8 6. Backup and Data Recovery
9 7. Interoperability with other Disk Encryption Tools
10 8. Issues with Specific Versions of cryptsetup
19 This is the FAQ (Frequently Asked Questions) for cryptsetup. It
20 covers Linux disk encryption with plain dm-crypt (one passphrase,
21 no management, no metadata on disk) and LUKS (multiple user keys
22 with one master key, anti-forensic features, metadata block at
23 start of device, ...). The latest version of this FAQ should
24 usually be available at
25 http://code.google.com/p/cryptsetup/wiki/FrequentlyAskedQuestions
30 ATTENTION: If you are going to read just one thing, make it the
31 section on Backup and Data Recovery. By far the most questions on
32 the cryptsetup mailing list are from people that managed to damage
33 the start of their LUKS partitions, i.e. the LUKS header. In
34 most cases, there is nothing that can be done to help these poor
35 souls recover their data. Make sure you understand the problem and
36 limitations imposed by the LUKS security model BEFORE you face
37 such a disaster! In particular, make sure you have a current header
38 backup before doing any potentially dangerous operations.
40 BACKUP: Yes, encrypted disks die, just as normal ones do. A full
41 backup is mandatory, see Section "6. Backup and Data Recovery" on
42 options for doing encrypted backup.
44 DISTRIBUTION INSTALLERS: Some distribution installers offer to
45 create LUKS containers in a way that can be mistaken as activation
46 of an existing container. Creating a new LUKS container on top of
47 an existing one leads to permanent, complete and irreversible data
48 loss. It is strongly recommended to only use distribution
49 installers after a complete backup of all LUKS containers has been
52 LUKS PASSPHRASE IS NOT THE MASTER KEY: The LUKS passphrase is not
53 used in deriving the master key. It is used in decrypting a master
54 key that is randomly selected on header creation. This means that
55 if you create a new LUKS header on top of an old one with
56 exactly the same parameters and exactly the same passphrase as the
57 old one, it will still have a different master key and your data
58 will be permanently lost.
60 PASSPHRASE CHARACTER SET: Some people have had difficulties with
61 this when upgrading distributions. It is highly advisable to only
62 use the 94 printable characters from the first 128 characters of
63 the ASCII table, as they will always have the same binary
64 representation. Other characters may have different encoding
65 depending on system configuration and your passphrase will not
66 work with a different encoding. A table of the standardized first
67 128 ASCII caracters can, e.g. be found on
68 http://en.wikipedia.org/wiki/ASCII
71 * System Specific warnings
73 - Ubuntu as of 4/2011: It seems the installer offers to create
74 LUKS partitions in a way that several people mistook for an offer
75 to activate their existing LUKS partition. The installer gives no
76 or an inadequate warning and will destroy your old LUKS header,
77 causing permanent data loss. See also the section on Backup and
80 This issue has been acknowledged by the Ubuntu dev team, see here:
81 http://launchpad.net/bugs/420080
86 Current FAQ maintainer is Arno Wagner <arno@wagner.name>. Other
87 contributors are listed at the end. If you want to contribute, send
88 your article, including a descriptive headline, to the maintainer,
89 or the dm-crypt mailing list with something like "FAQ ..." in the
90 subject. You can also send more raw information and have me write
91 the section. Please note that by contributing to this FAQ, you
92 accept the license described below.
94 This work is under the "Attribution-Share Alike 3.0 Unported"
95 license, which means distribution is unlimited, you may create
96 derived works, but attributions to original authors and this
97 license statement must be retained and the derived work must be
98 under the same license. See
99 http://creativecommons.org/licenses/by-sa/3.0/ for more details of
102 Side note: I did text license research some time ago and I think
103 this license is best suited for the purpose at hand and creates the
107 * Where is the project website?
109 There is the project website at http://code.google.com/p/cryptsetup/
110 Please do not post questions there, nobody will read them. Use
111 the mailing-list instead.
114 * Is there a mailing-list?
116 Instructions on how to subscribe to the mailing-list are at on the
117 project website. People are generally helpful and friendly on the
120 The question of how to unsubscribe from the list does crop up
121 sometimes. For this you need your list management URL, which is
122 sent to you initially and once at the start of each month. Go to
123 the URL mentioned in the email and select "unsubscribe". This page
124 also allows you to request a password reminder.
126 Alternatively, you can send an Email to dm-crypt-request@saout.de
127 with just the word "help" in the subject or message body. Make sure
128 to send it from your list address.
130 The mailing list archive is here:
131 http://dir.gmane.org/gmane.linux.kernel.device-mapper.dm-crypt
137 * What is the difference between "plain" and LUKS format?
139 Plain format is just that: It has no metadata on disk, reads all
140 paramters from the commandline (or the defaults), derives a
141 master-key from the passphrase and then uses that to de-/encrypt
142 the sectors of the device, with a direct 1:1 mapping between
143 encrypted and decrypted sectors.
145 Primary advantage is high resilience to damage, as one damaged
146 encrypted sector results in exactly one damaged decrypted sector.
147 Also, it is not readily apparent that there even is encrypted data
148 on the device, as an overwrite with crypto-grade randomness (e.g.
149 from /dev/urandom) looks exactly the same on disk.
151 Side-note: That has limited value against the authorities. In
152 civilized countries, they cannot force you to give up a crypto-key
153 anyways. In the US, the UK and dictatorships around the world,
154 they can force you to give up the keys (using imprisonment or worse
155 to pressure you), and in the worst case, they only need a
156 nebulous "suspicion" about the presence of encrypted data. My
157 advice is to either be ready to give up the keys or to not have
158 encrypted data when traveling to those countries, especially when
159 crossing the borders.
161 Disadvantages are that you do not have all the nice features that
162 the LUKS metadata offers, like multiple passphrases that can be
163 changed, the cipher being stored in the metadata, anti-forensic
164 properties like key-slot diffusion and salts, etc..
166 LUKS format uses a metadata header and 8 key-slot areas that are
167 being placed ath the begining of the disk, see below under "What
168 does the LUKS on-disk format looks like?". The passphrases are used
169 to decryt a single master key that is stored in the anti-forensic
172 Advantages are a higher usability, automatic configuration of
173 non-default crypto parameters, defenses against low-entropy
174 passphrases like salting and iterated PBKDF2 passphrase hashing,
175 the ability to change passhrases, and others.
177 Disadvantages are that it is readily obvious there is encrypted
178 data on disk (but see side note above) and that damage to the
179 header or key-slots usually results in permanent data-loss. See
180 below under "6. Backup and Data Recovery" on how to reduce that
181 risk. Also the sector numbers get shifted by the length of the
182 header and key-slots and there is a loss of that size in capacity
183 (1MB+4096B for defaults and 2MB for the most commonly used
184 non-default XTS mode).
187 * Can I encrypt an already existing, non-empty partition to use LUKS?
189 There is no converter, and it is not really needed. The way to do
190 this is to make a backup of the device in question, securely wipe
191 the device (as LUKS device initialization does not clear away old
192 data), do a luksFormat, optionally overwrite the encrypted device,
193 create a new filesystem and restore your backup on the now
194 encrypted device. Also refer to sections "Security Aspects" and
195 "Backup and Data Recovery".
197 For backup, plain GNU tar works well and backs up anything likely
198 to be in a filesystem.
201 * How do I use LUKS with a loop-device?
203 This can be very handy for experiments. Setup is just the same as
204 with any block device. If you want, for example, to use a 100MiB
205 file as LUKS container, do something like this:
207 head -c 100M /dev/zero > luksfile # create empty file
208 losetup /dev/loop0 luksfile # map luksfile to /dev/loop0
209 cryptsetup luksFormat /dev/loop0 # create LUKS on loop device
211 Afterwards just use /dev/loop0 as a you would use a LUKS partition.
212 To unmap the file when done, use "losetup -d /dev/loop0".
215 * When I add a new key-slot to LUKS, it asks for a passphrase but
216 then complains about there not being a key-slot with that
219 That is as intended. You are asked a passphrase of an existing
220 key-slot first, before you can enter the passphrase for the new
221 key-slot. Otherwise you could break the encryption by just adding a
222 new key-slot. This way, you have to know the passphrase of one of
223 the already configured key-slots in order to be able to configure a
227 * Encrytion on top of RAID or the other way round?
229 Unless you have special needs, place encryption between RAID and
230 filesystem, i.e. encryption on top of RAID. You can do it the other
231 way round, but you have to be aware that you then need to give the
232 pasphrase for each individual disk and RAID autotetection will not
233 work anymore. Therefore it is better to encrypt the RAID device,
237 * How do I read a dm-crypt key from file?
239 Note that the file will still be hashed first, just like keyboard
240 input. Use the --key-file option, like this:
242 cryptsetup create --key-file keyfile e1 /dev/loop0
245 * How do I read a LUKS slot key from file?
247 What you really do here is to read a passphrase from file, just as
248 you would with manual entry of a passphrase for a key-slot. You can
249 add a new passphrase to a free key-slot, set the passphrase of an
250 specific key-slot or put an already configured passphrase into a
251 file. In the last case make sure no trailing newline (0x0a) is
252 contained in the key file, or the passphrase will not work because
253 the whole file is used as input.
255 To add a new passphrase to a free key slot from file, use something
258 cryptsetup luksAddKey /dev/loop0 keyfile
260 To add a new passphrase to a specific key-slot, use something like
263 cryptsetup luksAddKey --key-slot 7 /dev/loop0 keyfile
265 To supply a key from file to any LUKS command, use the --key-file
266 option, e.g. like this:
268 cryptsetup luksOpen --key-file keyfile /dev/loop0 e1
271 * How do I read the LUKS master key from file?
273 The question you should ask yourself first is why you would want to
274 do this. The only legitimate reason I can think of is if you want
275 to have two LUKS devices with the same master key. Even then, I
276 think it would be preferable to just use key-slots with the same
277 passphrase, or to use plain dm-crypt instead. If you really have a
278 good reason, please tell me. If I am convinced, I will add how to
282 * What are the security requirements for a key read from file?
284 A file-stored key or passphrase has the same security requirements
285 as one entered interactively, however you can use random bytes and
286 thereby use bytes you cannot type on the keyboard. You can use any
287 file you like as key file, for example a plain text file with a
288 human readable passphrase. To generate a file with random bytes,
289 use something like this:
291 head -c 256 /dev/random > keyfile
294 * If I map a journaled file system using dm-crypt/LUKS, does it still
295 provide its usual transactional guarantees?
297 As far as I know it does (but I may be wrong), but please note that
298 these "guarantees" are far weaker than they appear to be. For
299 example, you may not get a hard flush to disk surface even on a
300 call to fsync. In addition, the HDD itself may do independent
301 write reordering. Some other things can go wrong as well. The
302 filesystem developers are aware of these problems and typically
303 can make it work anyways. That said, dm-crypt/LUKS should not make
306 Personally, I have several instances of ext3 on dm-crypt and have
307 not noticed any specific problems.
309 Update: I did run into frequent small freezes (1-2 sec) when putting
310 a vmware image on ext3 over dm-crypt. This does indicate that the
311 transactional guarantees are in place, but at a cost. When I went
312 back to ext2, the problem went away. This also seems to have gotten
313 better with kernel 2.6.36 and the reworking of filesystem flush
314 locking. Kernel 2.6.38 is expected to have more improvements here.
317 * Can I use LUKS or cryptsetup with a more secure (external) medium
318 for key storage, e.g. TPM or a smartcard?
320 Yes, see the answers on using a file-supplied key. You do have to
321 write the glue-logic yourself though. Basically you can have
322 cryptsetup read the key from STDIN and write it there with your
323 own tool that in turn gets the key from the more secure key
327 * Can I resize a dm-crypt or LUKS partition?
329 Yes, you can, as neither dm-crypt nor LUKS stores partition size.
330 Whether you should is a different question. Personally I recommend
331 backup, recreation of the encrypted partition with new size,
332 recreation of the filesystem and restore. This gets around the
333 tricky business of resizing the filesystem. Resizing a dm-crypt or
334 LUKS container does not resize the filesystem in it. The backup is
335 really non-optional here, as a lot can go wrong, resulting in
336 partial or complete data loss. Using something like gparted to
337 resize an encrypted partition is slow, but typicaly works. This
338 will not change the size of the filesystem hidden under the
341 You also need to be aware of size-based limitations. The one
342 currently relevant is that aes-xts-plain should not be used for
343 encrypted container sizes larger than 2TiB. Use aes-xts-plain64
350 * My dm-crypt/LUKS mapping does not work! What general steps are
351 there to investigate the problem?
353 If you get a specific error message, investigate what it claims
354 first. If not, you may want to check the following things.
356 - Check that "/dev", including "/dev/mapper/control" is there. If it
357 is missing, you may have a problem with the "/dev" tree itself or
358 you may have broken udev rules.
360 - Check that you have the device mapper and the crypt target in your
361 kernel. The output of "dmsetup targets" should list a "crypt"
362 target. If it is not there or the command fails, add device mapper
363 and crypt-target to the kernel.
365 - Check that the hash-functions and ciphers you want to use are in
366 the kernel. The output of "cat /proc/crypto" needs to list them.
369 * My dm-crypt mapping suddenly stopped when upgrading cryptsetup.
371 The default cipher, hash or mode may have changed (the mode changed
372 from 1.0.x to 1.1.x). See under "Issues With Specific Versions of
376 * When I call cryptsetup from cron/CGI, I get errors about unknown
379 If you get errors about unknown parameters or the like that are not
380 present when cryptsetup is called from the shell, make sure you
381 have no older version of cryptsetup on your system that then gets
382 called by cron/CGI. For example some distributions install
383 cryptsetup into /usr/sbin, while a manual install could go to
384 /usr/local/sbin. As a debugging aid, call "cryptsetup --version"
385 from cron/CGI or the non-shell mechanism to be sure the right
389 * Unlocking a LUKS device takes very long. Why?
391 The iteration time for a key-slot (see Section 5 for an explanation
392 what iteration does) is calculated when setting a passphrase. By
393 default it is 1 second on the machine where the passphrase is set.
394 If you set a passphrase on a fast machine and then unlock it on a
395 slow machine, the unlocking time can be much longer. Also take into
396 account that up to 8 key-slots have to be tried in order to find the
399 If this is problem, you can add another key-slot using the slow
400 machine with the same passphrase and then remove the old key-slot.
401 The new key-slot will have an iteration count adjusted to 1 second
402 on the slow machine. Use luksKeyAdd and then luksKillSlot or
405 However, this operation will not change volume key iteration count
406 (MK iterations in output of "cryptsetup luksDump"). In order to
407 change that, you will have to backup the data in the LUKS
408 container (i.e. your encrypted data), luksFormat on the slow
409 machine and restore the data. Note that in the original LUKS
410 specification this value was fixed to 10, but it is now derived
411 from the PBKDF2 benchmark as well and set to iterations in 0.125
412 sec or 1000, whichever is larger. Also note that MK iterations
413 are not very security relevant. But as each key-slot already takes
414 1 second, spending the additional 0.125 seconds really does not
418 * "blkid" sees a LUKS UUID and an ext2/swap UUID on the same device.
421 Some old versions of cryptsetup have a bug where the header does
422 not get completely wiped during LUKS format and an older ext2/swap
423 signature remains on the device. This confuses blkid.
425 Fix: Wipe the unused header areas by doing a backup and restore of
426 the header with cryptsetup 1.1.x:
428 cryptsetup luksHeaderBackup --header-backup-file <file> <device>
429 cryptsetup luksHeaderRestore --header-backup-file <file> <device>
432 * cryptsetup segfaults on Gentoo amd64 hardened ...
434 There seems to be some inteference between the hardening and and
435 the way cryptsetup benchmarks PBKDF2. The solution to this is
436 currently not quite clear for an encrypted root filesystem. For
437 other uses, you can apparently specify USE="dynamic" as compile
438 flag, see http://bugs.gentoo.org/show_bug.cgi?id=283470
444 * I get the error "LUKS keyslot x is invalid." What does that mean?
446 This means that the given keyslot has an offset that points
447 outside the valid keyslot area. Typically, the reason is a
448 corrupted LUKS header because something was written to the start of
449 the device the LUKS contaner is on. Refer to Section "Backup and
450 Data Recovery" and ask on the mailing list if you have trouble
451 diagnosing and (if still possible) repairing this.
454 * Can a bad RAM module cause problems?
456 LUKS and dm-crypt can give the RAM quite a workout, especially when
457 combined with software RAID. In particular the combination RAID5 +
458 LUKS + XFS seems to uncover RAM problems that never caused obvious
459 problems before. Symptoms vary, but often the problem manifest
460 itself when copying large amounts of data, typically several times
461 larger than your main memory.
463 Side note: One thing you should always do on large data
464 copy/movements is to run a verify, for example with the "-d"
465 option of "tar" or by doing a set of MD5 checksums on the source
468 find . -type f -exec md5sum \{\} \; > checksum-file
470 and then a "md5sum -c checksum-file" on the other side. If you get
471 mismatches here, RAM is the primary suspect. A lesser suspect is
472 an overclocked CPU. I have found countless hardware problems in
473 verify runs after copying or making backups. Bit errors are much
474 more common than most people think.
476 Some RAM issues are even worse and corrupt structures in one of the
477 layers. This typically results in lockups, CPU state dumps in the
478 system logs, kernel panic or other things. It is quite possible to
479 have the problem with an encrypted device, but not with an
480 otherwise the same unencrypted device. The reason for that is that
481 encryption has an error amplification property: You flip one bit
482 in an encrypted data block, and the decrypted version has half of
483 its bits flipped. This is an important security property for modern
484 ciphers. With the usual modes in cryptsetup (CBC, ESSIV, XTS), you
485 get up to a completely changed 512 byte block per bit error. A
486 corrupt block causes a lot more havoc than the occasionally
487 flipped single bit and can result in various obscure errors.
489 Note, that a verify run on copying between encrypted or
490 unencrypted devices will reliably detect corruption, even when the
491 copying itself did not report any problems. If you find defect
492 RAM, assume all backups and copied data to be suspect, unless you
498 First you should know that overclocking often makes memory
499 problems worse. So if you overclock (which I strongly recommend
500 against in a system holding data that has some worth), run the
501 tests with the overclocking active.
503 There are two good options. One is Memtest86+ and the other is
504 "memtester" by Charles Cazabon. Memtest86+ requires a reboot and
505 then takes over the machine, while memtester runs from a
506 root-shell. Both use different testing methods and I have found
507 problems fast with each one that the other needed long to find. I
508 recommend running the following procedure until the first error is
511 - Run Memtest86+ for one cycle
513 - Run memterster for one cycle (shut down as many other applications
516 - Run Memtest86+ for 24h or more
518 - Run memtester for 24h or more
520 If all that does not produce error messages, your RAM may be sound,
521 but I have had one weak bit that Memtest86+ needed around 60 hours
522 to find. If you can reproduce the original problem reliably, a good
523 additional test may be to remove half of the RAM (if you have more
524 than one module) and try whether the problem is still there and if
525 so, try with the other half. If you just have one module, get a
526 different one and try with that. If you do overclocking, reduce
527 the settings to the most conservative ones available and try with
534 * Is LUKS insecure? Everybody can see I have encrypted data!
536 In practice it does not really matter. In most civilized countries
537 you can just refuse to hand over the keys, no harm done. In some
538 countries they can force you to hand over the keys, if they suspect
539 encryption. However the suspicion is enough, they do not have to
540 prove anything. This is for practical reasons, as even the presence
541 of a header (like the LUKS header) is not enough to prove that you
542 have any keys. It might have been an experiment, for example. Or it
543 was used as encrypted swap with a key from /dev/random. So they
544 make you prove you do not have encrypted data. Of course that is
545 just as impossible as the other way round.
547 This means that if you have a large set of random-looking data,
548 they can already lock you up. Hidden containers (encryption hidden
549 within encryption), as possible with Truecrypt, do not help
550 either. They will just assume the hidden container is there and
551 unless you hand over the key, you will stay locked up. Don't have
552 a hidden container? Though luck. Anybody could claim that.
554 Still, if you are concerned about the LUKS header, use plain
555 dm-crypt with a good passphrase. See also Section 2, "What is the
556 difference between "plain" and LUKS format?"
559 * Should I initialize (overwrite) a new LUKS/dm-crypt partition?
561 If you just create a filesystem on it, most of the old data will
562 still be there. If the old data is sensitive, you should overwrite
563 it before encrypting. In any case, not initializing will leave the
564 old data there until the specific sector gets written. That may
565 enable an attacker to determine how much and where on the
566 partition data was written. If you think this is a risk, you can
567 prevent this by overwriting the encrypted device (here assumed to
568 be named "e1") with zeros like this:
570 dd_rescue -w /dev/zero /dev/mapper/e1
572 or alternatively with one of the following more standard commands:
574 cat /dev/zero > /dev/mapper/e1
575 dd if=/dev/zero of=/dev/mapper/e1
578 * How do I securely erase a LUKS (or other) partition?
580 For LUKS, if you are in a desperate hurry, overwrite the LUKS
581 header and key-slot area. This means overwriting the first
582 (keyslots x stripes x keysize) + offset bytes. For the default
583 parameters, this is the 1'052'672 bytes, i.e. 1MiB + 4096 of the
584 LUKS partition. For 512 bit key length (e.g. for aes-xts-plain with
585 512 bit key) this is 2MiB. (The diferent offset stems from
586 differences in the sector alignment of the key-slots.) If in doubt,
587 just be generous and overwrite the first 10MB or so, it will likely
588 still be fast enough. A single overwrite with zeros should be
589 enough. If you anticipate being in a desperate hurry, prepare the
590 command beforehand. Example with /dev/sde1 as the LUKS partition
591 and default parameters:
593 head -c 1052672 /dev/zero > /dev/sde1; sync
595 A LUKS header backup or full backup will still grant access to
596 most or all data, so make sure that an attacker does not have
597 access to backups or destroy them as well.
599 If you have time, overwrite the whole LUKS partition with a single
600 pass of zeros. This is enough for current HDDs. For SSDs or FLASH
601 (USB sticks) you may want to overwrite the whole drive several
602 times to be sure data is not retained by wear leveling. This is
603 possibly still insecure as SSD technology is not fully understood
604 in this regard. Still, due to the anti-forensic properties of the
605 LUKS key-slots, a single overwrite of an SSD or FLASH drive could
606 be enough. If in doubt, use physical destruction in addition. Here
607 is a link to some current reseach results on erasing SSDs and FLASH
609 http://www.usenix.org/events/fast11/tech/full_papers/Wei.pdf
611 Keep in mind to also erase all backups.
613 Example for a zero-overwrite erase of partition sde1 done with
616 dd_rescue -w /dev/zero /dev/sde1
619 * How do I securely erase a backup of a LUKS partition or header?
621 That depends on the medium it is stored on. For HDD and SSD, use
622 overwrite with zeros. For an SSD or FLASH drive (USB stick), you
623 may want to overwrite the complete SSD several times and use
624 physical destruction in addition, see last item. For re-writable
625 CD/DVD, a single overwrite should also be enough, due to the
626 anti-forensic properties of the LUKS keyslots. For write-once
627 media, use physical destruction. For low security requirements,
628 just cut the CD/DVD into several parts. For high security needs,
629 shred or burn the medium. If your backup is on magnetic tape, I
630 advise physical destruction by shredding or burning, after
631 overwriting . The problem with magnetic tape is that it has a
632 higher dynamic range than HDDs and older data may well be
633 recoverable after overwrites. Also write-head alignment issues can
634 lead to data not actually being deleted at all during overwrites.
637 * What about backup? Does it compromise security?
639 That depends. See next section.
642 * Why is all my data permanently gone if I overwrite the LUKS header?
644 Overwriting the LUKS header in part or in full is the most common
645 reason why access to LUKS containers is lost permanently.
646 Overwriting can be done in a number of fashions, like creating a
647 new filesystem on the raw LUKS partition, making the raw partition
648 part of a raid array and just writing to the raw partition.
650 The LUKS header contains a 256 bit "salt" value and without that no
651 decryption is possible. While the salt is not secret, it is
652 key-grade material and cannot be reconstructed. This is a
653 cryptographically strong "cannot". From observations on the
654 cryptsetup mailing-list, people typically go though the usual
655 stages of grief (Denial, Anger, Bargaining, Depression, Acceptance)
656 when this happens to them. Observed times vary between 1 day and 2
657 weeks to complete the cycle. Seeking help on the mailing-list is
658 fine. Even if we usually cannot help with getting back your data,
659 most people found the feedback comforting.
661 If your header does not contain an intact salt, best go directly
662 to the last stage ("Acceptance") and think about what to do now.
663 There is one exception that I know of: If your LUKS container is
664 still open, then it may be possible to extract the master key from
665 the running system. See Item "How do I recover the master key from
666 a mapped LUKS container?" in Section "Backup and Data Recovery".
671 A salt is a random key-grade value added to the passphrase before
672 it is processed. It is not kept secret. The reason for using salts
673 is as follows: If an attacker wants to crack the password for a
674 single LUKS container, then every possible passphrase has to be
675 tried. Typically an attacker will not try every binary value, but
676 will try words and sentences from a dictionary.
678 If an attacker wants to attack several LUKS containers with the
679 same dictionary, then a different approach makes sense: Compute the
680 resulting slot-key for each dictionary element and store it on
681 disk. Then the test for each entry is just the slow unlocking with
682 the slot key (say 0.00001 sec) instead of calculating the slot-key
683 first (1 sec). For a single attack, this does not help. But if you
684 have more than one container to attack, this helps tremendously,
685 also because you can prepare your table before you even have the
686 container to attack! The calculation is also very simple to
687 parallelize. You could, for example, use the night-time unused CPU
688 power of your desktop PCs for this.
690 This is where the salt comes in. If the salt is combined with the
691 passphrase (in the simplest form, just appended to it), you
692 suddenly need a separate table for each salt value. With a
693 reasonably-sized salt value (256 bit, e.g.) this is quite
697 * Is LUKS secure with a low-entropy (bad) passphrase?
699 Note: You should only use the 94 printable characters from 7 bit
700 ASCII code to prevent your passphrase from failing when the
701 character encoding changes, e.g. because of a system upgrade, see
702 also the note at the very start of this FAQ under "WARNINGS".
704 This needs a bit of theory. The quality of your passphrase is
705 directly related to its entropy (information theoretic, not
706 thermodynamic). The entropy says how many bits of "uncertainty" or
707 "randomness" are in you passphrase. In other words, that is how
708 difficult guessing the passphrase is.
710 Example: A random English sentence has about 1 bit of entropy per
711 character. A random lowercase (or uppercase) character has about
714 Now, if n is the number of bits of entropy in your passphrase and t
715 is the time it takes to process a passphrase in order to open the
716 LUKS container, then an attacker has to spend at maximum
718 attack_time_max = 2^n * t
720 time for a successful attack and on average half that. There is no
721 way getting around that relationship. However, there is one thing
722 that does help, namely increasing t, the time it takes to use a
723 passphrase, see next FAQ item.
725 Still, if you want good security, a high-entropy passphrase is the
726 only option. Use at least 64 bits for secret stuff. That is 64
727 characters of English text (but only if randomly chosen) or a
728 combination of 12 truly random letters and digits.
730 For passphrase generation, do not use lines from very well-known
731 texts (religious texts, Harry potter, etc.) as they are to easy to
732 guess. For example, the total Harry Potter has about 1'500'000
733 words (my estimation). Trying every 64 character sequence starting
734 and ending at a word boundary would take only something like 20
735 days on a single CPU and is entirely feasible. To put that into
736 perspective, using a number of Amazon EC2 High-CPU Extra Large
737 instances (each gives about 8 real cores), this tests costs
738 currently about 50USD/EUR, but can be made to run arbitrarily fast.
740 On the other hand, choosing 1.5 lines from, say, the Wheel of Time
741 is in itself not more secure, but the book selection adds quite a
742 bit of entropy. (Now that I have mentioned it here, don't use tWoT
743 either!) If you add 2 or 3 typos or switch some words around, then
744 this is good passphrase material.
747 * What is "iteration count" and why is decreasing it a bad idea?
749 Iteration count is the number of PBKDF2 iterations a passphrase is
750 put through before it is used to unlock a key-slot. Iterations are
751 done with the explicit purpose to increase the time that it takes
752 to unlock a key-slot. This provides some protection against use of
753 low-entropy passphrases.
755 The idea is that an attacker has to try all possible passphrases.
756 Even if the attacker knows the passphrase is low-entropy (see last
757 item), it is possible to make each individual try take longer. The
758 way to do this is to repeatedly hash the passphrase for a certain
759 time. The attacker then has to spend the same time (given the same
760 computing power) as the user per try. With LUKS, the default is 1
761 second of PBKDF2 hashing.
763 Example 1: Lets assume we have a really bad passphrase (e.g. a
764 girlfriends name) with 10 bits of entropy. With the same CPU, an
765 attacker would need to spend around 500 seconds on average to
766 break that passphrase. Without iteration, it would be more like
767 0.0001 seconds on a modern CPU.
769 Example 2: The user did a bit better and has 32 chars of English
770 text. That would be about 32 bits of entropy. With 1 second
771 iteration, that means an attacker on the same CPU needs around 136
772 years. That is pretty impressive for such a weak passphrase.
773 Without the iterations, it would be more like 50 days on a modern
774 CPU, and possibly far less.
776 In addition, the attacker can both parallelize and use special
777 hardware like GPUs to speed up the attack. The attack can also
778 happen quite some time after the luksFormat operation and CPUs can
779 have become faster and cheaper. For that reason you want a bit
780 of extra security. Anyways, in Example 1 your are screwed. In
781 example 2, not necessarily. Even if the attack is faster, it still
782 has a certain cost associated with it, say 10000 EUR/USD with
783 iteration and 1 EUR/USD without iteration. The first can be
784 prohibitively expensive, while the second is something you try
785 even without solid proof that the decryption will yield something
788 The numbers above are mostly made up, but show the idea. Of course
789 the best thing is to have a high-entropy passphrase.
791 Would a 100 sec iteration time be even better? Yes and no.
792 Cryptographically it would be a lot better, namely 100 times better.
793 However, usability is a very important factor for security
794 technology and one that gets overlooked surprisingly often. For
795 LUKS, if you have to wait 2 minutes to unlock the LUKS container,
796 most people will not bother and use less secure storage instead. It
797 is better to have less protection against low-entropy passphrases
798 and people actually use LUKS, than having them do without
799 encryption altogether.
801 Now, what about decreasing the iteration time? This is generally a
802 very bad idea, unless you know and can enforce that the users only
803 use high-entropy passphrases. If you decrease the iteration time
804 without ensuring that, then you put your users at increased risk,
805 and considering how rarely LUKS containers are unlocked in a
806 typical work-flow, you do so without a good reason. Don't do it.
807 The iteration time is already low enough that users with entropy
808 low passphrases are vulnerable. Lowering it even further increases
809 this danger significantly.
812 * What about iteration count with plain dm-crypt?
814 Simple: There is none. There is also no salting. If you use plain
815 dm-crypt, the only way to be secure is to use a high entropy
816 passphrase. If in doubt, use LUKS instead.
819 * Is LUKS with default parameters less secure on a slow CPU?
821 Unfortunately, yes. However the only aspect affected is the
822 protection for low-entropy passphrase or master-key. All other
823 security aspects are independent of CPU speed.
825 The master key is less critical, as you really have to work at it
826 to give it low entropy. One possibility is to supply the master key
827 yourself. If that key is low-entropy, then you get what you
828 deserve. The other known possibility is to use /dev/urandom for
829 key generation in an entropy-startved situation (e.g. automatic
830 installation on an embedded device without network and other entropy
833 For the passphrase, don't use a low-entropy passphrase. If your
834 passphrase is good, then a slow CPU will not matter. If you insist
835 on a low-entropy passphrase on a slow CPU, use something like
836 "--iter-time=10" or higher and wait a long time on each LUKS unlock
837 and pray that the attacker does not find out in which way exactly
838 your passphrase is low entropy. This also applies to low-entropy
839 passphrases on fast CPUs. Technology can do only so much to
840 compensate for problems in front of the keyboard.
843 * Why was the default aes-cbc-plain replaced with aes-cbc-essiv?
845 The problem is that cbc-plain has a fingerprint vulnerability, where
846 a specially crafted file placed into the crypto-container can be
847 recognized from the outside. The issue here is that for cbc-plain
848 the initialization vector (IV) is the sector number. The IV gets
849 XORed to the first data chunk of the sector to be encrypted. If you
850 make sure that the first data block to be stored in a sector
851 contains the sector number as well, the first data block to be
852 encrypted is all zeros and always encrypted to the same ciphertext.
853 This also works if the first data chunk just has a constant XOR
854 with the sector number. By having several shifted patterns you can
855 take care of the case of a non-power-of-two start sector number of
858 This mechanism allows you to create a pattern of sectors that have
859 the same first ciphertext block and signal one bit per sector to the
860 outside, allowing you to e.g. mark media files that way for
861 recognition without decryption. For large files this is a
862 practical attack. For small ones, you do not have enough blocks to
863 signal and take care of different file starting offsets.
865 In order to prevent this attack, the default was changed to
866 cbc-essiv. ESSIV uses a keyed hash of the sector number, with the
867 encryption key as key. This makes the IV unpredictable without
868 knowing the encryption key and the watermarking attack fails.
871 * Are there any problems with "plain" IV? What is "plain64"?
873 First, "plain" and "plain64" are both not secure to use with CBC,
874 see previous FAQ item.
876 However there are modes, like XTS, that are secure with "plain" IV.
877 The next limit is that "plain" is 64 bit, with the upper 32 bit set
878 to zero. This means that on volumes larger than 2TiB, the IV
879 repeats, creating a vulnerability that potentially leaks some
880 data. To avoid this, use "plain64", which uses the full sector
881 number up to 64 bit. Note that "plain64" requires a kernel >=
882 2.6.33. Also note that "plain64" is backwards compatible for
883 volume sizes <= 2TiB, but not for those > 2TiB. Finally, "plain64"
884 does not cause any performance penalty compared to "plain".
887 * What about XTS mode?
889 XTS mode is potentially even more secure than cbc-essiv (but only if
890 cbc-essiv is insecure in your scenario). It is a NIST standard and
891 used, e.g. in Truecrypt. At the moment, if you want to use it, you
892 have to specify it manually as "aes-xts-plain", i.e.
894 cryptsetup -c aes-xts-plain luksFormat <device>
896 For volumes >2TiB and kernels >= 2.6.33 use "plain64" (see FAQ
897 item on "plain" and "plain64"):
899 cryptsetup -c aes-xts-plain64 luksFormat <device>
901 There is a potential security issue with XTS mode and large blocks.
902 LUKS and dm-crypt always use 512B blocks and the issue does not
906 6. Backup and Data Recovery
909 * Why do I need Backup?
911 First, disks die. The rate for well-treated (!) disk is about 5%
912 per year, which is high enough to worry about. There is some
913 indication that this may be even worse for some SSDs. This applies
914 both to LUKS and plain dm-crypt partitions.
916 Second, for LUKS, if anything damages the LUKS header or the
917 key-stripe area then decrypting the LUKS device can become
918 impossible. This is a frequent occuurence. For example an
919 accidental format as FAT or some software overwriting the first
920 sector where it suspects a partition boot sector typically makes a
921 LUKS partition permanently inacessible. See more below on LUKS
924 So, data-backup in some form is non-optional. For LUKS, you may
925 also want to store a header backup in some secure location. This
926 only needs an update if you change passphrases.
929 * How do I backup a LUKS header?
931 While you could just copy the appropriate number of bytes from the
932 start of the LUKS partition, the best way is to use command option
933 "luksHeaderBackup" of cryptsetup. This protects also against
934 errors when non-standard parameters have been used in LUKS
935 partition creation. Example:
938 cryptsetup luksHeaderBackup --header-backup-file h /dev/mapper/c1
940 To restore, use the inverse command, i.e.
942 cryptsetup luksHeaderRestore --header-backup-file h /dev/mapper/c1
945 * How do I backup a LUKS or dm-crypt partition?
947 There are two options, a sector-image and a plain file or
948 filesystem backup of the contents of the partition. The sector
949 image is already encrypted, but cannot be compressed and contains
950 all empty space. The filesystem backup can be compressed, can
951 contain only part of the encrypted device, but needs to be
952 encrypted separately if so desired.
954 A sector-image will contain the whole partition in encrypted form,
955 for LUKS the LUKS header, the keys-slots and the data area. It can
956 be done under Linux e.g. with dd_rescue (for a direct image copy)
957 and with "cat" or "dd". Example:
959 cat /dev/sda10 > sda10.img
960 dd_rescue /dev/sda10 sda10.img
962 You can also use any other backup software that is capable of making
963 a sector image of a partition. Note that compression is
964 ineffective for encrypted data, hence it does not make sense to
967 For a filesystem backup, you decrypt and mount the encrypted
968 partition and back it up as you would a normal filesystem. In this
969 case the backup is not encrypted, unless your encryption method
970 does that. For example you can encrypt a backup with "tar" as
973 tar cjf - <path> | gpg --cipher-algo AES -c - > backup.tbz2.gpg
975 And verify the backup like this if you are at "path":
977 cat backup.tbz2.gpg | gpg - | tar djf -
979 Note: Allways verify backups, especially encrypted ones.
981 In both cases GnuPG will ask you interactively for your symmetric
982 key. The verify will only output errors. Use "tar dvjf -" to get
983 all comparison results. To make sure no data is written to disk
984 unencrypted, turn off swap if it is not encrypted before doing the
987 You can of course use different or no compression and you can use
988 an asymmetric key if you have one and have a backup of the secret
989 key that belongs to it.
991 A second option for a filestem-level backup that can be used when
992 the backup is also on local disk (e.g. an external USB drive) is
993 to use a LUKS container there and copy the files to be backed up
994 between both mounted containers. Also see next item.
997 * Do I need a backup of the full partition? Would the header and
998 key-slots not be enough?
1000 Backup protects you against two things: Disk loss or corruption
1001 and user error. By far the most questions on the dm-crypt mailing
1002 list about how to recover a damaged LUKS partition are related
1003 to user error. For example, if you create a new filesystem on a
1004 LUKS partition, chances are good that all data is lost
1007 For this case, a header+key-slot backup would often be enough. But
1008 keep in mind that a well-treated (!) HDD has roughly a failure
1009 risk of 5% per year. It is highly advisable to have a complete
1010 backup to protect against this case.
1013 * *What do I need to backup if I use "decrypt_derived"?
1015 This is a script in Debian, intended for mounting /tmp or swap with
1016 a key derived from the master key of an already decrypted device.
1017 If you use this for an device with data that should be persistent,
1018 you need to make sure you either do not lose access to that master
1019 key or have a backup of the data. If you derive from a LUKS
1020 device, a header backup of that device would cover backing up the
1021 master key. Keep in mind that this does not protect against disk
1024 Note: If you recreate the LUKS header of the device you derive from
1025 (using luksFormat), the master key changes even if you use the same
1026 passphrase(s) and you will not be able to decrypt the derived
1027 device with the new LUKS header.
1030 * Does a backup compromise security?
1032 Depends on how you do it. However if you do not have one, you are
1033 going to eventually lose your encrypted data.
1035 There are risks introduced by backups. For example if you
1036 change/disable a key-slot in LUKS, a binary backup of the partition
1037 will still have the old key-slot. To deal with this, you have to
1038 be able to change the key-slot on the backup as well, securely
1039 erase the backup or do a filesystem-level backup instead of a binary
1042 If you use dm-crypt, backup is simpler: As there is no key
1043 management, the main risk is that you cannot wipe the backup when
1044 wiping the original. However wiping the original for dm-crypt
1045 should consist of forgetting the passphrase and that you can do
1046 without actual access to the backup.
1048 In both cases, there is an additional (usually small) risk with
1049 binary backups: An attacker can see how many sectors and which
1050 ones have been changed since the backup. To prevent this, use a
1051 filesystem level backup methid that encrypts the whole backup in
1052 one go, e.g. as described above with tar and GnuPG.
1054 My personal advice is to use one USB disk (low value data) or
1055 three disks (high value data) in rotating order for backups, and
1056 either use independent LUKS partitions on them, or use encrypted
1057 backup with tar and GnuPG.
1059 If you do network-backup or tape-backup, I strongly recommend to
1060 go the filesystem backup path with independent encryption, as you
1061 typically cannot reliably delete data in these scenarios,
1062 especially in a cloud setting. (Well, you can burn the tape if it
1063 is under your control...)
1066 * What happens if I overwrite the start of a LUKS partition or damage
1067 the LUKS header or key-slots?
1069 There are two critical components for decryption: The salt values
1070 in the header itself and the key-slots. If the salt values are
1071 overwritten or changed, nothing (in the cryptographically strong
1072 sense) can be done to access the data, unless there is a backup
1073 of the LUKS header. If a key-slot is damaged, the data can still
1074 be read with a different key-slot, if there is a remaining
1075 undamaged and used key-slot. Note that in order to make a key-slot
1076 unrecoverable in a cryptographically strong sense, changing about
1077 4-6 bits in random locations of its 128kiB size is quite enough.
1080 * What happens if I (quick) format a LUKS partition?
1082 I have not tried the different ways to do this, but very likely you
1083 will have written a new boot-sector, which in turn overwrites the
1084 LUKS header, including the salts, making your data permanently
1085 irretrivable, unless you have a LUKS header backup. You may also
1086 damage the key-slots in part or in full. See also last item.
1089 * How do I recover the master key from a mapped LUKS container?
1091 This is typically only needed if you managed to damage your LUKS
1092 header, but the container is still mapped, i.e. "luksOpen"ed. It
1093 also helps if you have a mapped container that you forgot or do not
1094 know a passphrase for (e.g. on a long running server.)
1096 WARNING: Things go wrong, do a full backup before trying this!
1098 WARNING: This exposes the master key of the LUKS container. Note
1099 that both ways to recreate a LUKS header with the old master key
1100 described below will write the master key to disk. Unless you are
1101 sure you have securely erased it afterwards, e.g. by writing it to
1102 an encrypted partition, RAM disk or by erasing the filesystem you
1103 wrote it to by a complete overwrite, you should change the master
1104 key afterwards. Changing the master key requires a full data
1105 backup, luksFormat and then restore of the backup.
1107 First, there is a script by Milan that automatizes the whole
1108 process, except generating a new LUKS header with the old master
1109 key (it prints the command for that though):
1111 http://code.google.com/p/cryptsetup/source/browse/trunk/misc/luks-header-from-active
1113 You can also do this manually. Here is how:
1115 - Get the master key from the device mapper. This is done by the
1116 following command. Substitute c5 for whatever you mapped to:
1118 # dmsetup table --target crypt --showkey /dev/mapper/c5
1120 0 200704 crypt aes-cbc-essiv:sha256
1121 a1704d9715f73a1bb4db581dcacadaf405e700d591e93e2eaade13ba653d0d09
1124 The result is actually one line, wrapped here for clarity. The long
1125 hex string is the master key.
1127 - Convert the master key to a binary file representation. You can
1128 do this manually, e.g. with hexedit. You can also use the tool
1129 "xxd" from vim like this:
1131 echo "a1704d9....53d0d09" | xxd -r -p > <master-key-file>
1133 - Do a luksFormat to create a new LUKS header.
1135 NOTE: If your header is intact and you just forgot the
1136 passphrase, you can just set a new passphrase, see next subitem.
1138 Unmap the device before you do that (luksClose). Then do
1140 cryptsetup luksFormat --master-key-file=<master-key-file> <luks device>
1142 Note that if the container was created with other than the default
1143 settings of the cryptsetup version you are using, you need to give
1144 additional parameters specifying the deviations. If in doubt, try
1145 the script by Milan. It does recover the other parameters as well.
1147 Side note: This is the way the decrypt_derived script gets at the
1148 master key. It just omits the conversion and hashes the master key
1151 - If the header is intact and you just forgot the passphrase, just
1152 set a new passphrase like this:
1154 cryptsetup luksAddKey --master-key-file=<master-key-file> <luks device>
1156 You may want to disable the old one afterwards.
1159 * What does the on-disk structure of dm-crypt look like?
1161 There is none. dm-crypt takes a block device and gives encrypted
1162 access to each of its blocks with a key derived from the passphrase
1163 given. If you use a cipher different than the default, you have to
1164 specify that as a parameter to cryptsetup too. If you want to
1165 change the password, you basically have to create a second
1166 encrypted device with the new passphrase and copy your data over.
1167 On the plus side, if you accidentally overwrite any part of a
1168 dm-crypt device, the damage will be limited to the are you
1172 * What does the on-disk structure of LUKS look like?
1174 A LUKS partition consists of a header, followed by 8 key-slot
1175 descriptors, followed by 8 key slots, followed by the encrypted
1178 Header and key-slot descriptors fill the first 592 bytes. The
1179 key-slot size depends on the creation parameters, namely on the
1180 number of anti-forensic stripes, key material offset and master
1183 With the default parameters, each key-slot is a bit less than
1184 128kiB in size. Due to sector alignment of the key-slot start,
1185 that means the key block 0 is at offset 0x1000-0x20400, key
1186 block 1 at offset 0x21000-0x40400, and key block 7 at offset
1187 0xc1000-0xe0400. The space to the next full sector address is
1188 padded with zeros. Never used key-slots are filled with what the
1189 disk originally contained there, a key-slot removed with
1190 "luksRemoveKey" or "luksKillSlot" gets filled with 0xff. Due to
1191 2MiB default alignment, start of the data area for cryptsetup 1.3
1192 and later is at 2MiB, i.e. at 0x200000. For older versions, it is
1193 at 0x101000, i.e. at 1'052'672 bytes, i.e. at 1MiB + 4096 bytes
1194 from the start of the partition. Incidentally, "luksHeaderBackup"
1195 for a LUKS container created with default parameters dumps exactly
1196 the first 2MiB (or 1'052'672 bytes for headers created with
1197 cryptsetup versions < 1.3) to file and "luksHeaderRestore" restores
1200 For non-default parameters, you have to figure out placement
1201 yourself. "luksDump" helps. See also next item. For the most common
1202 non-default settings, namely aes-xts-plain with 512 bit key, the
1203 offsets are: 1st keyslot 0x1000-0x3f800, 2nd keyslot
1204 0x40000-0x7e000, 3rd keyslot 0x7e000-0xbd800, ..., and start of
1205 bulk data at 0x200000.
1207 The exact specification of the format is here:
1208 http://code.google.com/p/cryptsetup/wiki/Specification
1211 * What is the smallest possible LUKS container?
1213 Note: From cryptsetup 1.3 onwards, alignment is set to 1MB. With
1214 modern Linux partitioning tools that also align to 1MB, this will
1215 result in aligmnet to 2k secors and typical Flash/SSD sectors,
1216 which is highly desirable for a number of reasons. Changing the
1217 alignment is not recomended.
1219 That said, with default parameters, the data area starts at
1220 exactly 2MB offset (at 0x101000 for crptsetup versions before 1.3).
1221 The smallest data area you can have is one sector of 512 bytes.
1222 Data areas of 0 bytes can be created, but fail on mapping.
1224 While you cannot put a filesystem into something this small, it may
1225 still be used to contain, for eamcple, key. Note that with current
1226 formatting tools, a partition for a container this size will be
1227 3MiB anyways. If you put the LUKS container into a file (via
1228 losetup and a loopback device), the file needs to be 2097664 bytes
1229 in size, i.e. 2MiB + 512B.
1231 The two ways to influence the start of the data area are key-size
1234 For alignment, you can go down to 1 on the parameter. This will
1235 still leave you with a data-area starting at 0x101000, i.e.
1236 1MiB+4096B (default parameters) as alignment will be rounded up to
1237 the next multiple of 8 (i.e. 4096 bytes) (TODO: need to verify
1240 For key-size, you can use 128 bit (e.g. AES-128 with CBC), 256 bit
1241 (e.g. AES-256 with CBC) or 512 bit (e.g. AES-256 with XTS mode).
1242 You can do 64 bit (e.g. blofish-64 with CBC), but anything below
1243 128 bit has to be considered insecure today.
1245 Example 1 - AES 128 bit with CBC:
1247 cryptsetup luksFormat -s 128 --align-payload=8 <device>
1249 This results in a data offset of 0x81000, i.e. 516KiB or 528384
1250 bytes. Add one 512 byte sector and the smallest LUKS container size
1251 with these parameters is 516KiB + 512B or 528896 bytes.
1253 Example 2 - Blowfish 64 bit with CBC (WARNING: insecure):
1255 cryptsetup luksFormat -c blowfish -s 64 --align-payload=8 /dev/loop0
1257 This results in a data offset of 0x41000, i.e. 260kiB or 266240
1258 bytes, with a minimal LUKS conatiner size of 260kiB + 512B or
1262 * I think this is overly complicated. Is there an alternative?
1264 Not really. Encryption comes at a price. You can use plain
1265 dm-crypt to simplify things a bit. It does not allow multiple
1266 passphrases, but on the plus side, it has zero on disk description
1267 and if you overwrite some part of a plain dm-crypt partition,
1268 exactly the overwritten parts are lost (rounded up to sector
1272 7. Interoperability with other Disk Encryption Tools
1275 * What is this section about?
1277 Cryptsetup for plain dm-crypt can be used to access a number of
1278 on-disk formats created by tools like loop-aes patched into
1279 losetup. This somtimes works and sometimes does not. This section
1280 collects insights into what works, what does not and where more
1281 information is required.
1283 Additional information may be found in the mailing-list archives,
1284 mentioned at the start of this FAQ document. If you have a
1285 solution working that is not yet documented here and think a wider
1286 audience may be intertested, please email the FAQ maintainer.
1289 * loop-aes: General observations.
1291 One problem is that there are different versions of losetup around.
1292 loop-aes is a patch for losetup. Possible problems and deviations
1293 from cryptsetup option syntax include:
1295 - Offsets specifed in bytes (cryptsetup: 512 byte sectors)
1297 - The need to specify an IV offset
1299 - Encryption mode needs specifying (e.g. "-c twofish-cbc-plain")
1301 - Key size needs specifying (e.g. "-s 128" for 128 bit keys)
1303 - Passphrase hash algorithm needs specifying
1305 Also note that because plain dm-crypt and loop-aes format does not
1306 have metadata, autodetection, while feasible in most cases, would
1307 be a lot of work that nobody really wants to do. If you still have
1308 the old set-up, using a verbosity option (-v) on mapping with the
1309 old tool or having a look into the system logs after setup could
1310 give you the information you need.
1313 * loop-aes patched into losetup on debian 5.x, kernel 2.6.32
1315 In this case, the main problem seems to be that this variant of
1316 losetup takes the offset (-o option) in bytes, while cryptsetup
1317 takes it in sectors of 512 bytes each. Example: The losetupp
1320 losetup -e twofish -o 2560 /dev/loop0 /dev/sdb1
1321 mount /dev/loop0 mountpoint
1325 cryptsetup create -c twofish -o 5 --skip 5 e1 /dev/sdb1
1326 mount /dev/mapper/e1 mountpoint
1329 * loop-aes with 160 bit key
1331 This seems to be sometimes used with twofish and blowfish and
1332 represents a 160 bit ripemed160 hash output padded to 196 bit key
1333 length. It seems the corresponding options for cryptsetup are
1335 --cipher twofish-cbc-null -s 192 -h ripemd160:20
1338 8. Issues with Specific Versions of cryptsetup
1341 * When using the create command for plain dm-crypt with cryptsetup
1342 1.1.x, the mapping is incompatible and my data is not accessible
1345 With cryptsetup 1.1.x, the distro maintainer can define different
1346 default encryption modes for LUKS and plain devices. You can check
1347 these compiled-in defaults using "cryptsetup --help". Moreover, the
1348 plain device default changed because the old IV mode was
1349 vulnerable to a watermarking attack.
1351 If you are using a plain device and you need a compatible mode, just
1352 specify cipher, key size and hash algorithm explicitly. For
1353 compatibility with cryptsetup 1.0.x defaults, simple use the
1356 cryptsetup create -c aes-cbc-plain -s 256 -h ripemd160 <name> <dev>
1358 LUKS stores cipher and mode in the metadata on disk, avoiding this
1362 * cryptsetup on SLED 10 has problems...
1364 SLED 10 is missing an essential kernel patch for dm-crypt, which
1365 is broken in its kernel as a result. There may be a very old
1366 version of cryptsetup (1.0.x) provided by SLED, which should also
1367 not be used anymore as well. My advice would be to drop SLED 10.
1369 A. Contributors In no particular order: