is that key-slots and LUKS
1100 header may end up in these internal pools. This means that password
1101 management functionality is compromised (the old passwords may
1102 still be around, potentially for a very long time) and that fast
1103 erase by overwriting the header and key-slot area is insecure.
1105 Also keep in mind that the discarded/used pool may be large. For
1106 example, a 240GB SSD has about 16GB of spare area in the chips that
1107 it is free to do with as it likes. You would need to make each
1108 individual key-slot larger than that to allow reliable overwriting.
1109 And that assumes the disk thinks all other space is in use.
1110 Reading the internal pools using forensic tools is not that hard,
1111 but may involve some soldering.
1115 If you trust the device vendor (you probably should not...) you can
1116 try an ATA "secure erase" command for SSDs. That does not work for
1117 USB keys though. And if it finishes after a few seconds, it was
1118 possibly faked by the SSD.
1120 If you can do without password management and are fine with doing
1121 physical destruction for permenently deleting data (allways after
1122 one or several full overwrites!), you can use plain dm-crypt or
1125 If you want or need the original LUKS security features to work,
1126 you can use a detached LUKS header and put that on a conventional,
1127 magnetic disk. That leaves potentially old encrypted data in the
1128 pools on the disk, but otherwise you get LUKS with the same
1129 security as on a magnetic disk.
1131 If you are concerned about your laptop being stolen, you are likely
1132 fine using LUKS on an SSD. An attacker would need to have access
1133 to an old passphrase (and the key-slot for this old passphrase
1134 would actually need to still be somewhere in the SSD) for your
1135 data to be at risk. So unless you pasted your old passphrase all
1136 over the Internet or the attacker has knowledge of it from some
1137 other source and does a targetted laptop theft to get at your
1138 data, you should be fine.
1141 6. Backup and Data Recovery
1144 * 6.1 Why do I need Backup?
1146 First, disks die. The rate for well-treated (!) disk is about 5%
1147 per year, which is high enough to worry about. There is some
1148 indication that this may be even worse for some SSDs. This applies
1149 both to LUKS and plain dm-crypt partitions.
1151 Second, for LUKS, if anything damages the LUKS header or the
1152 key-stripe area then decrypting the LUKS device can become
1153 impossible. This is a frequent occurrence. For example an
1154 accidental format as FAT or some software overwriting the first
1155 sector where it suspects a partition boot sector typically makes a
1156 LUKS partition permanently inaccessible. See more below on LUKS
1159 So, data-backup in some form is non-optional. For LUKS, you may
1160 also want to store a header backup in some secure location. This
1161 only needs an update if you change passphrases.
1164 * 6.2 How do I backup a LUKS header?
1166 While you could just copy the appropriate number of bytes from the
1167 start of the LUKS partition, the best way is to use command option
1168 "luksHeaderBackup" of cryptsetup. This protects also against
1169 errors when non-standard parameters have been used in LUKS
1170 partition creation. Example:
1173 cryptsetup luksHeaderBackup --header-backup-file <file> <device>
1175 To restore, use the inverse command, i.e.
1177 cryptsetup luksHeaderRestore --header-backup-file <file> <device>
1180 * 6.3 How do I test a LUKS header?
1184 cryptsetup -v isLuks <device>
1186 on the device. Without the "-v" it just signals its result via
1187 exit-status. You can also use the more general test
1191 which will also detect other types and give some more info. Omit
1192 "-p" for old versions of blkid that do not support it.
1195 * 6.4 How do I backup a LUKS or dm-crypt partition?
1197 There are two options, a sector-image and a plain file or
1198 filesystem backup of the contents of the partition. The sector
1199 image is already encrypted, but cannot be compressed and contains
1200 all empty space. The filesystem backup can be compressed, can
1201 contain only part of the encrypted device, but needs to be
1202 encrypted separately if so desired.
1204 A sector-image will contain the whole partition in encrypted form,
1205 for LUKS the LUKS header, the keys-slots and the data area. It can
1206 be done under Linux e.g. with dd_rescue (for a direct image copy)
1207 and with "cat" or "dd". Example:
1209 cat /dev/sda10 > sda10.img
1210 dd_rescue /dev/sda10 sda10.img
1212 You can also use any other backup software that is capable of making
1213 a sector image of a partition. Note that compression is
1214 ineffective for encrypted data, hence it does not make sense to
1217 For a filesystem backup, you decrypt and mount the encrypted
1218 partition and back it up as you would a normal filesystem. In this
1219 case the backup is not encrypted, unless your encryption method
1220 does that. For example you can encrypt a backup with "tar" as
1223 tar cjf - <path> | gpg --cipher-algo AES -c - > backup.tbz2.gpg
1225 And verify the backup like this if you are at "path":
1227 cat backup.tbz2.gpg | gpg - | tar djf -
1229 Note: Always verify backups, especially encrypted ones.
1231 In both cases GnuPG will ask you interactively for your symmetric
1232 key. The verify will only output errors. Use "tar dvjf -" to get
1233 all comparison results. To make sure no data is written to disk
1234 unencrypted, turn off swap if it is not encrypted before doing the
1237 You can of course use different or no compression and you can use
1238 an asymmetric key if you have one and have a backup of the secret
1239 key that belongs to it.
1241 A second option for a filesystem-level backup that can be used when
1242 the backup is also on local disk (e.g. an external USB drive) is
1243 to use a LUKS container there and copy the files to be backed up
1244 between both mounted containers. Also see next item.
1247 * 6.5 Do I need a backup of the full partition? Would the header and
1248 key-slots not be enough?
1250 Backup protects you against two things: Disk loss or corruption
1251 and user error. By far the most questions on the dm-crypt mailing
1252 list about how to recover a damaged LUKS partition are related
1253 to user error. For example, if you create a new filesystem on a
1254 LUKS partition, chances are good that all data is lost
1257 For this case, a header+key-slot backup would often be enough. But
1258 keep in mind that a well-treated (!) HDD has roughly a failure
1259 risk of 5% per year. It is highly advisable to have a complete
1260 backup to protect against this case.
1263 * *6.6 What do I need to backup if I use "decrypt_derived"?
1265 This is a script in Debian, intended for mounting /tmp or swap with
1266 a key derived from the master key of an already decrypted device.
1267 If you use this for an device with data that should be persistent,
1268 you need to make sure you either do not lose access to that master
1269 key or have a backup of the data. If you derive from a LUKS
1270 device, a header backup of that device would cover backing up the
1271 master key. Keep in mind that this does not protect against disk
1274 Note: If you recreate the LUKS header of the device you derive from
1275 (using luksFormat), the master key changes even if you use the same
1276 passphrase(s) and you will not be able to decrypt the derived
1277 device with the new LUKS header.
1280 * 6.7 Does a backup compromise security?
1282 Depends on how you do it. However if you do not have one, you are
1283 going to eventually lose your encrypted data.
1285 There are risks introduced by backups. For example if you
1286 change/disable a key-slot in LUKS, a binary backup of the partition
1287 will still have the old key-slot. To deal with this, you have to
1288 be able to change the key-slot on the backup as well, securely
1289 erase the backup or do a filesystem-level backup instead of a binary
1292 If you use dm-crypt, backup is simpler: As there is no key
1293 management, the main risk is that you cannot wipe the backup when
1294 wiping the original. However wiping the original for dm-crypt
1295 should consist of forgetting the passphrase and that you can do
1296 without actual access to the backup.
1298 In both cases, there is an additional (usually small) risk with
1299 binary backups: An attacker can see how many sectors and which
1300 ones have been changed since the backup. To prevent this, use a
1301 filesystem level backup method that encrypts the whole backup in
1302 one go, e.g. as described above with tar and GnuPG.
1304 My personal advice is to use one USB disk (low value data) or
1305 three disks (high value data) in rotating order for backups, and
1306 either use independent LUKS partitions on them, or use encrypted
1307 backup with tar and GnuPG.
1309 If you do network-backup or tape-backup, I strongly recommend to
1310 go the filesystem backup path with independent encryption, as you
1311 typically cannot reliably delete data in these scenarios,
1312 especially in a cloud setting. (Well, you can burn the tape if it
1313 is under your control...)
1316 * 6.8 What happens if I overwrite the start of a LUKS partition or
1317 damage the LUKS header or key-slots?
1319 There are two critical components for decryption: The salt values
1320 in the header itself and the key-slots. If the salt values are
1321 overwritten or changed, nothing (in the cryptographically strong
1322 sense) can be done to access the data, unless there is a backup
1323 of the LUKS header. If a key-slot is damaged, the data can still
1324 be read with a different key-slot, if there is a remaining
1325 undamaged and used key-slot. Note that in order to make a key-slot
1326 unrecoverable in a cryptographically strong sense, changing about
1327 4-6 bits in random locations of its 128kiB size is quite enough.
1330 * 6.9 What happens if I (quick) format a LUKS partition?
1332 I have not tried the different ways to do this, but very likely you
1333 will have written a new boot-sector, which in turn overwrites the
1334 LUKS header, including the salts, making your data permanently
1335 irretrievable, unless you have a LUKS header backup. You may also
1336 damage the key-slots in part or in full. See also last item.
1339 * 6.10 How do I recover the master key from a mapped LUKS container?
1341 This is typically only needed if you managed to damage your LUKS
1342 header, but the container is still mapped, i.e. "luksOpen"ed. It
1343 also helps if you have a mapped container that you forgot or do not
1344 know a passphrase for (e.g. on a long running server.)
1346 WARNING: Things go wrong, do a full backup before trying this!
1348 WARNING: This exposes the master key of the LUKS container. Note
1349 that both ways to recreate a LUKS header with the old master key
1350 described below will write the master key to disk. Unless you are
1351 sure you have securely erased it afterwards, e.g. by writing it to
1352 an encrypted partition, RAM disk or by erasing the filesystem you
1353 wrote it to by a complete overwrite, you should change the master
1354 key afterwards. Changing the master key requires a full data
1355 backup, luksFormat and then restore of the backup.
1357 First, there is a script by Milan that automates the whole
1358 process, except generating a new LUKS header with the old master
1359 key (it prints the command for that though):
1361 http://code.google.com/p/cryptsetup/source/browse/misc/luks-header-from-active
1363 You can also do this manually. Here is how:
1365 - Get the master key from the device mapper. This is done by the
1366 following command. Substitute c5 for whatever you mapped to:
1368 # dmsetup table --target crypt --showkey /dev/mapper/c5
1370 0 200704 crypt aes-cbc-essiv:sha256
1371 a1704d9715f73a1bb4db581dcacadaf405e700d591e93e2eaade13ba653d0d09
1374 The result is actually one line, wrapped here for clarity. The long
1375 hex string is the master key.
1377 - Convert the master key to a binary file representation. You can
1378 do this manually, e.g. with hexedit. You can also use the tool
1379 "xxd" from vim like this:
1381 echo "a1704d9....53d0d09" | xxd -r -p > <master-key-file>
1383 - Do a luksFormat to create a new LUKS header.
1385 NOTE: If your header is intact and you just forgot the
1386 passphrase, you can just set a new passphrase, see next
1389 Unmap the device before you do that (luksClose). Then do
1391 cryptsetup luksFormat --master-key-file=<master-key-file> <luks device>
1393 Note that if the container was created with other than the default
1394 settings of the cryptsetup version you are using, you need to give
1395 additional parameters specifying the deviations. If in doubt, try
1396 the script by Milan. It does recover the other parameters as well.
1398 Side note: This is the way the decrypt_derived script gets at the
1399 master key. It just omits the conversion and hashes the master key
1402 - If the header is intact and you just forgot the passphrase, just
1403 set a new passphrase like this:
1405 cryptsetup luksAddKey --master-key-file=<master-key-file> <luks device>
1407 You may want to disable the old one afterwards.
1410 * 6.11 What does the on-disk structure of dm-crypt look like?
1412 There is none. dm-crypt takes a block device and gives encrypted
1413 access to each of its blocks with a key derived from the passphrase
1414 given. If you use a cipher different than the default, you have to
1415 specify that as a parameter to cryptsetup too. If you want to
1416 change the password, you basically have to create a second
1417 encrypted device with the new passphrase and copy your data over.
1418 On the plus side, if you accidentally overwrite any part of a
1419 dm-crypt device, the damage will be limited to the are you
1423 * 6.12 What does the on-disk structure of LUKS look like?
1425 A LUKS partition consists of a header, followed by 8 key-slot
1426 descriptors, followed by 8 key slots, followed by the encrypted
1429 Header and key-slot descriptors fill the first 592 bytes. The
1430 key-slot size depends on the creation parameters, namely on the
1431 number of anti-forensic stripes, key material offset and master
1434 With the default parameters, each key-slot is a bit less than
1435 128kiB in size. Due to sector alignment of the key-slot start,
1436 that means the key block 0 is at offset 0x1000-0x20400, key
1437 block 1 at offset 0x21000-0x40400, and key block 7 at offset
1438 0xc1000-0xe0400. The space to the next full sector address is
1439 padded with zeros. Never used key-slots are filled with what the
1440 disk originally contained there, a key-slot removed with
1441 "luksRemoveKey" or "luksKillSlot" gets filled with 0xff. Due to
1442 2MiB default alignment, start of the data area for cryptsetup 1.3
1443 and later is at 2MiB, i.e. at 0x200000. For older versions, it is
1444 at 0x101000, i.e. at 1'052'672 bytes, i.e. at 1MiB + 4096 bytes
1445 from the start of the partition. Incidentally, "luksHeaderBackup"
1446 for a LUKS container created with default parameters dumps exactly
1447 the first 2MiB (or 1'052'672 bytes for headers created with
1448 cryptsetup versions < 1.3) to file and "luksHeaderRestore" restores
1451 For non-default parameters, you have to figure out placement
1452 yourself. "luksDump" helps. See also next item. For the most common
1453 non-default settings, namely aes-xts-plain with 512 bit key, the
1454 offsets are: 1st keyslot 0x1000-0x3f800, 2nd keyslot
1455 0x40000-0x7e000, 3rd keyslot 0x7e000-0xbd800, ..., and start of
1456 bulk data at 0x200000.
1458 The exact specification of the format is here:
1459 http://code.google.com/p/cryptsetup/wiki/Specification
1462 * 6.13 What is the smallest possible LUKS container?
1464 Note: From cryptsetup 1.3 onwards, alignment is set to 1MB. With
1465 modern Linux partitioning tools that also align to 1MB, this will
1466 result in alignment to 2k sectors and typical Flash/SSD sectors,
1467 which is highly desirable for a number of reasons. Changing the
1468 alignment is not recommended.
1470 That said, with default parameters, the data area starts at
1471 exactly 2MB offset (at 0x101000 for cryptsetup versions before
1472 1.3). The smallest data area you can have is one sector of 512
1473 bytes. Data areas of 0 bytes can be created, but fail on mapping.
1475 While you cannot put a filesystem into something this small, it may
1476 still be used to contain, for example, key. Note that with current
1477 formatting tools, a partition for a container this size will be
1478 3MiB anyways. If you put the LUKS container into a file (via
1479 losetup and a loopback device), the file needs to be 2097664 bytes
1480 in size, i.e. 2MiB + 512B.
1482 There two ways to influence the start of the data area are key-size
1485 For alignment, you can go down to 1 on the parameter. This will
1486 still leave you with a data-area starting at 0x101000, i.e.
1487 1MiB+4096B (default parameters) as alignment will be rounded up to
1488 the next multiple of 8 (i.e. 4096 bytes) If in doubt, do a dry-run
1489 on a larger file and dump the LUKS header to get actual
1492 For key-size, you can use 128 bit (e.g. AES-128 with CBC), 256 bit
1493 (e.g. AES-256 with CBC) or 512 bit (e.g. AES-256 with XTS mode).
1494 You can do 64 bit (e.g. blowfish-64 with CBC), but anything below
1495 128 bit has to be considered insecure today.
1497 Example 1 - AES 128 bit with CBC:
1499 cryptsetup luksFormat -s 128 --align-payload=8 <device>
1501 This results in a data offset of 0x81000, i.e. 516KiB or 528384
1502 bytes. Add one 512 byte sector and the smallest LUKS container size
1503 with these parameters is 516KiB + 512B or 528896 bytes.
1505 Example 2 - Blowfish 64 bit with CBC (WARNING: insecure):
1507 cryptsetup luksFormat -c blowfish -s 64 --align-payload=8 /dev/loop0
1509 This results in a data offset of 0x41000, i.e. 260kiB or 266240
1510 bytes, with a minimal LUKS container size of 260kiB + 512B or
1514 * 6.14 I think this is overly complicated. Is there an alternative?
1516 Not really. Encryption comes at a price. You can use plain
1517 dm-crypt to simplify things a bit. It does not allow multiple
1518 passphrases, but on the plus side, it has zero on disk description
1519 and if you overwrite some part of a plain dm-crypt partition,
1520 exactly the overwritten parts are lost (rounded up to sector
1524 * 6.15 Can I clone a LUKS container?
1526 You can, but it breaks security, because the cloned container has
1527 the same header and hence the same master key. You cannot change
1528 the master key on a LUKS container, even if you change the
1529 passphrase(s), the master key stays the same. That means whoever
1530 has access to one of the clones can decrypt them all, completely
1531 bypassing the passphrases.
1533 The right way to do this is to first luksFormat the target
1534 container, then to clone the contents of the source container, with
1535 both containers mapped, i.e. decrypted. You can clone the decrypted
1536 contents of a LUKS container in binary mode, although you may run
1537 into secondary issues with GUIDs in filesystems, partition tables,
1538 RAID-components and the like. These are just the normal problems
1539 binary cloning causes.
1541 Note that if you need to ship (e.g.) cloned LUKS containers with a
1542 default passphrase, that is fine as long as each container was
1543 individually created (and hence has its own master key). In this
1544 case, changing the default passphrase will make it secure again.
1547 7. Interoperability with other Disk Encryption Tools
1550 * 7.1 What is this section about?
1552 Cryptsetup for plain dm-crypt can be used to access a number of
1553 on-disk formats created by tools like loop-aes patched into
1554 losetup. This sometimes works and sometimes does not. This
1555 section collects insights into what works, what does not and where
1556 more information is required.
1558 Additional information may be found in the mailing-list archives,
1559 mentioned at the start of this FAQ document. If you have a
1560 solution working that is not yet documented here and think a wider
1561 audience may be interested, please email the FAQ maintainer.
1564 * 7.2 loop-aes: General observations.
1566 One problem is that there are different versions of losetup around.
1567 loop-aes is a patch for losetup. Possible problems and deviations
1568 from cryptsetup option syntax include:
1570 - Offsets specified in bytes (cryptsetup: 512 byte sectors)
1572 - The need to specify an IV offset
1574 - Encryption mode needs specifying (e.g. "-c twofish-cbc-plain")
1576 - Key size needs specifying (e.g. "-s 128" for 128 bit keys)
1578 - Passphrase hash algorithm needs specifying
1580 Also note that because plain dm-crypt and loop-aes format does not
1581 have metadata, and while the loopAES extension for cryptsetup tries
1582 autodetection (see command loopaesOpen), it may not always work.
1583 If you still have the old set-up, using a verbosity option (-v)
1584 on mapping with the old tool or having a look into the system logs
1585 after setup could give you the information you need. Below, there
1586 are also some things that worked for somebody.
1589 * 7.3 loop-aes patched into losetup on Debian 5.x, kernel 2.6.32
1591 In this case, the main problem seems to be that this variant of
1592 losetup takes the offset (-o option) in bytes, while cryptsetup
1593 takes it in sectors of 512 bytes each. Example: The losetup command
1595 losetup -e twofish -o 2560 /dev/loop0 /dev/sdb1
1596 mount /dev/loop0 mount-point
1600 cryptsetup create -c twofish -o 5 --skip 5 e1 /dev/sdb1
1601 mount /dev/mapper/e1 mount-point
1604 * 7.4 loop-aes with 160 bit key
1606 This seems to be sometimes used with twofish and blowfish and
1607 represents a 160 bit ripemed160 hash output padded to 196 bit key
1608 length. It seems the corresponding options for cryptsetup are
1610 --cipher twofish-cbc-null -s 192 -h ripemd160:20
1613 * 7.5 loop-aes v1 format OpenSUSE
1615 Apparently this is done by older OpenSUSE distros and stopped
1616 working from OpenSUSE 12.1 to 12.2. One user had success with the
1619 cryptsetup create <target> <device> -c aes -s 128 -h sha256
1622 8. Issues with Specific Versions of cryptsetup
1625 * 8.1 When using the create command for plain dm-crypt with
1626 cryptsetup 1.1.x, the mapping is incompatible and my data is not
1629 With cryptsetup 1.1.x, the distro maintainer can define different
1630 default encryption modes for LUKS and plain devices. You can check
1631 these compiled-in defaults using "cryptsetup --help". Moreover, the
1632 plain device default changed because the old IV mode was
1633 vulnerable to a watermarking attack.
1635 If you are using a plain device and you need a compatible mode, just
1636 specify cipher, key size and hash algorithm explicitly. For
1637 compatibility with cryptsetup 1.0.x defaults, simple use the
1640 cryptsetup create -c aes-cbc-plain -s 256 -h ripemd160 <name> <dev>
1642 LUKS stores cipher and mode in the metadata on disk, avoiding this
1646 * 8.2 cryptsetup on SLED 10 has problems...
1648 SLED 10 is missing an essential kernel patch for dm-crypt, which
1649 is broken in its kernel as a result. There may be a very old
1650 version of cryptsetup (1.0.x) provided by SLED, which should also
1651 not be used anymore as well. My advice would be to drop SLED 10.
1654 9. References and Further Reading
1657 * Purpose of this Section
1659 The purpose of this section is to collect references to all
1660 materials that do not fit the FAQ but are relevant in some fashion.
1661 This can be core topics like the LUKS spec or disk encryption, but
1662 it can also be more tangential, like secure storage management or
1663 cryptography used in LUKS. It should still have relevance to
1664 cryptsetup and its applications.
1666 If you wan to see something added here, send email to the
1667 maintainer (or the cryptsetup mailing list) giving an URL, a
1668 description (1-3 lines preferred) and a section to put it in. You
1669 can also propose new sections.
1671 At this time I would like to limit the references to things that
1672 are available on the web.
1677 - LUKS on-disk format spec:
1678 http://code.google.com/p/cryptsetup/wiki/Specification
1683 - Some code examples are in the source package under docs/examples
1689 * SSD and Flash Disk Related
1695 * Attacks Against Disk Encryption
1698 * Risk Management as Relevant for Disk Encryption
1706 A. Contributors In no particular order: