1 This is gcrypt.info, produced by makeinfo version 6.3 from gcrypt.texi.
3 This manual is for Libgcrypt (version 1.8.4, 24 October 2018), which is
4 GNU's library of cryptographic building blocks.
6 Copyright (C) 2000, 2002, 2003, 2004, 2006, 2007, 2008, 2009, 2011, 2012
7 Free Software Foundation, Inc.
8 Copyright (C) 2012, 2013, 2016, 2017 g10 Code GmbH
10 Permission is granted to copy, distribute and/or modify this
11 document under the terms of the GNU General Public License as
12 published by the Free Software Foundation; either version 2 of the
13 License, or (at your option) any later version. The text of the
14 license can be found in the section entitled "GNU General Public
16 INFO-DIR-SECTION GNU Libraries
18 * libgcrypt: (gcrypt). Cryptographic function library.
22 File: gcrypt.info, Node: Top, Next: Introduction, Up: (dir)
27 This manual is for Libgcrypt (version 1.8.4, 24 October 2018), which is
28 GNU's library of cryptographic building blocks.
30 Copyright (C) 2000, 2002, 2003, 2004, 2006, 2007, 2008, 2009, 2011, 2012
31 Free Software Foundation, Inc.
32 Copyright (C) 2012, 2013, 2016, 2017 g10 Code GmbH
34 Permission is granted to copy, distribute and/or modify this
35 document under the terms of the GNU General Public License as
36 published by the Free Software Foundation; either version 2 of the
37 License, or (at your option) any later version. The text of the
38 license can be found in the section entitled "GNU General Public
43 * Introduction:: What is Libgcrypt.
44 * Preparation:: What you should do before using the library.
45 * Generalities:: General library functions and data types.
46 * Handler Functions:: Working with handler functions.
47 * Symmetric cryptography:: How to use symmetric cryptography.
48 * Public Key cryptography:: How to use public key cryptography.
49 * Hashing:: How to use hash algorithms.
50 * Message Authentication Codes:: How to use MAC algorithms.
51 * Key Derivation:: How to derive keys from strings
52 * Random Numbers:: How to work with random numbers.
53 * S-expressions:: How to manage S-expressions.
54 * MPI library:: How to work with multi-precision-integers.
55 * Prime numbers:: How to use the Prime number related functions.
56 * Utilities:: Utility functions.
57 * Tools:: Utility tools.
58 * Configuration:: Configuration files and environment variables.
59 * Architecture:: How Libgcrypt works internally.
63 * Self-Tests:: Description of the self-tests.
64 * FIPS Mode:: Description of the FIPS mode.
65 * Library Copying:: The GNU Lesser General Public License
66 says how you can copy and share Libgcrypt.
67 * Copying:: The GNU General Public License says how you
68 can copy and share some parts of Libgcrypt.
72 * Figures and Tables:: Index of figures and tables.
73 * Concept Index:: Index of concepts and programs.
74 * Function and Data Index:: Index of functions, variables and data types.
77 File: gcrypt.info, Node: Introduction, Next: Preparation, Prev: Top, Up: Top
82 Libgcrypt is a library providing cryptographic building blocks.
86 * Getting Started:: How to use this manual.
87 * Features:: A glance at Libgcrypt's features.
88 * Overview:: Overview about the library.
91 File: gcrypt.info, Node: Getting Started, Next: Features, Up: Introduction
96 This manual documents the Libgcrypt library application programming
97 interface (API). All functions and data types provided by the library
100 The reader is assumed to possess basic knowledge about applied
103 This manual can be used in several ways. If read from the beginning
104 to the end, it gives a good introduction into the library and how it can
105 be used in an application. Forward references are included where
106 necessary. Later on, the manual can be used as a reference manual to
107 get just the information needed about any particular interface of the
108 library. Experienced programmers might want to start looking at the
109 examples at the end of the manual, and then only read up those parts of
110 the interface which are unclear.
113 File: gcrypt.info, Node: Features, Next: Overview, Prev: Getting Started, Up: Introduction
118 Libgcrypt might have a couple of advantages over other libraries doing a
122 Anybody can use, modify, and redistribute it under the terms of the
123 GNU Lesser General Public License (*note Library Copying::). Note,
124 that some parts (which are in general not needed by applications)
125 are subject to the terms of the GNU General Public License (*note
126 Copying::); please see the README file of the distribution for of
129 It encapsulates the low level cryptography
130 Libgcrypt provides a high level interface to cryptographic building
131 blocks using an extensible and flexible API.
134 File: gcrypt.info, Node: Overview, Prev: Features, Up: Introduction
139 The Libgcrypt library is fully thread-safe, where it makes sense to be
140 thread-safe. Not thread-safe are some cryptographic functions that
141 modify a certain context stored in handles. If the user really intents
142 to use such functions from different threads on the same handle, he has
143 to take care of the serialization of such functions himself. If not
144 described otherwise, every function is thread-safe.
146 Libgcrypt depends on the library 'libgpg-error', which contains some
147 common code used by other GnuPG components.
150 File: gcrypt.info, Node: Preparation, Next: Generalities, Prev: Introduction, Up: Top
155 To use Libgcrypt, you have to perform some changes to your sources and
156 the build system. The necessary changes are small and explained in the
157 following sections. At the end of this chapter, it is described how the
158 library is initialized, and how the requirements of the library are
163 * Header:: What header file you need to include.
164 * Building sources:: How to build sources using the library.
165 * Building sources using Automake:: How to build sources with the help of Automake.
166 * Initializing the library:: How to initialize the library.
167 * Multi-Threading:: How Libgcrypt can be used in a MT environment.
168 * Enabling FIPS mode:: How to enable the FIPS mode.
169 * Hardware features:: How to disable hardware features.
172 File: gcrypt.info, Node: Header, Next: Building sources, Up: Preparation
177 All interfaces (data types and functions) of the library are defined in
178 the header file 'gcrypt.h'. You must include this in all source files
179 using the library, either directly or through some other header file,
184 The name space of Libgcrypt is 'gcry_*' for function and type names
185 and 'GCRY*' for other symbols. In addition the same name prefixes with
186 one prepended underscore are reserved for internal use and should never
187 be used by an application. Note that Libgcrypt uses libgpg-error, which
188 uses 'gpg_*' as name space for function and type names and 'GPG_*' for
189 other symbols, including all the error codes.
191 Certain parts of gcrypt.h may be excluded by defining these macros:
193 'GCRYPT_NO_MPI_MACROS'
194 Do not define the shorthand macros 'mpi_*' for 'gcry_mpi_*'.
196 'GCRYPT_NO_DEPRECATED'
197 Do not include definitions for deprecated features. This is useful
198 to make sure that no deprecated features are used.
201 File: gcrypt.info, Node: Building sources, Next: Building sources using Automake, Prev: Header, Up: Preparation
206 If you want to compile a source file including the 'gcrypt.h' header
207 file, you must make sure that the compiler can find it in the directory
208 hierarchy. This is accomplished by adding the path to the directory in
209 which the header file is located to the compilers include file search
210 path (via the '-I' option).
212 However, the path to the include file is determined at the time the
213 source is configured. To solve this problem, Libgcrypt ships with a
214 small helper program 'libgcrypt-config' that knows the path to the
215 include file and other configuration options. The options that need to
216 be added to the compiler invocation at compile time are output by the
217 '--cflags' option to 'libgcrypt-config'. The following example shows
218 how it can be used at the command line:
220 gcc -c foo.c `libgcrypt-config --cflags`
222 Adding the output of 'libgcrypt-config --cflags' to the compiler’s
223 command line will ensure that the compiler can find the Libgcrypt header
226 A similar problem occurs when linking the program with the library.
227 Again, the compiler has to find the library files. For this to work,
228 the path to the library files has to be added to the library search path
229 (via the '-L' option). For this, the option '--libs' to
230 'libgcrypt-config' can be used. For convenience, this option also
231 outputs all other options that are required to link the program with the
232 Libgcrypt libraries (in particular, the '-lgcrypt' option). The example
233 shows how to link 'foo.o' with the Libgcrypt library to a program 'foo'.
235 gcc -o foo foo.o `libgcrypt-config --libs`
237 Of course you can also combine both examples to a single command by
238 specifying both options to 'libgcrypt-config':
240 gcc -o foo foo.c `libgcrypt-config --cflags --libs`
243 File: gcrypt.info, Node: Building sources using Automake, Next: Initializing the library, Prev: Building sources, Up: Preparation
245 2.3 Building sources using Automake
246 ===================================
248 It is much easier if you use GNU Automake instead of writing your own
249 Makefiles. If you do that, you do not have to worry about finding and
250 invoking the 'libgcrypt-config' script at all. Libgcrypt provides an
251 extension to Automake that does all the work for you.
253 -- Macro: AM_PATH_LIBGCRYPT ([MINIMUM-VERSION], [ACTION-IF-FOUND],
254 [ACTION-IF-NOT-FOUND])
255 Check whether Libgcrypt (at least version MINIMUM-VERSION, if
256 given) exists on the host system. If it is found, execute
257 ACTION-IF-FOUND, otherwise do ACTION-IF-NOT-FOUND, if given.
259 Additionally, the function defines 'LIBGCRYPT_CFLAGS' to the flags
260 needed for compilation of the program to find the 'gcrypt.h' header
261 file, and 'LIBGCRYPT_LIBS' to the linker flags needed to link the
262 program to the Libgcrypt library. If the used helper script does
263 not match the target type you are building for a warning is printed
264 and the string 'libgcrypt' is appended to the variable
265 'gpg_config_script_warn'.
267 This macro searches for 'libgcrypt-config' along the PATH. If you
268 are cross-compiling, it is useful to set the environment variable
269 'SYSROOT' to the top directory of your target. The macro will then
270 first look for the helper program in the 'bin' directory below that
271 top directory. An absolute directory name must be used for
272 'SYSROOT'. Finally, if the configure command line option
273 '--with-libgcrypt-prefix' is used, only its value is used for the
274 top directory below which the helper script is expected.
276 You can use the defined Autoconf variables like this in your
279 AM_CPPFLAGS = $(LIBGCRYPT_CFLAGS)
280 LDADD = $(LIBGCRYPT_LIBS)
283 File: gcrypt.info, Node: Initializing the library, Next: Multi-Threading, Prev: Building sources using Automake, Up: Preparation
285 2.4 Initializing the library
286 ============================
288 Before the library can be used, it must initialize itself. This is
289 achieved by invoking the function 'gcry_check_version' described below.
291 Also, it is often desirable to check that the version of Libgcrypt
292 used is indeed one which fits all requirements. Even with binary
293 compatibility, new features may have been introduced, but due to problem
294 with the dynamic linker an old version may actually be used. So you may
295 want to check that the version is okay right after program startup.
297 -- Function: const char * gcry_check_version (const char *REQ_VERSION)
299 The function 'gcry_check_version' initializes some subsystems used
300 by Libgcrypt and must be invoked before any other function in the
301 library. *Note Multi-Threading::.
303 Furthermore, this function returns the version number of the
304 library. It can also verify that the version number is higher than
305 a certain required version number REQ_VERSION, if this value is not
308 Libgcrypt uses a concept known as secure memory, which is a region of
309 memory set aside for storing sensitive data. Because such memory is a
310 scarce resource, it needs to be setup in advanced to a fixed size.
311 Further, most operating systems have special requirements on how that
312 secure memory can be used. For example, it might be required to install
313 an application as "setuid(root)" to allow allocating such memory.
314 Libgcrypt requires a sequence of initialization steps to make sure that
315 this works correctly. The following examples show the necessary steps.
317 If you don't have a need for secure memory, for example if your
318 application does not use secret keys or other confidential data or it
319 runs in a controlled environment where key material floating around in
320 memory is not a problem, you should initialize Libgcrypt this way:
322 /* Version check should be the very first call because it
323 makes sure that important subsystems are initialized. */
324 if (!gcry_check_version (GCRYPT_VERSION))
326 fputs ("libgcrypt version mismatch\n", stderr);
330 /* Disable secure memory. */
331 gcry_control (GCRYCTL_DISABLE_SECMEM, 0);
333 /* ... If required, other initialization goes here. */
335 /* Tell Libgcrypt that initialization has completed. */
336 gcry_control (GCRYCTL_INITIALIZATION_FINISHED, 0);
338 If you have to protect your keys or other information in memory
339 against being swapped out to disk and to enable an automatic overwrite
340 of used and freed memory, you need to initialize Libgcrypt this way:
342 /* Version check should be the very first call because it
343 makes sure that important subsystems are initialized. */
344 if (!gcry_check_version (GCRYPT_VERSION))
346 fputs ("libgcrypt version mismatch\n", stderr);
350 /* We don't want to see any warnings, e.g. because we have not yet
351 parsed program options which might be used to suppress such
353 gcry_control (GCRYCTL_SUSPEND_SECMEM_WARN);
355 /* ... If required, other initialization goes here. Note that the
356 process might still be running with increased privileges and that
357 the secure memory has not been initialized. */
359 /* Allocate a pool of 16k secure memory. This makes the secure memory
360 available and also drops privileges where needed. Note that by
361 using functions like gcry_xmalloc_secure and gcry_mpi_snew Libgcrypt
362 may expand the secure memory pool with memory which lacks the
363 property of not being swapped out to disk. */
364 gcry_control (GCRYCTL_INIT_SECMEM, 16384, 0);
366 /* It is now okay to let Libgcrypt complain when there was/is
367 a problem with the secure memory. */
368 gcry_control (GCRYCTL_RESUME_SECMEM_WARN);
370 /* ... If required, other initialization goes here. */
372 /* Tell Libgcrypt that initialization has completed. */
373 gcry_control (GCRYCTL_INITIALIZATION_FINISHED, 0);
375 It is important that these initialization steps are not done by a
376 library but by the actual application. A library using Libgcrypt might
377 want to check for finished initialization using:
379 if (!gcry_control (GCRYCTL_INITIALIZATION_FINISHED_P))
381 fputs ("libgcrypt has not been initialized\n", stderr);
385 Instead of terminating the process, the library may instead print a
386 warning and try to initialize Libgcrypt itself. See also the section on
387 multi-threading below for more pitfalls.
390 File: gcrypt.info, Node: Multi-Threading, Next: Enabling FIPS mode, Prev: Initializing the library, Up: Preparation
395 As mentioned earlier, the Libgcrypt library is thread-safe if you adhere
396 to the following requirements:
398 * If you use pthread and your applications forks and does not
399 directly call exec (even calling stdio functions), all kind of
400 problems may occur. Future versions of Libgcrypt will try to
401 cleanup using pthread_atfork but even that may lead to problems.
402 This is a common problem with almost all applications using pthread
405 * The function 'gcry_check_version' must be called before any other
406 function in the library. To achieve this in multi-threaded
407 programs, you must synchronize the memory with respect to other
408 threads that also want to use Libgcrypt. For this, it is
409 sufficient to call 'gcry_check_version' before creating the other
410 threads using Libgcrypt(1).
412 * Just like the function 'gpg_strerror', the function 'gcry_strerror'
413 is not thread safe. You have to use 'gpg_strerror_r' instead.
415 ---------- Footnotes ----------
417 (1) At least this is true for POSIX threads, as 'pthread_create' is a
418 function that synchronizes memory with respects to other threads. There
419 are many functions which have this property, a complete list can be
420 found in POSIX, IEEE Std 1003.1-2003, Base Definitions, Issue 6, in the
421 definition of the term "Memory Synchronization". For other thread
422 packages, more relaxed or more strict rules may apply.
425 File: gcrypt.info, Node: Enabling FIPS mode, Next: Hardware features, Prev: Multi-Threading, Up: Preparation
427 2.6 How to enable the FIPS mode
428 ===============================
430 Libgcrypt may be used in a FIPS 140-2 mode. Note, that this does not
431 necessary mean that Libcgrypt is an appoved FIPS 140-2 module. Check
432 the NIST database at <http://csrc.nist.gov/groups/STM/cmvp/> to see what
433 versions of Libgcrypt are approved.
435 Because FIPS 140 has certain restrictions on the use of cryptography
436 which are not always wanted, Libgcrypt needs to be put into FIPS mode
437 explicitly. Three alternative mechanisms are provided to switch
438 Libgcrypt into this mode:
440 * If the file '/proc/sys/crypto/fips_enabled' exists and contains a
441 numeric value other than '0', Libgcrypt is put into FIPS mode at
442 initialization time. Obviously this works only on systems with a
443 'proc' file system (i.e. GNU/Linux).
445 * If the file '/etc/gcrypt/fips_enabled' exists, Libgcrypt is put
446 into FIPS mode at initialization time. Note that this filename is
447 hardwired and does not depend on any configuration options.
449 * If the application requests FIPS mode using the control command
450 'GCRYCTL_FORCE_FIPS_MODE'. This must be done prior to any
451 initialization (i.e. before 'gcry_check_version').
453 In addition to the standard FIPS mode, Libgcrypt may also be put into
454 an Enforced FIPS mode by writing a non-zero value into the file
455 '/etc/gcrypt/fips_enabled' or by using the control command
456 'GCRYCTL_SET_ENFORCED_FIPS_FLAG' before any other calls to libgcrypt.
457 The Enforced FIPS mode helps to detect applications which don't fulfill
458 all requirements for using Libgcrypt in FIPS mode (*note FIPS Mode::).
460 Once Libgcrypt has been put into FIPS mode, it is not possible to
461 switch back to standard mode without terminating the process first. If
462 the logging verbosity level of Libgcrypt has been set to at least 2, the
463 state transitions and the self-tests are logged.
466 File: gcrypt.info, Node: Hardware features, Prev: Enabling FIPS mode, Up: Preparation
468 2.7 How to disable hardware features
469 ====================================
471 Libgcrypt makes use of certain hardware features. If the use of a
472 feature is not desired it may be either be disabled by a program or
473 globally using a configuration file. The currently supported features
492 To disable a feature for all processes using Libgcrypt 1.6 or newer,
493 create the file '/etc/gcrypt/hwf.deny' and put each feature not to be
494 used on a single line. Empty lines, white space, and lines prefixed
495 with a hash mark are ignored. The file should be world readable.
497 To disable a feature specifically for a program that program must
498 tell it Libgcrypt before before calling 'gcry_check_version'.
501 gcry_control (GCRYCTL_DISABLE_HWF, "intel-rdrand", NULL);
503 To print the list of active features you may use this command:
505 mpicalc --print-config | grep ^hwflist: | tr : '\n' | tail -n +2
507 ---------- Footnotes ----------
509 (1) NB. Libgcrypt uses the RDRAND feature only as one source of
510 entropy. A CPU with a broken RDRAND will thus not compromise of the
511 random number generator
514 File: gcrypt.info, Node: Generalities, Next: Handler Functions, Prev: Preparation, Up: Top
521 * Controlling the library:: Controlling Libgcrypt's behavior.
522 * Error Handling:: Error codes and such.
525 File: gcrypt.info, Node: Controlling the library, Next: Error Handling, Up: Generalities
527 3.1 Controlling the library
528 ===========================
530 -- Function: gcry_error_t gcry_control (enum gcry_ctl_cmds CMD, ...)
532 This function can be used to influence the general behavior of
533 Libgcrypt in several ways. Depending on CMD, more arguments can or
536 'GCRYCTL_ENABLE_M_GUARD; Arguments: none'
537 This command enables the built-in memory guard. It must not
538 be used to activate the memory guard after the memory
539 management has already been used; therefore it can ONLY be
540 used before 'gcry_check_version'. Note that the memory guard
541 is NOT used when the user of the library has set his own
542 memory management callbacks.
544 'GCRYCTL_ENABLE_QUICK_RANDOM; Arguments: none'
545 This command inhibits the use the very secure random quality
546 level ('GCRY_VERY_STRONG_RANDOM') and degrades all request
547 down to 'GCRY_STRONG_RANDOM'. In general this is not
548 recommended. However, for some applications the extra quality
549 random Libgcrypt tries to create is not justified and this
550 option may help to get better performance. Please check with
551 a crypto expert whether this option can be used for your
554 This option can only be used at initialization time.
556 'GCRYCTL_DUMP_RANDOM_STATS; Arguments: none'
557 This command dumps random number generator related statistics
558 to the library's logging stream.
560 'GCRYCTL_DUMP_MEMORY_STATS; Arguments: none'
561 This command dumps memory management related statistics to the
562 library's logging stream.
564 'GCRYCTL_DUMP_SECMEM_STATS; Arguments: none'
565 This command dumps secure memory management related statistics
566 to the library's logging stream.
568 'GCRYCTL_DROP_PRIVS; Arguments: none'
569 This command disables the use of secure memory and drops the
570 privileges of the current process. This command has not much
571 use; the suggested way to disable secure memory is to use
572 'GCRYCTL_DISABLE_SECMEM' right after initialization.
574 'GCRYCTL_DISABLE_SECMEM; Arguments: none'
575 This command disables the use of secure memory. If this
576 command is used in FIPS mode, FIPS mode will be disabled and
577 the function 'gcry_fips_mode_active' returns false. However,
578 in Enforced FIPS mode this command has no effect at all.
580 Many applications do not require secure memory, so they should
581 disable it right away. This command should be executed right
582 after 'gcry_check_version'.
584 'GCRYCTL_DISABLE_LOCKED_SECMEM; Arguments: none'
585 This command disables the use of the mlock call for secure
586 memory. Disabling the use of mlock may for example be done if
587 an encrypted swap space is in use. This command should be
588 executed right after 'gcry_check_version'. Note that by using
589 functions like gcry_xmalloc_secure and gcry_mpi_snew Libgcrypt
590 may expand the secure memory pool with memory which lacks the
591 property of not being swapped out to disk (but will still be
594 'GCRYCTL_DISABLE_PRIV_DROP; Arguments: none'
595 This command sets a global flag to tell the secure memory
596 subsystem that it shall not drop privileges after secure
597 memory has been allocated. This command is commonly used
598 right after 'gcry_check_version' but may also be used right
599 away at program startup. It won't have an effect after the
600 secure memory pool has been initialized. WARNING: A process
601 running setuid(root) is a severe security risk. Processes
602 making use of Libgcrypt or other complex code should drop
603 these extra privileges as soon as possible. If this command
604 has been used the caller is responsible for dropping the
607 'GCRYCTL_INIT_SECMEM; Arguments: unsigned int nbytes'
608 This command is used to allocate a pool of secure memory and
609 thus enabling the use of secure memory. It also drops all
610 extra privileges the process has (i.e. if it is run as setuid
611 (root)). If the argument NBYTES is 0, secure memory will be
612 disabled. The minimum amount of secure memory allocated is
613 currently 16384 bytes; you may thus use a value of 1 to
614 request that default size.
616 'GCRYCTL_TERM_SECMEM; Arguments: none'
617 This command zeroises the secure memory and destroys the
618 handler. The secure memory pool may not be used anymore after
619 running this command. If the secure memory pool as already
620 been destroyed, this command has no effect. Applications
621 might want to run this command from their exit handler to make
622 sure that the secure memory gets properly destroyed. This
623 command is not necessarily thread-safe but that should not be
624 needed in cleanup code. It may be called from a signal
627 'GCRYCTL_DISABLE_SECMEM_WARN; Arguments: none'
628 Disable warning messages about problems with the secure memory
629 subsystem. This command should be run right after
630 'gcry_check_version'.
632 'GCRYCTL_SUSPEND_SECMEM_WARN; Arguments: none'
633 Postpone warning messages from the secure memory subsystem.
634 *Note the initialization example: sample-use-suspend-secmem,
637 'GCRYCTL_RESUME_SECMEM_WARN; Arguments: none'
638 Resume warning messages from the secure memory subsystem.
639 *Note the initialization example: sample-use-resume-secmem, on
642 'GCRYCTL_USE_SECURE_RNDPOOL; Arguments: none'
643 This command tells the PRNG to store random numbers in secure
644 memory. This command should be run right after
645 'gcry_check_version' and not later than the command
646 GCRYCTL_INIT_SECMEM. Note that in FIPS mode the secure memory
649 'GCRYCTL_SET_RANDOM_SEED_FILE; Arguments: const char *filename'
650 This command specifies the file, which is to be used as seed
651 file for the PRNG. If the seed file is registered prior to
652 initialization of the PRNG, the seed file's content (if it
653 exists and seems to be valid) is fed into the PRNG pool.
654 After the seed file has been registered, the PRNG can be
655 signalled to write out the PRNG pool's content into the seed
656 file with the following command.
658 'GCRYCTL_UPDATE_RANDOM_SEED_FILE; Arguments: none'
659 Write out the PRNG pool's content into the registered seed
662 Multiple instances of the applications sharing the same random
663 seed file can be started in parallel, in which case they will
664 read out the same pool and then race for updating it (the last
665 update overwrites earlier updates). They will differentiate
666 only by the weak entropy that is added in read_seed_file based
667 on the PID and clock, and up to 16 bytes of weak random
668 non-blockingly. The consequence is that the output of these
669 different instances is correlated to some extent. In a
670 perfect attack scenario, the attacker can control (or at least
671 guess) the PID and clock of the application, and drain the
672 system's entropy pool to reduce the "up to 16 bytes" above to
673 0. Then the dependencies of the initial states of the pools
674 are completely known. Note that this is not an issue if
675 random of 'GCRY_VERY_STRONG_RANDOM' quality is requested as in
676 this case enough extra entropy gets mixed. It is also not an
677 issue when using Linux (rndlinux driver), because this one
678 guarantees to read full 16 bytes from /dev/urandom and thus
679 there is no way for an attacker without kernel access to
680 control these 16 bytes.
682 'GCRYCTL_CLOSE_RANDOM_DEVICE; Arguments: none'
683 Try to close the random device. If on Unix system you call
684 fork(), the child process does no call exec(), and you do not
685 intend to use Libgcrypt in the child, it might be useful to
686 use this control code to close the inherited file descriptors
687 of the random device. If Libgcrypt is later used again by the
688 child, the device will be re-opened. On non-Unix systems this
689 control code is ignored.
691 'GCRYCTL_SET_VERBOSITY; Arguments: int level'
692 This command sets the verbosity of the logging. A level of 0
693 disables all extra logging whereas positive numbers enable
694 more verbose logging. The level may be changed at any time
695 but be aware that no memory synchronization is done so the
696 effect of this command might not immediately show up in other
697 threads. This command may even be used prior to
698 'gcry_check_version'.
700 'GCRYCTL_SET_DEBUG_FLAGS; Arguments: unsigned int flags'
701 Set the debug flag bits as given by the argument. Be aware
702 that that no memory synchronization is done so the effect of
703 this command might not immediately show up in other threads.
704 The debug flags are not considered part of the API and thus
705 may change without notice. As of now bit 0 enables debugging
706 of cipher functions and bit 1 debugging of
707 multi-precision-integers. This command may even be used prior
708 to 'gcry_check_version'.
710 'GCRYCTL_CLEAR_DEBUG_FLAGS; Arguments: unsigned int flags'
711 Set the debug flag bits as given by the argument. Be aware
712 that that no memory synchronization is done so the effect of
713 this command might not immediately show up in other threads.
714 This command may even be used prior to 'gcry_check_version'.
716 'GCRYCTL_DISABLE_INTERNAL_LOCKING; Arguments: none'
717 This command does nothing. It exists only for backward
720 'GCRYCTL_ANY_INITIALIZATION_P; Arguments: none'
721 This command returns true if the library has been basically
722 initialized. Such a basic initialization happens implicitly
723 with many commands to get certain internal subsystems running.
724 The common and suggested way to do this basic initialization
725 is by calling gcry_check_version.
727 'GCRYCTL_INITIALIZATION_FINISHED; Arguments: none'
728 This command tells the library that the application has
729 finished the initialization.
731 'GCRYCTL_INITIALIZATION_FINISHED_P; Arguments: none'
732 This command returns true if the command
733 GCRYCTL_INITIALIZATION_FINISHED has already been run.
735 'GCRYCTL_SET_THREAD_CBS; Arguments: struct ath_ops *ath_ops'
736 This command is obsolete since version 1.6.
738 'GCRYCTL_FAST_POLL; Arguments: none'
739 Run a fast random poll.
741 'GCRYCTL_SET_RNDEGD_SOCKET; Arguments: const char *filename'
742 This command may be used to override the default name of the
743 EGD socket to connect to. It may be used only during
744 initialization as it is not thread safe. Changing the socket
745 name again is not supported. The function may return an error
746 if the given filename is too long for a local socket name.
748 EGD is an alternative random gatherer, used only on systems
749 lacking a proper random device.
751 'GCRYCTL_PRINT_CONFIG; Arguments: FILE *stream'
752 This command dumps information pertaining to the configuration
753 of the library to the given stream. If NULL is given for
754 STREAM, the log system is used. This command may be used
755 before the initialization has been finished but not before a
756 'gcry_check_version'. Note that the macro 'estream_t' can be
757 used instead of 'gpgrt_stream_t'.
759 'GCRYCTL_OPERATIONAL_P; Arguments: none'
760 This command returns true if the library is in an operational
761 state. This information makes only sense in FIPS mode. In
762 contrast to other functions, this is a pure test function and
763 won't put the library into FIPS mode or change the internal
764 state. This command may be used before the initialization has
765 been finished but not before a 'gcry_check_version'.
767 'GCRYCTL_FIPS_MODE_P; Arguments: none'
768 This command returns true if the library is in FIPS mode.
769 Note, that this is no indication about the current state of
770 the library. This command may be used before the
771 initialization has been finished but not before a
772 'gcry_check_version'. An application may use this command or
773 the convenience macro below to check whether FIPS mode is
776 -- Function: int gcry_fips_mode_active (void)
778 Returns true if the FIPS mode is active. Note that this
779 is implemented as a macro.
781 'GCRYCTL_FORCE_FIPS_MODE; Arguments: none'
782 Running this command puts the library into FIPS mode. If the
783 library is already in FIPS mode, a self-test is triggered and
784 thus the library will be put into operational state. This
785 command may be used before a call to 'gcry_check_version' and
786 that is actually the recommended way to let an application
787 switch the library into FIPS mode. Note that Libgcrypt will
788 reject an attempt to switch to fips mode during or after the
791 'GCRYCTL_SET_ENFORCED_FIPS_FLAG; Arguments: none'
792 Running this command sets the internal flag that puts the
793 library into the enforced FIPS mode during the FIPS mode
794 initialization. This command does not affect the library if
795 the library is not put into the FIPS mode and it must be used
796 before any other libgcrypt library calls that initialize the
797 library such as 'gcry_check_version'. Note that Libgcrypt
798 will reject an attempt to switch to the enforced fips mode
799 during or after the initialization.
801 'GCRYCTL_SET_PREFERRED_RNG_TYPE; Arguments: int'
802 These are advisory commands to select a certain random number
803 generator. They are only advisory because libraries may not
804 know what an application actually wants or vice versa. Thus
805 Libgcrypt employs a priority check to select the actually used
806 RNG. If an applications selects a lower priority RNG but a
807 library requests a higher priority RNG Libgcrypt will switch
808 to the higher priority RNG. Applications and libraries should
809 use these control codes before 'gcry_check_version'. The
810 available generators are:
811 'GCRY_RNG_TYPE_STANDARD'
812 A conservative standard generator based on the
813 "Continuously Seeded Pseudo Random Number Generator"
814 designed by Peter Gutmann.
816 A deterministic random number generator conforming to he
817 document "NIST-Recommended Random Number Generator Based
818 on ANSI X9.31 Appendix A.2.4 Using the 3-Key Triple DES
819 and AES Algorithms" (2005-01-31). This implementation
820 uses the AES variant.
821 'GCRY_RNG_TYPE_SYSTEM'
822 A wrapper around the system's native RNG. On Unix system
823 these are usually the /dev/random and /dev/urandom
825 The default is 'GCRY_RNG_TYPE_STANDARD' unless FIPS mode as
826 been enabled; in which case 'GCRY_RNG_TYPE_FIPS' is used and
827 locked against further changes.
829 'GCRYCTL_GET_CURRENT_RNG_TYPE; Arguments: int *'
830 This command stores the type of the currently used RNG as an
831 integer value at the provided address.
833 'GCRYCTL_SELFTEST; Arguments: none'
834 This may be used at anytime to have the library run all
835 implemented self-tests. It works in standard and in FIPS
836 mode. Returns 0 on success or an error code on failure.
838 'GCRYCTL_DISABLE_HWF; Arguments: const char *name'
840 Libgcrypt detects certain features of the CPU at startup time.
841 For performance tests it is sometimes required not to use such
842 a feature. This option may be used to disable a certain
843 feature; i.e. Libgcrypt behaves as if this feature has not
844 been detected. This call can be used several times to disable
845 a set of features, or features may be given as a colon or
846 comma delimited string. The special feature "all" can be used
847 to disable all available features.
849 Note that the detection code might be run if the feature has
850 been disabled. This command must be used at initialization
851 time; i.e. before calling 'gcry_check_version'.
853 'GCRYCTL_REINIT_SYSCALL_CLAMP; Arguments: none'
855 Libgcrypt wraps blocking system calls with two functions calls
856 ("system call clamp") to give user land threading libraries a
857 hook for re-scheduling. This works by reading the system call
858 clamp from Libgpg-error at initialization time. However
859 sometimes Libgcrypt needs to be initialized before the user
860 land threading systems and at that point the system call clamp
861 has not been registered with Libgpg-error and in turn
862 Libgcrypt would not use them. The control code can be used to
863 tell Libgcrypt that a system call clamp has now been
864 registered with Libgpg-error and advised it to read the clamp
865 again. Obviously this control code may only be used before a
866 second thread is started in a process.
869 File: gcrypt.info, Node: Error Handling, Prev: Controlling the library, Up: Generalities
874 Many functions in Libgcrypt can return an error if they fail. For this
875 reason, the application should always catch the error condition and take
876 appropriate measures, for example by releasing the resources and passing
877 the error up to the caller, or by displaying a descriptive message to
878 the user and cancelling the operation.
880 Some error values do not indicate a system error or an error in the
881 operation, but the result of an operation that failed properly. For
882 example, if you try to decrypt a tempered message, the decryption will
883 fail. Another error value actually means that the end of a data buffer
884 or list has been reached. The following descriptions explain for many
885 error codes what they mean usually. Some error values have specific
886 meanings if returned by a certain functions. Such cases are described
887 in the documentation of those functions.
889 Libgcrypt uses the 'libgpg-error' library. This allows to share the
890 error codes with other components of the GnuPG system, and to pass error
891 values transparently from the crypto engine, or some helper application
892 of the crypto engine, to the user. This way no information is lost. As
893 a consequence, Libgcrypt does not use its own identifiers for error
894 codes, but uses those provided by 'libgpg-error'. They usually start
897 However, Libgcrypt does provide aliases for the functions defined in
898 libgpg-error, which might be preferred for name space consistency.
900 Most functions in Libgcrypt return an error code in the case of
901 failure. For this reason, the application should always catch the error
902 condition and take appropriate measures, for example by releasing the
903 resources and passing the error up to the caller, or by displaying a
904 descriptive message to the user and canceling the operation.
906 Some error values do not indicate a system error or an error in the
907 operation, but the result of an operation that failed properly.
909 GnuPG components, including Libgcrypt, use an extra library named
910 libgpg-error to provide a common error handling scheme. For more
911 information on libgpg-error, see the according manual.
915 * Error Values:: The error value and what it means.
916 * Error Sources:: A list of important error sources.
917 * Error Codes:: A list of important error codes.
918 * Error Strings:: How to get a descriptive string from a value.
921 File: gcrypt.info, Node: Error Values, Next: Error Sources, Up: Error Handling
926 -- Data type: gcry_err_code_t
927 The 'gcry_err_code_t' type is an alias for the 'libgpg-error' type
928 'gpg_err_code_t'. The error code indicates the type of an error,
929 or the reason why an operation failed.
931 A list of important error codes can be found in the next section.
933 -- Data type: gcry_err_source_t
934 The 'gcry_err_source_t' type is an alias for the 'libgpg-error'
935 type 'gpg_err_source_t'. The error source has not a precisely
936 defined meaning. Sometimes it is the place where the error
937 happened, sometimes it is the place where an error was encoded into
938 an error value. Usually the error source will give an indication
939 to where to look for the problem. This is not always true, but it
940 is attempted to achieve this goal.
942 A list of important error sources can be found in the next section.
944 -- Data type: gcry_error_t
945 The 'gcry_error_t' type is an alias for the 'libgpg-error' type
946 'gpg_error_t'. An error value like this has always two components,
947 an error code and an error source. Both together form the error
950 Thus, the error value can not be directly compared against an error
951 code, but the accessor functions described below must be used.
952 However, it is guaranteed that only 0 is used to indicate success
953 ('GPG_ERR_NO_ERROR'), and that in this case all other parts of the
954 error value are set to 0, too.
956 Note that in Libgcrypt, the error source is used purely for
957 diagnostic purposes. Only the error code should be checked to test
958 for a certain outcome of a function. The manual only documents the
959 error code part of an error value. The error source is left
960 unspecified and might be anything.
962 -- Function: gcry_err_code_t gcry_err_code (gcry_error_t ERR)
963 The static inline function 'gcry_err_code' returns the
964 'gcry_err_code_t' component of the error value ERR. This function
965 must be used to extract the error code from an error value in order
966 to compare it with the 'GPG_ERR_*' error code macros.
968 -- Function: gcry_err_source_t gcry_err_source (gcry_error_t ERR)
969 The static inline function 'gcry_err_source' returns the
970 'gcry_err_source_t' component of the error value ERR. This
971 function must be used to extract the error source from an error
972 value in order to compare it with the 'GPG_ERR_SOURCE_*' error
975 -- Function: gcry_error_t gcry_err_make (gcry_err_source_t SOURCE,
976 gcry_err_code_t CODE)
977 The static inline function 'gcry_err_make' returns the error value
978 consisting of the error source SOURCE and the error code CODE.
980 This function can be used in callback functions to construct an
981 error value to return it to the library.
983 -- Function: gcry_error_t gcry_error (gcry_err_code_t CODE)
984 The static inline function 'gcry_error' returns the error value
985 consisting of the default error source and the error code CODE.
987 For GCRY applications, the default error source is
988 'GPG_ERR_SOURCE_USER_1'. You can define 'GCRY_ERR_SOURCE_DEFAULT'
989 before including 'gcrypt.h' to change this default.
991 This function can be used in callback functions to construct an
992 error value to return it to the library.
994 The 'libgpg-error' library provides error codes for all system error
995 numbers it knows about. If ERR is an unknown error number, the error
996 code 'GPG_ERR_UNKNOWN_ERRNO' is used. The following functions can be
997 used to construct error values from system errno numbers.
999 -- Function: gcry_error_t gcry_err_make_from_errno
1000 (gcry_err_source_t SOURCE, int ERR)
1001 The function 'gcry_err_make_from_errno' is like 'gcry_err_make',
1002 but it takes a system error like 'errno' instead of a
1003 'gcry_err_code_t' error code.
1005 -- Function: gcry_error_t gcry_error_from_errno (int ERR)
1006 The function 'gcry_error_from_errno' is like 'gcry_error', but it
1007 takes a system error like 'errno' instead of a 'gcry_err_code_t'
1010 Sometimes you might want to map system error numbers to error codes
1011 directly, or map an error code representing a system error back to the
1012 system error number. The following functions can be used to do that.
1014 -- Function: gcry_err_code_t gcry_err_code_from_errno (int ERR)
1015 The function 'gcry_err_code_from_errno' returns the error code for
1016 the system error ERR. If ERR is not a known system error, the
1017 function returns 'GPG_ERR_UNKNOWN_ERRNO'.
1019 -- Function: int gcry_err_code_to_errno (gcry_err_code_t ERR)
1020 The function 'gcry_err_code_to_errno' returns the system error for
1021 the error code ERR. If ERR is not an error code representing a
1022 system error, or if this system error is not defined on this
1023 system, the function returns '0'.
1026 File: gcrypt.info, Node: Error Sources, Next: Error Codes, Prev: Error Values, Up: Error Handling
1031 The library 'libgpg-error' defines an error source for every component
1032 of the GnuPG system. The error source part of an error value is not
1033 well defined. As such it is mainly useful to improve the diagnostic
1034 error message for the user.
1036 If the error code part of an error value is '0', the whole error
1037 value will be '0'. In this case the error source part is of course
1038 'GPG_ERR_SOURCE_UNKNOWN'.
1040 The list of error sources that might occur in applications using
1043 'GPG_ERR_SOURCE_UNKNOWN'
1044 The error source is not known. The value of this error source is
1047 'GPG_ERR_SOURCE_GPGME'
1048 The error source is GPGME itself.
1050 'GPG_ERR_SOURCE_GPG'
1051 The error source is GnuPG, which is the crypto engine used for the
1054 'GPG_ERR_SOURCE_GPGSM'
1055 The error source is GPGSM, which is the crypto engine used for the
1058 'GPG_ERR_SOURCE_GCRYPT'
1059 The error source is 'libgcrypt', which is used by crypto engines to
1060 perform cryptographic operations.
1062 'GPG_ERR_SOURCE_GPGAGENT'
1063 The error source is 'gpg-agent', which is used by crypto engines to
1064 perform operations with the secret key.
1066 'GPG_ERR_SOURCE_PINENTRY'
1067 The error source is 'pinentry', which is used by 'gpg-agent' to
1068 query the passphrase to unlock a secret key.
1070 'GPG_ERR_SOURCE_SCD'
1071 The error source is the SmartCard Daemon, which is used by
1072 'gpg-agent' to delegate operations with the secret key to a
1075 'GPG_ERR_SOURCE_KEYBOX'
1076 The error source is 'libkbx', a library used by the crypto engines
1077 to manage local keyrings.
1079 'GPG_ERR_SOURCE_USER_1'
1080 'GPG_ERR_SOURCE_USER_2'
1081 'GPG_ERR_SOURCE_USER_3'
1082 'GPG_ERR_SOURCE_USER_4'
1083 These error sources are not used by any GnuPG component and can be
1084 used by other software. For example, applications using Libgcrypt
1085 can use them to mark error values coming from callback handlers.
1086 Thus 'GPG_ERR_SOURCE_USER_1' is the default for errors created with
1087 'gcry_error' and 'gcry_error_from_errno', unless you define
1088 'GCRY_ERR_SOURCE_DEFAULT' before including 'gcrypt.h'.
1091 File: gcrypt.info, Node: Error Codes, Next: Error Strings, Prev: Error Sources, Up: Error Handling
1096 The library 'libgpg-error' defines many error values. The following
1097 list includes the most important error codes.
1100 This value indicates the end of a list, buffer or file.
1103 This value indicates success. The value of this error code is '0'.
1104 Also, it is guaranteed that an error value made from the error code
1105 '0' will be '0' itself (as a whole). This means that the error
1106 source information is lost for this error code, however, as this
1107 error code indicates that no error occurred, this is generally not
1111 This value means that something went wrong, but either there is not
1112 enough information about the problem to return a more useful error
1113 value, or there is no separate error value for this type of
1117 This value means that an out-of-memory condition occurred.
1120 System errors are mapped to GPG_ERR_EFOO where FOO is the symbol
1121 for the system error.
1124 This value means that some user provided data was out of range.
1126 'GPG_ERR_UNUSABLE_PUBKEY'
1127 This value means that some recipients for a message were invalid.
1129 'GPG_ERR_UNUSABLE_SECKEY'
1130 This value means that some signers were invalid.
1133 This value means that data was expected where no data was found.
1136 This value means that a conflict of some sort occurred.
1138 'GPG_ERR_NOT_IMPLEMENTED'
1139 This value indicates that the specific function (or operation) is
1140 not implemented. This error should never happen. It can only
1141 occur if you use certain values or configuration options which do
1142 not work, but for which we think that they should work at some
1145 'GPG_ERR_DECRYPT_FAILED'
1146 This value indicates that a decryption operation was unsuccessful.
1148 'GPG_ERR_WRONG_KEY_USAGE'
1149 This value indicates that a key is not used appropriately.
1152 This value indicates that no secret key for the user ID is
1155 'GPG_ERR_UNSUPPORTED_ALGORITHM'
1156 This value means a verification failed because the cryptographic
1157 algorithm is not supported by the crypto backend.
1159 'GPG_ERR_BAD_SIGNATURE'
1160 This value means a verification failed because the signature is
1164 This value means a verification failed because the public key is
1167 'GPG_ERR_NOT_OPERATIONAL'
1168 This value means that the library is not yet in state which allows
1169 to use this function. This error code is in particular returned if
1170 Libgcrypt is operated in FIPS mode and the internal state of the
1171 library does not yet or not anymore allow the use of a service.
1173 This error code is only available with newer libgpg-error versions,
1174 thus you might see "invalid error code" when passing this to
1175 'gpg_strerror'. The numeric value of this error code is 176.
1181 These error codes are not used by any GnuPG component and can be
1182 freely used by other software. Applications using Libgcrypt might
1183 use them to mark specific errors returned by callback handlers if
1184 no suitable error codes (including the system errors) for these
1185 errors exist already.
1188 File: gcrypt.info, Node: Error Strings, Prev: Error Codes, Up: Error Handling
1193 -- Function: const char * gcry_strerror (gcry_error_t ERR)
1194 The function 'gcry_strerror' returns a pointer to a statically
1195 allocated string containing a description of the error code
1196 contained in the error value ERR. This string can be used to
1197 output a diagnostic message to the user.
1199 -- Function: const char * gcry_strsource (gcry_error_t ERR)
1200 The function 'gcry_strsource' returns a pointer to a statically
1201 allocated string containing a description of the error source
1202 contained in the error value ERR. This string can be used to
1203 output a diagnostic message to the user.
1205 The following example illustrates the use of the functions described
1209 gcry_cipher_hd_t handle;
1210 gcry_error_t err = 0;
1212 err = gcry_cipher_open (&handle, GCRY_CIPHER_AES,
1213 GCRY_CIPHER_MODE_CBC, 0);
1216 fprintf (stderr, "Failure: %s/%s\n",
1217 gcry_strsource (err),
1218 gcry_strerror (err));
1223 File: gcrypt.info, Node: Handler Functions, Next: Symmetric cryptography, Prev: Generalities, Up: Top
1228 Libgcrypt makes it possible to install so called 'handler functions',
1229 which get called by Libgcrypt in case of certain events.
1233 * Progress handler:: Using a progress handler function.
1234 * Allocation handler:: Using special memory allocation functions.
1235 * Error handler:: Using error handler functions.
1236 * Logging handler:: Using a special logging function.
1239 File: gcrypt.info, Node: Progress handler, Next: Allocation handler, Up: Handler Functions
1241 4.1 Progress handler
1242 ====================
1244 It is often useful to retrieve some feedback while long running
1245 operations are performed.
1247 -- Data type: gcry_handler_progress_t
1248 Progress handler functions have to be of the type
1249 'gcry_handler_progress_t', which is defined as:
1251 'void (*gcry_handler_progress_t) (void *, const char *, int, int,
1254 The following function may be used to register a handler function for
1257 -- Function: void gcry_set_progress_handler (gcry_handler_progress_t
1260 This function installs CB as the 'Progress handler' function. It
1261 may be used only during initialization. CB must be defined as
1265 my_progress_handler (void *CB_DATA, const char *WHAT,
1266 int PRINTCHAR, int CURRENT, int TOTAL)
1271 A description of the arguments of the progress handler function
1275 The argument provided in the call to
1276 'gcry_set_progress_handler'.
1278 A string identifying the type of the progress output. The
1279 following values for WHAT are defined:
1282 Not enough entropy is available. TOTAL holds the number
1286 Waiting to re-open a random device. TOTAL gives the
1287 number of seconds until the next try.
1290 Values for PRINTCHAR:
1294 Need to refresh the pool of prime numbers.
1296 Number of bits adjusted.
1298 Searching for a generator.
1300 Fermat test on 10 candidates failed.
1302 Restart with a new random value.
1304 Rabin Miller test passed.
1307 File: gcrypt.info, Node: Allocation handler, Next: Error handler, Prev: Progress handler, Up: Handler Functions
1309 4.2 Allocation handler
1310 ======================
1312 It is possible to make Libgcrypt use special memory allocation functions
1313 instead of the built-in ones.
1315 Memory allocation functions are of the following types:
1316 -- Data type: gcry_handler_alloc_t
1317 This type is defined as: 'void *(*gcry_handler_alloc_t) (size_t
1319 -- Data type: gcry_handler_secure_check_t
1320 This type is defined as: 'int *(*gcry_handler_secure_check_t)
1322 -- Data type: gcry_handler_realloc_t
1323 This type is defined as: 'void *(*gcry_handler_realloc_t) (void *p,
1325 -- Data type: gcry_handler_free_t
1326 This type is defined as: 'void *(*gcry_handler_free_t) (void *)'.
1328 Special memory allocation functions can be installed with the
1331 -- Function: void gcry_set_allocation_handler (gcry_handler_alloc_t
1332 FUNC_ALLOC, gcry_handler_alloc_t FUNC_ALLOC_SECURE,
1333 gcry_handler_secure_check_t FUNC_SECURE_CHECK,
1334 gcry_handler_realloc_t FUNC_REALLOC, gcry_handler_free_t
1336 Install the provided functions and use them instead of the built-in
1337 functions for doing memory allocation. Using this function is in
1338 general not recommended because the standard Libgcrypt allocation
1339 functions are guaranteed to zeroize memory if needed.
1341 This function may be used only during initialization and may not be
1345 File: gcrypt.info, Node: Error handler, Next: Logging handler, Prev: Allocation handler, Up: Handler Functions
1350 The following functions may be used to register handler functions that
1351 are called by Libgcrypt in case certain error conditions occur. They
1352 may and should be registered prior to calling 'gcry_check_version'.
1354 -- Data type: gcry_handler_no_mem_t
1355 This type is defined as: 'int (*gcry_handler_no_mem_t) (void *,
1356 size_t, unsigned int)'
1357 -- Function: void gcry_set_outofcore_handler (gcry_handler_no_mem_t
1358 FUNC_NO_MEM, void *CB_DATA)
1359 This function registers FUNC_NO_MEM as 'out-of-core handler', which
1360 means that it will be called in the case of not having enough
1361 memory available. The handler is called with 3 arguments: The
1362 first one is the pointer CB_DATA as set with this function, the
1363 second is the requested memory size and the last being a flag. If
1364 bit 0 of the flag is set, secure memory has been requested. The
1365 handler should either return true to indicate that Libgcrypt should
1366 try again allocating memory or return false to let Libgcrypt use
1367 its default fatal error handler.
1369 -- Data type: gcry_handler_error_t
1370 This type is defined as: 'void (*gcry_handler_error_t) (void *,
1373 -- Function: void gcry_set_fatalerror_handler (gcry_handler_error_t
1374 FUNC_ERROR, void *CB_DATA)
1375 This function registers FUNC_ERROR as 'error handler', which means
1376 that it will be called in error conditions.
1379 File: gcrypt.info, Node: Logging handler, Prev: Error handler, Up: Handler Functions
1384 -- Data type: gcry_handler_log_t
1385 This type is defined as: 'void (*gcry_handler_log_t) (void *, int,
1386 const char *, va_list)'
1388 -- Function: void gcry_set_log_handler (gcry_handler_log_t FUNC_LOG,
1390 This function registers FUNC_LOG as 'logging handler', which means
1391 that it will be called in case Libgcrypt wants to log a message.
1392 This function may and should be used prior to calling
1393 'gcry_check_version'.
1396 File: gcrypt.info, Node: Symmetric cryptography, Next: Public Key cryptography, Prev: Handler Functions, Up: Top
1398 5 Symmetric cryptography
1399 ************************
1401 The cipher functions are used for symmetrical cryptography, i.e.
1402 cryptography using a shared key. The programming model follows an
1403 open/process/close paradigm and is in that similar to other building
1404 blocks provided by Libgcrypt.
1408 * Available ciphers:: List of ciphers supported by the library.
1409 * Available cipher modes:: List of cipher modes supported by the library.
1410 * Working with cipher handles:: How to perform operations related to cipher handles.
1411 * General cipher functions:: General cipher functions independent of cipher handles.
1414 File: gcrypt.info, Node: Available ciphers, Next: Available cipher modes, Up: Symmetric cryptography
1416 5.1 Available ciphers
1417 =====================
1420 This is not a real algorithm but used by some functions as error
1421 return. The value always evaluates to false.
1424 This is the IDEA algorithm.
1427 Triple-DES with 3 Keys as EDE. The key size of this algorithm is
1428 168 but you have to pass 192 bits because the most significant bits
1429 of each byte are ignored.
1432 CAST128-5 block cipher algorithm. The key size is 128 bits.
1434 'GCRY_CIPHER_BLOWFISH'
1435 The blowfish algorithm. The current implementation allows only for
1436 a key size of 128 bits.
1438 'GCRY_CIPHER_SAFER_SK128'
1439 Reserved and not currently implemented.
1441 'GCRY_CIPHER_DES_SK'
1442 Reserved and not currently implemented.
1445 'GCRY_CIPHER_AES128'
1446 'GCRY_CIPHER_RIJNDAEL'
1447 'GCRY_CIPHER_RIJNDAEL128'
1448 AES (Rijndael) with a 128 bit key.
1450 'GCRY_CIPHER_AES192'
1451 'GCRY_CIPHER_RIJNDAEL192'
1452 AES (Rijndael) with a 192 bit key.
1454 'GCRY_CIPHER_AES256'
1455 'GCRY_CIPHER_RIJNDAEL256'
1456 AES (Rijndael) with a 256 bit key.
1458 'GCRY_CIPHER_TWOFISH'
1459 The Twofish algorithm with a 256 bit key.
1461 'GCRY_CIPHER_TWOFISH128'
1462 The Twofish algorithm with a 128 bit key.
1464 'GCRY_CIPHER_ARCFOUR'
1465 An algorithm which is 100% compatible with RSA Inc.'s RC4
1466 algorithm. Note that this is a stream cipher and must be used very
1467 carefully to avoid a couple of weaknesses.
1470 Standard DES with a 56 bit key. You need to pass 64 bit but the
1471 high bits of each byte are ignored. Note, that this is a weak
1472 algorithm which can be broken in reasonable time using a brute
1475 'GCRY_CIPHER_SERPENT128'
1476 'GCRY_CIPHER_SERPENT192'
1477 'GCRY_CIPHER_SERPENT256'
1478 The Serpent cipher from the AES contest.
1480 'GCRY_CIPHER_RFC2268_40'
1481 'GCRY_CIPHER_RFC2268_128'
1482 Ron's Cipher 2 in the 40 and 128 bit variants.
1485 A 128 bit cipher as described by RFC4269.
1487 'GCRY_CIPHER_CAMELLIA128'
1488 'GCRY_CIPHER_CAMELLIA192'
1489 'GCRY_CIPHER_CAMELLIA256'
1490 The Camellia cipher by NTT. See
1491 <http://info.isl.ntt.co.jp/crypt/eng/camellia/specifications.html>.
1493 'GCRY_CIPHER_SALSA20'
1494 This is the Salsa20 stream cipher.
1496 'GCRY_CIPHER_SALSA20R12'
1497 This is the Salsa20/12 - reduced round version of Salsa20 stream
1500 'GCRY_CIPHER_GOST28147'
1501 The GOST 28147-89 cipher, defined in the respective GOST standard.
1502 Translation of this GOST into English is provided in the RFC-5830.
1504 'GCRY_CIPHER_CHACHA20'
1505 This is the ChaCha20 stream cipher.
1508 File: gcrypt.info, Node: Available cipher modes, Next: Working with cipher handles, Prev: Available ciphers, Up: Symmetric cryptography
1510 5.2 Available cipher modes
1511 ==========================
1513 'GCRY_CIPHER_MODE_NONE'
1514 No mode specified. This should not be used. The only exception is
1515 that if Libgcrypt is not used in FIPS mode and if any debug flag
1516 has been set, this mode may be used to bypass the actual
1519 'GCRY_CIPHER_MODE_ECB'
1520 Electronic Codebook mode.
1522 'GCRY_CIPHER_MODE_CFB'
1523 'GCRY_CIPHER_MODE_CFB8'
1524 Cipher Feedback mode. For GCRY_CIPHER_MODE_CFB the shift size
1525 equals the block size of the cipher (e.g. for AES it is CFB-128).
1526 For GCRY_CIPHER_MODE_CFB8 the shift size is 8 bit but that variant
1527 is not yet available.
1529 'GCRY_CIPHER_MODE_CBC'
1530 Cipher Block Chaining mode.
1532 'GCRY_CIPHER_MODE_STREAM'
1533 Stream mode, only to be used with stream cipher algorithms.
1535 'GCRY_CIPHER_MODE_OFB'
1536 Output Feedback mode.
1538 'GCRY_CIPHER_MODE_CTR'
1541 'GCRY_CIPHER_MODE_AESWRAP'
1542 This mode is used to implement the AES-Wrap algorithm according to
1543 RFC-3394. It may be used with any 128 bit block length algorithm,
1544 however the specs require one of the 3 AES algorithms. These
1545 special conditions apply: If 'gcry_cipher_setiv' has not been used
1546 the standard IV is used; if it has been used the lower 64 bit of
1547 the IV are used as the Alternative Initial Value. On encryption
1548 the provided output buffer must be 64 bit (8 byte) larger than the
1549 input buffer; in-place encryption is still allowed. On decryption
1550 the output buffer may be specified 64 bit (8 byte) shorter than
1551 then input buffer. As per specs the input length must be at least
1552 128 bits and the length must be a multiple of 64 bits.
1554 'GCRY_CIPHER_MODE_CCM'
1555 Counter with CBC-MAC mode is an Authenticated Encryption with
1556 Associated Data (AEAD) block cipher mode, which is specified in
1557 'NIST Special Publication 800-38C' and RFC 3610.
1559 'GCRY_CIPHER_MODE_GCM'
1560 Galois/Counter Mode (GCM) is an Authenticated Encryption with
1561 Associated Data (AEAD) block cipher mode, which is specified in
1562 'NIST Special Publication 800-38D'.
1564 'GCRY_CIPHER_MODE_POLY1305'
1565 This mode implements the Poly1305 Authenticated Encryption with
1566 Associated Data (AEAD) mode according to RFC-7539. This mode can
1567 be used with ChaCha20 stream cipher.
1569 'GCRY_CIPHER_MODE_OCB'
1570 OCB is an Authenticated Encryption with Associated Data (AEAD)
1571 block cipher mode, which is specified in RFC-7253. Supported tag
1572 lengths are 128, 96, and 64 bit with the default being 128 bit. To
1573 switch to a different tag length 'gcry_cipher_ctl' using the
1574 command 'GCRYCTL_SET_TAGLEN' and the address of an 'int' variable
1575 set to 12 (for 96 bit) or 8 (for 64 bit) provided for the 'buffer'
1576 argument and 'sizeof(int)' for 'buflen'.
1578 Note that the use of 'gcry_cipher_final' is required.
1580 'GCRY_CIPHER_MODE_XTS'
1581 XEX-based tweaked-codebook mode with ciphertext stealing (XTS) mode
1582 is used to implement the AES-XTS as specified in IEEE 1619 Standard
1583 Architecture for Encrypted Shared Storage Media and NIST SP800-38E.
1585 The XTS mode requires doubling key-length, for example, using
1586 512-bit key with AES-256 ('GCRY_CIPHER_AES256'). The 128-bit tweak
1587 value is feed to XTS mode as little-endian byte array using
1588 'gcry_cipher_setiv' function. When encrypting or decrypting,
1589 full-sized data unit buffers needs to be passed to
1590 'gcry_cipher_encrypt' or 'gcry_cipher_decrypt'. The tweak value is
1591 automatically incremented after each call of 'gcry_cipher_encrypt'
1592 and 'gcry_cipher_decrypt'. Auto-increment allows avoiding need of
1593 setting IV between processing of sequential data units.
1596 File: gcrypt.info, Node: Working with cipher handles, Next: General cipher functions, Prev: Available cipher modes, Up: Symmetric cryptography
1598 5.3 Working with cipher handles
1599 ===============================
1601 To use a cipher algorithm, you must first allocate an according handle.
1602 This is to be done using the open function:
1604 -- Function: gcry_error_t gcry_cipher_open (gcry_cipher_hd_t *HD, int
1605 ALGO, int MODE, unsigned int FLAGS)
1607 This function creates the context handle required for most of the
1608 other cipher functions and returns a handle to it in 'hd'. In case
1609 of an error, an according error code is returned.
1611 The ID of algorithm to use must be specified via ALGO. See *note
1612 Available ciphers::, for a list of supported ciphers and the
1613 according constants.
1615 Besides using the constants directly, the function
1616 'gcry_cipher_map_name' may be used to convert the textual name of
1617 an algorithm into the according numeric ID.
1619 The cipher mode to use must be specified via MODE. See *note
1620 Available cipher modes::, for a list of supported cipher modes and
1621 the according constants. Note that some modes are incompatible
1622 with some algorithms - in particular, stream mode
1623 ('GCRY_CIPHER_MODE_STREAM') only works with stream ciphers.
1624 Poly1305 AEAD mode ('GCRY_CIPHER_MODE_POLY1305') only works with
1625 ChaCha20 stream cipher. The block cipher modes
1626 ('GCRY_CIPHER_MODE_ECB', 'GCRY_CIPHER_MODE_CBC',
1627 'GCRY_CIPHER_MODE_CFB', 'GCRY_CIPHER_MODE_OFB' and
1628 'GCRY_CIPHER_MODE_CTR') will work with any block cipher algorithm.
1629 GCM mode ('GCRY_CIPHER_MODE_CCM'), CCM mode
1630 ('GCRY_CIPHER_MODE_GCM'), OCB mode ('GCRY_CIPHER_MODE_OCB'), and
1631 XTS mode ('GCRY_CIPHER_MODE_XTS') will only work with block cipher
1632 algorithms which have the block size of 16 bytes.
1634 The third argument FLAGS can either be passed as '0' or as the
1635 bit-wise OR of the following constants.
1637 'GCRY_CIPHER_SECURE'
1638 Make sure that all operations are allocated in secure memory.
1639 This is useful when the key material is highly confidential.
1640 'GCRY_CIPHER_ENABLE_SYNC'
1641 This flag enables the CFB sync mode, which is a special
1642 feature of Libgcrypt's CFB mode implementation to allow for
1643 OpenPGP's CFB variant. See 'gcry_cipher_sync'.
1644 'GCRY_CIPHER_CBC_CTS'
1645 Enable cipher text stealing (CTS) for the CBC mode. Cannot be
1646 used simultaneous as GCRY_CIPHER_CBC_MAC. CTS mode makes it
1647 possible to transform data of almost arbitrary size (only
1648 limitation is that it must be greater than the algorithm's
1650 'GCRY_CIPHER_CBC_MAC'
1651 Compute CBC-MAC keyed checksums. This is the same as CBC
1652 mode, but only output the last block. Cannot be used
1653 simultaneous as GCRY_CIPHER_CBC_CTS.
1655 Use the following function to release an existing handle:
1657 -- Function: void gcry_cipher_close (gcry_cipher_hd_t H)
1659 This function releases the context created by 'gcry_cipher_open'.
1660 It also zeroises all sensitive information associated with this
1663 In order to use a handle for performing cryptographic operations, a
1664 'key' has to be set first:
1666 -- Function: gcry_error_t gcry_cipher_setkey (gcry_cipher_hd_t H, const
1669 Set the key K used for encryption or decryption in the context
1670 denoted by the handle H. The length L (in bytes) of the key K must
1671 match the required length of the algorithm set for this context or
1672 be in the allowed range for algorithms with variable key size. The
1673 function checks this and returns an error if there is a problem. A
1674 caller should always check for an error.
1676 Most crypto modes requires an initialization vector (IV), which
1677 usually is a non-secret random string acting as a kind of salt value.
1678 The CTR mode requires a counter, which is also similar to a salt value.
1679 To set the IV or CTR, use these functions:
1681 -- Function: gcry_error_t gcry_cipher_setiv (gcry_cipher_hd_t H, const
1684 Set the initialization vector used for encryption or decryption.
1685 The vector is passed as the buffer K of length L bytes and copied
1686 to internal data structures. The function checks that the IV
1687 matches the requirement of the selected algorithm and mode.
1689 This function is also used by AEAD modes and with Salsa20 and
1690 ChaCha20 stream ciphers to set or update the required nonce. In
1691 these cases it needs to be called after setting the key.
1693 -- Function: gcry_error_t gcry_cipher_setctr (gcry_cipher_hd_t H, const
1696 Set the counter vector used for encryption or decryption. The
1697 counter is passed as the buffer C of length L bytes and copied to
1698 internal data structures. The function checks that the counter
1699 matches the requirement of the selected algorithm (i.e., it must be
1700 the same size as the block size).
1702 -- Function: gcry_error_t gcry_cipher_reset (gcry_cipher_hd_t H)
1704 Set the given handle's context back to the state it had after the
1705 last call to gcry_cipher_setkey and clear the initialization
1708 Note that gcry_cipher_reset is implemented as a macro.
1710 Authenticated Encryption with Associated Data (AEAD) block cipher
1711 modes require the handling of the authentication tag and the additional
1712 authenticated data, which can be done by using the following functions:
1714 -- Function: gcry_error_t gcry_cipher_authenticate (gcry_cipher_hd_t H,
1715 const void *ABUF, size_t ABUFLEN)
1717 Process the buffer ABUF of length ABUFLEN as the additional
1718 authenticated data (AAD) for AEAD cipher modes.
1720 -- Function: gcry_error_t gcry_cipher_gettag (gcry_cipher_hd_t H,
1721 void *TAG, size_t TAGLEN)
1723 This function is used to read the authentication tag after
1724 encryption. The function finalizes and outputs the authentication
1725 tag to the buffer TAG of length TAGLEN bytes.
1727 Depending on the used mode certain restrictions for TAGLEN are
1728 enforced: For GCM TAGLEN must be at least 16 or one of the allowed
1729 truncated lengths (4, 8, 12, 13, 14, or 15).
1731 -- Function: gcry_error_t gcry_cipher_checktag (gcry_cipher_hd_t H,
1732 const void *TAG, size_t TAGLEN)
1734 Check the authentication tag after decryption. The authentication
1735 tag is passed as the buffer TAG of length TAGLEN bytes and compared
1736 to internal authentication tag computed during decryption. Error
1737 code 'GPG_ERR_CHECKSUM' is returned if the authentication tag in
1738 the buffer TAG does not match the authentication tag calculated
1741 Depending on the used mode certain restrictions for TAGLEN are
1742 enforced: For GCM TAGLEN must either be 16 or one of the allowed
1743 truncated lengths (4, 8, 12, 13, 14, or 15).
1745 The actual encryption and decryption is done by using one of the
1746 following functions. They may be used as often as required to process
1749 -- Function: gcry_error_t gcry_cipher_encrypt (gcry_cipher_hd_t H,
1750 unsigned char *out, size_t OUTSIZE, const unsigned char *IN,
1753 'gcry_cipher_encrypt' is used to encrypt the data. This function
1754 can either work in place or with two buffers. It uses the cipher
1755 context already setup and described by the handle H. There are 2
1756 ways to use the function: If IN is passed as 'NULL' and INLEN is
1757 '0', in-place encryption of the data in OUT of length OUTSIZE takes
1758 place. With IN being not 'NULL', INLEN bytes are encrypted to the
1759 buffer OUT which must have at least a size of INLEN. OUTSIZE must
1760 be set to the allocated size of OUT, so that the function can check
1761 that there is sufficient space. Note that overlapping buffers are
1764 Depending on the selected algorithms and encryption mode, the
1765 length of the buffers must be a multiple of the block size.
1767 Some encryption modes require that 'gcry_cipher_final' is used
1768 before the final data chunk is passed to this function.
1770 The function returns '0' on success or an error code.
1772 -- Function: gcry_error_t gcry_cipher_decrypt (gcry_cipher_hd_t H,
1773 unsigned char *out, size_t OUTSIZE, const unsigned char *IN,
1776 'gcry_cipher_decrypt' is used to decrypt the data. This function
1777 can either work in place or with two buffers. It uses the cipher
1778 context already setup and described by the handle H. There are 2
1779 ways to use the function: If IN is passed as 'NULL' and INLEN is
1780 '0', in-place decryption of the data in OUT or length OUTSIZE takes
1781 place. With IN being not 'NULL', INLEN bytes are decrypted to the
1782 buffer OUT which must have at least a size of INLEN. OUTSIZE must
1783 be set to the allocated size of OUT, so that the function can check
1784 that there is sufficient space. Note that overlapping buffers are
1787 Depending on the selected algorithms and encryption mode, the
1788 length of the buffers must be a multiple of the block size.
1790 Some encryption modes require that 'gcry_cipher_final' is used
1791 before the final data chunk is passed to this function.
1793 The function returns '0' on success or an error code.
1795 The OCB mode features integrated padding and must thus be told about
1796 the end of the input data. This is done with:
1798 -- Function: gcry_error_t gcry_cipher_final (gcry_cipher_hd_t H)
1800 Set a flag in the context to tell the encrypt and decrypt functions
1801 that their next call will provide the last chunk of data. Only the
1802 first call to this function has an effect and only for modes which
1803 support it. Checking the error is in general not necessary. This
1804 is implemented as a macro.
1806 OpenPGP (as defined in RFC-4880) requires a special sync operation in
1807 some places. The following function is used for this:
1809 -- Function: gcry_error_t gcry_cipher_sync (gcry_cipher_hd_t H)
1811 Perform the OpenPGP sync operation on context H. Note that this is
1812 a no-op unless the context was created with the flag
1813 'GCRY_CIPHER_ENABLE_SYNC'
1815 Some of the described functions are implemented as macros utilizing a
1816 catch-all control function. This control function is rarely used
1817 directly but there is nothing which would inhibit it:
1819 -- Function: gcry_error_t gcry_cipher_ctl (gcry_cipher_hd_t H, int CMD,
1820 void *BUFFER, size_t BUFLEN)
1822 'gcry_cipher_ctl' controls various aspects of the cipher module and
1823 specific cipher contexts. Usually some more specialized functions
1824 or macros are used for this purpose. The semantics of the function
1825 and its parameters depends on the the command CMD and the passed
1826 context handle H. Please see the comments in the source code
1827 ('src/global.c') for details.
1829 -- Function: gcry_error_t gcry_cipher_info (gcry_cipher_hd_t H, int
1830 WHAT, void *BUFFER, size_t *NBYTES)
1832 'gcry_cipher_info' is used to retrieve various information about a
1833 cipher context or the cipher module in general.
1835 'GCRYCTL_GET_TAGLEN:'
1836 Return the length of the tag for an AE algorithm mode. An
1837 error is returned for modes which do not support a tag.
1838 BUFFER must be given as NULL. On success the result is stored
1839 NBYTES. The taglen is returned in bytes.
1842 File: gcrypt.info, Node: General cipher functions, Prev: Working with cipher handles, Up: Symmetric cryptography
1844 5.4 General cipher functions
1845 ============================
1847 To work with the algorithms, several functions are available to map
1848 algorithm names to the internal identifiers, as well as ways to retrieve
1849 information about an algorithm or the current cipher context.
1851 -- Function: gcry_error_t gcry_cipher_algo_info (int ALGO, int WHAT,
1852 void *BUFFER, size_t *NBYTES)
1854 This function is used to retrieve information on a specific
1855 algorithm. You pass the cipher algorithm ID as ALGO and the type
1856 of information requested as WHAT. The result is either returned as
1857 the return code of the function or copied to the provided BUFFER
1858 whose allocated length must be available in an integer variable
1859 with the address passed in NBYTES. This variable will also receive
1860 the actual used length of the buffer.
1862 Here is a list of supported codes for WHAT:
1864 'GCRYCTL_GET_KEYLEN:'
1865 Return the length of the key. If the algorithm supports
1866 multiple key lengths, the maximum supported value is returned.
1867 The length is returned as number of octets (bytes) and not as
1868 number of bits in NBYTES; BUFFER must be zero. Note that it
1869 is usually better to use the convenience function
1870 'gcry_cipher_get_algo_keylen'.
1872 'GCRYCTL_GET_BLKLEN:'
1873 Return the block length of the algorithm. The length is
1874 returned as a number of octets in NBYTES; BUFFER must be zero.
1875 Note that it is usually better to use the convenience function
1876 'gcry_cipher_get_algo_blklen'.
1878 'GCRYCTL_TEST_ALGO:'
1879 Returns '0' when the specified algorithm is available for use.
1880 BUFFER and NBYTES must be zero.
1882 -- Function: size_t gcry_cipher_get_algo_keylen (ALGO)
1884 This function returns length of the key for algorithm ALGO. If the
1885 algorithm supports multiple key lengths, the maximum supported key
1886 length is returned. On error '0' is returned. The key length is
1887 returned as number of octets.
1889 This is a convenience functions which should be preferred over
1890 'gcry_cipher_algo_info' because it allows for proper type checking.
1892 -- Function: size_t gcry_cipher_get_algo_blklen (int ALGO)
1894 This functions returns the block-length of the algorithm ALGO
1895 counted in octets. On error '0' is returned.
1897 This is a convenience functions which should be preferred over
1898 'gcry_cipher_algo_info' because it allows for proper type checking.
1900 -- Function: const char * gcry_cipher_algo_name (int ALGO)
1902 'gcry_cipher_algo_name' returns a string with the name of the
1903 cipher algorithm ALGO. If the algorithm is not known or another
1904 error occurred, the string '"?"' is returned. This function should
1905 not be used to test for the availability of an algorithm.
1907 -- Function: int gcry_cipher_map_name (const char *NAME)
1909 'gcry_cipher_map_name' returns the algorithm identifier for the
1910 cipher algorithm described by the string NAME. If this algorithm
1911 is not available '0' is returned.
1913 -- Function: int gcry_cipher_mode_from_oid (const char *STRING)
1915 Return the cipher mode associated with an ASN.1 object identifier.
1916 The object identifier is expected to be in the IETF-style dotted
1917 decimal notation. The function returns '0' for an unknown object
1918 identifier or when no mode is associated with it.
1921 File: gcrypt.info, Node: Public Key cryptography, Next: Hashing, Prev: Symmetric cryptography, Up: Top
1923 6 Public Key cryptography
1924 *************************
1926 Public key cryptography, also known as asymmetric cryptography, is an
1927 easy way for key management and to provide digital signatures.
1928 Libgcrypt provides two completely different interfaces to public key
1929 cryptography, this chapter explains the one based on S-expressions.
1933 * Available algorithms:: Algorithms supported by the library.
1934 * Used S-expressions:: Introduction into the used S-expression.
1935 * Cryptographic Functions:: Functions for performing the cryptographic actions.
1936 * General public-key related Functions:: General functions, not implementing any cryptography.
1939 File: gcrypt.info, Node: Available algorithms, Next: Used S-expressions, Up: Public Key cryptography
1941 6.1 Available algorithms
1942 ========================
1944 Libgcrypt supports the RSA (Rivest-Shamir-Adleman) algorithms as well as
1945 DSA (Digital Signature Algorithm) and Elgamal. The versatile interface
1946 allows to add more algorithms in the future.
1949 File: gcrypt.info, Node: Used S-expressions, Next: Cryptographic Functions, Prev: Available algorithms, Up: Public Key cryptography
1951 6.2 Used S-expressions
1952 ======================
1954 Libgcrypt's API for asymmetric cryptography is based on data structures
1955 called S-expressions (see
1956 <http://people.csail.mit.edu/rivest/sexp.html>) and does not work with
1957 contexts as most of the other building blocks of Libgcrypt do.
1959 The following information are stored in S-expressions:
1969 To describe how Libgcrypt expect keys, we use examples. Note that words
1970 in uppercase indicate parameters whereas lowercase words are literals.
1972 Note that all MPI (multi-precision-integers) values are expected to
1973 be in 'GCRYMPI_FMT_USG' format. An easy way to create S-expressions is
1974 by using 'gcry_sexp_build' which allows to pass a string with
1975 printf-like escapes to insert MPI values.
1979 * RSA key parameters:: Parameters used with an RSA key.
1980 * DSA key parameters:: Parameters used with a DSA key.
1981 * ECC key parameters:: Parameters used with ECC keys.
1984 File: gcrypt.info, Node: RSA key parameters, Next: DSA key parameters, Up: Used S-expressions
1986 6.2.1 RSA key parameters
1987 ------------------------
1989 An RSA private key is described by this S-expression:
2000 An RSA public key is described by this S-expression:
2008 RSA public modulus n.
2010 RSA public exponent e.
2012 RSA secret exponent d = e^{-1} \bmod (p-1)(q-1).
2016 RSA secret prime q with p < q.
2018 Multiplicative inverse u = p^{-1} \bmod q.
2020 For signing and decryption the parameters (p, q, u) are optional but
2021 greatly improve the performance. Either all of these optional
2022 parameters must be given or none of them. They are mandatory for
2025 Note that OpenSSL uses slighly different parameters: q < p and u =
2026 q^{-1} \bmod p. To use these parameters you will need to swap the
2027 values and recompute u. Here is example code to do this:
2029 if (gcry_mpi_cmp (p, q) > 0)
2031 gcry_mpi_swap (p, q);
2032 gcry_mpi_invm (u, p, q);
2036 File: gcrypt.info, Node: DSA key parameters, Next: ECC key parameters, Prev: RSA key parameters, Up: Used S-expressions
2038 6.2.2 DSA key parameters
2039 ------------------------
2041 A DSA private key is described by this S-expression:
2054 DSA group order q (which is a prime divisor of p-1).
2056 DSA group generator g.
2058 DSA public key value y = g^x \bmod p.
2060 DSA secret exponent x.
2062 The public key is similar with "private-key" replaced by "public-key"
2066 File: gcrypt.info, Node: ECC key parameters, Prev: DSA key parameters, Up: Used S-expressions
2068 6.2.3 ECC key parameters
2069 ------------------------
2071 An ECC private key is described by this S-expression:
2084 Prime specifying the field GF(p).
2087 The two coefficients of the Weierstrass equation y^2 = x^3 + ax + b
2093 The point representing the public key Q = dG.
2097 All point values are encoded in standard format; Libgcrypt does in
2098 general only support uncompressed points, thus the first byte needs to
2099 be '0x04'. However "EdDSA" describes its own compression scheme which
2100 is used by default; the non-standard first byte '0x40' may optionally be
2101 used to explicit flag the use of the algorithm’s native compression
2104 The public key is similar with "private-key" replaced by "public-key"
2107 If the domain parameters are well-known, the name of this curve may
2108 be used. For example
2112 (curve "NIST P-192")
2116 Note that Q-POINT is optional for a private key. The 'curve'
2117 parameter may be given in any case and is used to replace missing
2120 Currently implemented curves are:
2122 '1.2.840.10045.3.1.1'
2125 The NIST 192 bit curve, its OID, X9.62 and SECP aliases.
2129 The NIST 224 bit curve and its SECP alias.
2132 '1.2.840.10045.3.1.7'
2135 The NIST 256 bit curve, its OID, X9.62 and SECP aliases.
2139 The NIST 384 bit curve and its SECP alias.
2143 The NIST 521 bit curve and its SECP alias.
2145 As usual the OIDs may optionally be prefixed with the string 'OID.'
2149 File: gcrypt.info, Node: Cryptographic Functions, Next: General public-key related Functions, Prev: Used S-expressions, Up: Public Key cryptography
2151 6.3 Cryptographic Functions
2152 ===========================
2154 Some functions operating on S-expressions support 'flags' to influence
2155 the operation. These flags have to be listed in a sub-S-expression
2156 named 'flags'. Flag names are case-sensitive. The following flags are
2161 If supported by the algorithm and curve the 'comp' flag requests
2162 that points are returned in compact (compressed) representation.
2163 The 'nocomp' flag requests that points are returned with full
2164 coordinates. The default depends on the the algorithm and curve.
2165 The compact representation requires a small overhead before a point
2166 can be used but halves the size of a to be conveyed public key. If
2167 'comp' is used with the "EdDSA" algorithm the key generation prefix
2168 the public key with a '0x40' byte.
2171 Use PKCS#1 block type 2 padding for encryption, block type 1
2172 padding for signing.
2175 Use RSA-OAEP padding for encryption.
2178 Use RSA-PSS padding for signing.
2181 Use the EdDSA scheme signing instead of the default ECDSA
2182 algorithm. Note that the EdDSA uses a special form of the public
2186 For DSA and ECDSA use a deterministic scheme for the k parameter.
2189 Do not use a technique called 'blinding', which is used by default
2190 in order to prevent leaking of secret information. Blinding is
2191 only implemented by RSA, but it might be implemented by other
2192 algorithms in the future as well, when necessary.
2195 For ECC key generation also return the domain parameters. For ECC
2196 signing and verification override default parameters by provided
2197 domain parameters of the public or private key.
2200 This flag is only meaningful for RSA, DSA, and ECC key generation.
2201 If given the key is created using a faster and a somewhat less
2202 secure random number generator. This flag may be used for keys
2203 which are only used for a short time or per-message and do not
2204 require full cryptographic strength.
2207 This flag skips internal failsafe tests to assert that a generated
2208 key is properly working. It currently has an effect only for
2209 standard ECC key generation. It is mostly useful along with
2210 transient-key to achieve fastest ECC key generation.
2213 Force the use of the ANSI X9.31 key generation algorithm instead of
2214 the default algorithm. This flag is only meaningful for RSA key
2215 generation and usually not required. Note that this algorithm is
2216 implicitly used if either 'derive-parms' is given or Libgcrypt is
2220 Force the use of the FIPS 186 key generation algorithm instead of
2221 the default algorithm. This flag is only meaningful for DSA and
2222 usually not required. Note that this algorithm is implicitly used
2223 if either 'derive-parms' is given or Libgcrypt is in FIPS mode. As
2224 of now FIPS 186-2 is implemented; after the approval of FIPS 186-3
2225 the code will be changed to implement 186-3.
2228 Force the use of the FIPS 186-2 key generation algorithm instead of
2229 the default algorithm. This algorithm is slightly different from
2230 FIPS 186-3 and allows only 1024 bit keys. This flag is only
2231 meaningful for DSA and only required for FIPS testing backward
2234 Now that we know the key basics, we can carry on and explain how to
2235 encrypt and decrypt data. In almost all cases the data is a random
2236 session key which is in turn used for the actual encryption of the real
2237 data. There are 2 functions to do this:
2239 -- Function: gcry_error_t gcry_pk_encrypt (gcry_sexp_t *R_CIPH,
2240 gcry_sexp_t DATA, gcry_sexp_t PKEY)
2242 Obviously a public key must be provided for encryption. It is
2243 expected as an appropriate S-expression (see above) in PKEY. The
2244 data to be encrypted can either be in the simple old format, which
2245 is a very simple S-expression consisting only of one MPI, or it may
2246 be a more complex S-expression which also allows to specify flags
2247 for operation, like e.g. padding rules.
2249 If you don't want to let Libgcrypt handle the padding, you must
2250 pass an appropriate MPI using this expression for DATA:
2256 This has the same semantics as the old style MPI only way. MPI is
2257 the actual data, already padded appropriate for your protocol.
2258 Most RSA based systems however use PKCS#1 padding and so you can
2259 use this S-expression for DATA:
2265 Here, the "flags" list has the "pkcs1" flag which let the function
2266 know that it should provide PKCS#1 block type 2 padding. The
2267 actual data to be encrypted is passed as a string of octets in
2268 BLOCK. The function checks that this data actually can be used
2269 with the given key, does the padding and encrypts it.
2271 If the function could successfully perform the encryption, the
2272 return value will be 0 and a new S-expression with the encrypted
2273 result is allocated and assigned to the variable at the address of
2274 R_CIPH. The caller is responsible to release this value using
2275 'gcry_sexp_release'. In case of an error, an error code is
2276 returned and R_CIPH will be set to 'NULL'.
2278 The returned S-expression has this format when used with RSA:
2284 Where A-MPI is an MPI with the result of the RSA operation. When
2285 using the Elgamal algorithm, the return value will have this
2293 Where A-MPI and B-MPI are MPIs with the result of the Elgamal
2294 encryption operation.
2296 -- Function: gcry_error_t gcry_pk_decrypt (gcry_sexp_t *R_PLAIN,
2297 gcry_sexp_t DATA, gcry_sexp_t SKEY)
2299 Obviously a private key must be provided for decryption. It is
2300 expected as an appropriate S-expression (see above) in SKEY. The
2301 data to be decrypted must match the format of the result as
2302 returned by 'gcry_pk_encrypt', but should be enlarged with a
2311 This function does not remove padding from the data by default. To
2312 let Libgcrypt remove padding, give a hint in 'flags' telling which
2313 padding method was used when encrypting:
2315 (flags PADDING-METHOD)
2317 Currently PADDING-METHOD is either 'pkcs1' for PKCS#1 block type 2
2318 padding, or 'oaep' for RSA-OAEP padding.
2320 The function returns 0 on success or an error code. The variable
2321 at the address of R_PLAIN will be set to NULL on error or receive
2322 the decrypted value on success. The format of R_PLAIN is a simple
2323 S-expression part (i.e. not a valid one) with just one MPI if
2324 there was no 'flags' element in DATA; if at least an empty 'flags'
2325 is passed in DATA, the format is:
2329 Another operation commonly performed using public key cryptography is
2330 signing data. In some sense this is even more important than encryption
2331 because digital signatures are an important instrument for key
2332 management. Libgcrypt supports digital signatures using 2 functions,
2333 similar to the encryption functions:
2335 -- Function: gcry_error_t gcry_pk_sign (gcry_sexp_t *R_SIG,
2336 gcry_sexp_t DATA, gcry_sexp_t SKEY)
2338 This function creates a digital signature for DATA using the
2339 private key SKEY and place it into the variable at the address of
2340 R_SIG. DATA may either be the simple old style S-expression with
2341 just one MPI or a modern and more versatile S-expression which
2342 allows to let Libgcrypt handle padding:
2346 (hash HASH-ALGO BLOCK))
2348 This example requests to sign the data in BLOCK after applying
2349 PKCS#1 block type 1 style padding. HASH-ALGO is a string with the
2350 hash algorithm to be encoded into the signature, this may be any
2351 hash algorithm name as supported by Libgcrypt. Most likely, this
2352 will be "sha256" or "sha1". It is obvious that the length of BLOCK
2353 must match the size of that message digests; the function checks
2354 that this and other constraints are valid.
2356 If PKCS#1 padding is not required (because the caller does already
2357 provide a padded value), either the old format or better the
2358 following format should be used:
2364 Here, the data to be signed is directly given as an MPI.
2366 For DSA the input data is expected in this format:
2372 Here, the data to be signed is directly given as an MPI. It is
2373 expect that this MPI is the the hash value. For the standard DSA
2374 using a MPI is not a problem in regard to leading zeroes because
2375 the hash value is directly used as an MPI. For better standard
2376 conformance it would be better to explicit use a memory string
2377 (like with pkcs1) but that is currently not supported. However,
2378 for deterministic DSA as specified in RFC6979 this can't be used.
2379 Instead the following input is expected.
2383 (hash HASH-ALGO BLOCK))
2385 Note that the provided hash-algo is used for the internal HMAC; it
2386 should match the hash-algo used to create BLOCK.
2388 The signature is returned as a newly allocated S-expression in
2389 R_SIG using this format for RSA:
2395 Where S-MPI is the result of the RSA sign operation. For DSA the
2396 S-expression returned is:
2403 Where R-MPI and S-MPI are the result of the DSA sign operation.
2405 For Elgamal signing (which is slow, yields large numbers, hard to
2406 use correctly and probably is not as secure as the other
2407 algorithms), the same format is used with "elg" replacing "dsa";
2408 for ECDSA signing, the same format is used with "ecdsa" replacing
2411 For the EdDSA algorithm (cf. Ed25515) the required input
2419 Note that the MESSAGE may be of any length; hashing is part of the
2420 algorithm. Using a large data block for MESSAGE is not suggested;
2421 in that case the used protocol should better require that a hash of
2422 the message is used as input to the EdDSA algorithm.
2424 The operation most commonly used is definitely the verification of a
2425 signature. Libgcrypt provides this function:
2427 -- Function: gcry_error_t gcry_pk_verify (gcry_sexp_t SIG,
2428 gcry_sexp_t DATA, gcry_sexp_t PKEY)
2430 This is used to check whether the signature SIG matches the DATA.
2431 The public key PKEY must be provided to perform this verification.
2432 This function is similar in its parameters to 'gcry_pk_sign' with
2433 the exceptions that the public key is used instead of the private
2434 key and that no signature is created but a signature, in a format
2435 as created by 'gcry_pk_sign', is passed to the function in SIG.
2437 The result is 0 for success (i.e. the data matches the signature),
2438 or an error code where the most relevant code is
2439 'GCRY_ERR_BAD_SIGNATURE' to indicate that the signature does not
2440 match the provided data.
2443 File: gcrypt.info, Node: General public-key related Functions, Prev: Cryptographic Functions, Up: Public Key cryptography
2445 6.4 General public-key related Functions
2446 ========================================
2448 A couple of utility functions are available to retrieve the length of
2449 the key, map algorithm identifiers and perform sanity checks:
2451 -- Function: const char * gcry_pk_algo_name (int ALGO)
2453 Map the public key algorithm id ALGO to a string representation of
2454 the algorithm name. For unknown algorithms this functions returns
2455 the string '"?"'. This function should not be used to test for the
2456 availability of an algorithm.
2458 -- Function: int gcry_pk_map_name (const char *NAME)
2460 Map the algorithm NAME to a public key algorithm Id. Returns 0 if
2461 the algorithm name is not known.
2463 -- Function: int gcry_pk_test_algo (int ALGO)
2465 Return 0 if the public key algorithm ALGO is available for use.
2466 Note that this is implemented as a macro.
2468 -- Function: unsigned int gcry_pk_get_nbits (gcry_sexp_t KEY)
2470 Return what is commonly referred as the key length for the given
2471 public or private in KEY.
2473 -- Function: unsigned char * gcry_pk_get_keygrip (gcry_sexp_t KEY,
2474 unsigned char *ARRAY)
2476 Return the so called "keygrip" which is the SHA-1 hash of the
2477 public key parameters expressed in a way depended on the algorithm.
2478 ARRAY must either provide space for 20 bytes or be 'NULL'. In the
2479 latter case a newly allocated array of that size is returned. On
2480 success a pointer to the newly allocated space or to ARRAY is
2481 returned. 'NULL' is returned to indicate an error which is most
2482 likely an unknown algorithm or one where a "keygrip" has not yet
2483 been defined. The function accepts public or secret keys in KEY.
2485 -- Function: gcry_error_t gcry_pk_testkey (gcry_sexp_t KEY)
2487 Return zero if the private key KEY is 'sane', an error code
2488 otherwise. Note that it is not possible to check the 'saneness' of
2491 -- Function: gcry_error_t gcry_pk_algo_info (int ALGO, int WHAT,
2492 void *BUFFER, size_t *NBYTES)
2494 Depending on the value of WHAT return various information about the
2495 public key algorithm with the id ALGO. Note that the function
2496 returns '-1' on error and the actual error code must be retrieved
2497 using the function 'gcry_errno'. The currently defined values for
2500 'GCRYCTL_TEST_ALGO:'
2501 Return 0 if the specified algorithm is available for use.
2502 BUFFER must be 'NULL', NBYTES may be passed as 'NULL' or point
2503 to a variable with the required usage of the algorithm. This
2504 may be 0 for "don't care" or the bit-wise OR of these flags:
2506 'GCRY_PK_USAGE_SIGN'
2507 Algorithm is usable for signing.
2508 'GCRY_PK_USAGE_ENCR'
2509 Algorithm is usable for encryption.
2511 Unless you need to test for the allowed usage, it is in
2512 general better to use the macro gcry_pk_test_algo instead.
2514 'GCRYCTL_GET_ALGO_USAGE:'
2515 Return the usage flags for the given algorithm. An invalid
2516 algorithm return 0. Disabled algorithms are ignored here
2517 because we want to know whether the algorithm is at all
2518 capable of a certain usage.
2520 'GCRYCTL_GET_ALGO_NPKEY'
2521 Return the number of elements the public key for algorithm
2522 ALGO consist of. Return 0 for an unknown algorithm.
2524 'GCRYCTL_GET_ALGO_NSKEY'
2525 Return the number of elements the private key for algorithm
2526 ALGO consist of. Note that this value is always larger than
2527 that of the public key. Return 0 for an unknown algorithm.
2529 'GCRYCTL_GET_ALGO_NSIGN'
2530 Return the number of elements a signature created with the
2531 algorithm ALGO consists of. Return 0 for an unknown algorithm
2532 or for an algorithm not capable of creating signatures.
2534 'GCRYCTL_GET_ALGO_NENC'
2535 Return the number of elements a encrypted message created with
2536 the algorithm ALGO consists of. Return 0 for an unknown
2537 algorithm or for an algorithm not capable of encryption.
2539 Please note that parameters not required should be passed as
2542 -- Function: gcry_error_t gcry_pk_ctl (int CMD, void *BUFFER,
2545 This is a general purpose function to perform certain control
2546 operations. CMD controls what is to be done. The return value is
2547 0 for success or an error code. Currently supported values for CMD
2550 'GCRYCTL_DISABLE_ALGO'
2551 Disable the algorithm given as an algorithm id in BUFFER.
2552 BUFFER must point to an 'int' variable with the algorithm id
2553 and BUFLEN must have the value 'sizeof (int)'. This function
2554 is not thread safe and should thus be used before any other
2555 threads are started.
2557 Libgcrypt also provides a function to generate public key pairs:
2559 -- Function: gcry_error_t gcry_pk_genkey (gcry_sexp_t *R_KEY,
2562 This function create a new public key pair using information given
2563 in the S-expression PARMS and stores the private and the public key
2564 in one new S-expression at the address given by R_KEY. In case of
2565 an error, R_KEY is set to 'NULL'. The return code is 0 for success
2566 or an error code otherwise.
2568 Here is an example for PARMS to create an 2048 bit RSA key:
2574 To create an Elgamal key, substitute "elg" for "rsa" and to create
2575 a DSA key use "dsa". Valid ranges for the key length depend on the
2576 algorithms; all commonly used key lengths are supported. Currently
2577 supported parameters are:
2580 This is always required to specify the length of the key. The
2581 argument is a string with a number in C-notation. The value
2582 should be a multiple of 8. Note that the S-expression syntax
2583 requires that a number is prefixed with its string length;
2584 thus the '4:' in the above example.
2587 For ECC a named curve may be used instead of giving the number
2588 of requested bits. This allows to request a specific curve to
2589 override a default selection Libgcrypt would have taken if
2590 'nbits' has been given. The available names are listed with
2591 the description of the ECC public key parameters.
2594 This is only used with RSA to give a hint for the public
2595 exponent. The VALUE will be used as a base to test for a
2596 usable exponent. Some values are special:
2599 Use a secure and fast value. This is currently the
2602 Use a value as required by some crypto policies. This is
2603 currently the number 65537.
2607 Use the given value.
2609 If this parameter is not used, Libgcrypt uses for historic
2610 reasons 65537. Note that the value must fit into a 32 bit
2611 unsigned variable and that the usual C prefixes are considered
2612 (e.g. 017 gives 15).
2615 This is only meanigful for DSA keys. If it is given the DSA
2616 key is generated with a Q parameyer of size N bits. If it is
2617 not given or zero Q is deduced from NBITS in this way:
2628 Note that in this case only the values for N, as given in the
2629 table, are allowed. When specifying Q all values of N in the
2630 range 512 to 15680 are valid as long as they are multiples of
2634 This is only meaningful for DLP algorithms. If specified keys
2635 are generated with domain parameters taken from this list.
2636 The exact format of this parameter depends on the actual
2637 algorithm. It is currently only implemented for DSA using
2647 'nbits' and 'qbits' may not be specified because they are
2648 derived from the domain parameters.
2651 This is currently only implemented for RSA and DSA keys. It
2652 is not allowed to use this together with a 'domain'
2653 specification. If given, it is used to derive the keys using
2654 the given parameters.
2656 If given for an RSA key the X9.31 key generation algorithm is
2657 used even if libgcrypt is not in FIPS mode. If given for a
2658 DSA key, the FIPS 186 algorithm is used even if libgcrypt is
2666 (Xp1 #1A1916DDB29B4EB7EB6732E128#)
2667 (Xp2 #192E8AAC41C576C822D93EA433#)
2668 (Xp #D8CD81F035EC57EFE822955149D3BFF70C53520D
2669 769D6D76646C7A792E16EBD89FE6FC5B605A6493
2670 39DFC925A86A4C6D150B71B9EEA02D68885F5009
2672 (Xq1 #1A5CF72EE770DE50CB09ACCEA9#)
2673 (Xq2 #134E4CAA16D2350A21D775C404#)
2674 (Xq #CC1092495D867E64065DEE3E7955F2EBC7D47A2D
2675 7C9953388F97DDDC3E1CA19C35CA659EDC2FC325
2676 6D29C2627479C086A699A49C4C9CEE7EF7BD1B34
2686 This is preferred way to define flags. FLAGLIST may contain
2687 any number of flags. See above for a specification of these
2690 Here is an example on how to create a key using curve Ed25519
2691 with the ECDSA signature algorithm. Note that the use of
2692 ECDSA with that curve is in general not recommended.
2695 (flags transient-key)))
2701 These are deprecated ways to set a flag with that name; see
2702 above for a description of each flag.
2704 The key pair is returned in a format depending on the algorithm.
2705 Both private and public keys are returned in one container and may
2706 be accompanied by some miscellaneous information.
2708 Here are two examples; the first for Elgamal and the second for
2709 elliptic curve key generation:
2724 (pm1-factors N1 N2 ... NN))
2739 As you can see, some of the information is duplicated, but this
2740 provides an easy way to extract either the public or the private
2741 key. Note that the order of the elements is not defined, e.g. the
2742 private key may be stored before the public key. N1 N2 ... NN is a
2743 list of prime numbers used to composite P-MPI; this is in general
2744 not a very useful information and only available if the key
2745 generation algorithm provides them.
2747 Future versions of Libgcrypt will have extended versions of the public
2748 key interfaced which will take an additional context to allow for
2749 pre-computations, special operations, and other optimization. As a
2750 first step a new function is introduced to help using the ECC algorithms
2753 -- Function: gcry_error_t gcry_pubkey_get_sexp (gcry_sexp_t *R_SEXP,
2754 int MODE, gcry_ctx_t CTX)
2756 Return an S-expression representing the context CTX. Depending on
2757 the state of that context, the S-expression may either be a public
2758 key, a private key or any other object used with public key
2759 operations. On success 0 is returned and a new S-expression is
2760 stored at R_SEXP; on error an error code is returned and NULL is
2761 stored at R_SEXP. MODE must be one of:
2764 Decide what to return depending on the context. For example
2765 if the private key parameter is available a private key is
2766 returned, if not a public key is returned.
2768 'GCRY_PK_GET_PUBKEY'
2769 Return the public key even if the context has the private key
2772 'GCRY_PK_GET_SECKEY'
2773 Return the private key or the error 'GPG_ERR_NO_SECKEY' if it
2776 As of now this function supports only certain ECC operations
2777 because a context object is right now only defined for ECC. Over
2778 time this function will be extended to cover more algorithms.
2781 File: gcrypt.info, Node: Hashing, Next: Message Authentication Codes, Prev: Public Key cryptography, Up: Top
2786 Libgcrypt provides an easy and consistent to use interface for hashing.
2787 Hashing is buffered and several hash algorithms can be updated at once.
2788 It is possible to compute a HMAC using the same routines. The
2789 programming model follows an open/process/close paradigm and is in that
2790 similar to other building blocks provided by Libgcrypt.
2792 For convenience reasons, a few cyclic redundancy check value
2793 operations are also supported.
2797 * Available hash algorithms:: List of hash algorithms supported by the library.
2798 * Working with hash algorithms:: List of functions related to hashing.
2801 File: gcrypt.info, Node: Available hash algorithms, Next: Working with hash algorithms, Up: Hashing
2803 7.1 Available hash algorithms
2804 =============================
2807 This is not a real algorithm but used by some functions as an error
2808 return value. This constant is guaranteed to have the value '0'.
2811 This is the SHA-1 algorithm which yields a message digest of 20
2812 bytes. Note that SHA-1 begins to show some weaknesses and it is
2813 suggested to fade out its use if strong cryptographic properties
2817 This is the 160 bit version of the RIPE message digest
2818 (RIPE-MD-160). Like SHA-1 it also yields a digest of 20 bytes.
2819 This algorithm share a lot of design properties with SHA-1 and thus
2820 it is advisable not to use it for new protocols.
2823 This is the well known MD5 algorithm, which yields a message digest
2824 of 16 bytes. Note that the MD5 algorithm has severe weaknesses,
2825 for example it is easy to compute two messages yielding the same
2826 hash (collision attack). The use of this algorithm is only
2827 justified for non-cryptographic application.
2830 This is the MD4 algorithm, which yields a message digest of 16
2831 bytes. This algorithm has severe weaknesses and should not be
2835 This is an reserved identifier for MD-2; there is no implementation
2836 yet. This algorithm has severe weaknesses and should not be used.
2839 This is the TIGER/192 algorithm which yields a message digest of 24
2840 bytes. Actually this is a variant of TIGER with a different output
2841 print order as used by GnuPG up to version 1.3.2.
2844 This is the TIGER variant as used by the NESSIE project. It uses
2845 the most commonly used output print order.
2848 This is another variant of TIGER with a different padding scheme.
2851 This is an reserved value for the HAVAL algorithm with 5 passes and
2852 160 bit. It yields a message digest of 20 bytes. Note that there
2853 is no implementation yet available.
2856 This is the SHA-224 algorithm which yields a message digest of 28
2857 bytes. See Change Notice 1 for FIPS 180-2 for the specification.
2860 This is the SHA-256 algorithm which yields a message digest of 32
2861 bytes. See FIPS 180-2 for the specification.
2864 This is the SHA-384 algorithm which yields a message digest of 48
2865 bytes. See FIPS 180-2 for the specification.
2868 This is the SHA-384 algorithm which yields a message digest of 64
2869 bytes. See FIPS 180-2 for the specification.
2872 This is the SHA3-224 algorithm which yields a message digest of 28
2873 bytes. See FIPS 202 for the specification.
2876 This is the SHA3-256 algorithm which yields a message digest of 32
2877 bytes. See FIPS 202 for the specification.
2880 This is the SHA3-384 algorithm which yields a message digest of 48
2881 bytes. See FIPS 202 for the specification.
2884 This is the SHA3-384 algorithm which yields a message digest of 64
2885 bytes. See FIPS 202 for the specification.
2888 This is the SHAKE128 extendable-output function (XOF) algorithm
2889 with 128 bit security strength. See FIPS 202 for the
2893 This is the SHAKE256 extendable-output function (XOF) algorithm
2894 with 256 bit security strength. See FIPS 202 for the
2898 This is the ISO 3309 and ITU-T V.42 cyclic redundancy check. It
2899 yields an output of 4 bytes. Note that this is not a hash
2900 algorithm in the cryptographic sense.
2902 'GCRY_MD_CRC32_RFC1510'
2903 This is the above cyclic redundancy check function, as modified by
2904 RFC 1510. It yields an output of 4 bytes. Note that this is not a
2905 hash algorithm in the cryptographic sense.
2907 'GCRY_MD_CRC24_RFC2440'
2908 This is the OpenPGP cyclic redundancy check function. It yields an
2909 output of 3 bytes. Note that this is not a hash algorithm in the
2910 cryptographic sense.
2913 This is the Whirlpool algorithm which yields a message digest of 64
2916 'GCRY_MD_GOSTR3411_94'
2917 This is the hash algorithm described in GOST R 34.11-94 which
2918 yields a message digest of 32 bytes.
2920 'GCRY_MD_STRIBOG256'
2921 This is the 256-bit version of hash algorithm described in GOST R
2922 34.11-2012 which yields a message digest of 32 bytes.
2924 'GCRY_MD_STRIBOG512'
2925 This is the 512-bit version of hash algorithm described in GOST R
2926 34.11-2012 which yields a message digest of 64 bytes.
2928 'GCRY_MD_BLAKE2B_512'
2929 This is the BLAKE2b-512 algorithm which yields a message digest of
2930 64 bytes. See RFC 7693 for the specification.
2932 'GCRY_MD_BLAKE2B_384'
2933 This is the BLAKE2b-384 algorithm which yields a message digest of
2934 48 bytes. See RFC 7693 for the specification.
2936 'GCRY_MD_BLAKE2B_256'
2937 This is the BLAKE2b-256 algorithm which yields a message digest of
2938 32 bytes. See RFC 7693 for the specification.
2940 'GCRY_MD_BLAKE2B_160'
2941 This is the BLAKE2b-160 algorithm which yields a message digest of
2942 20 bytes. See RFC 7693 for the specification.
2944 'GCRY_MD_BLAKE2S_256'
2945 This is the BLAKE2s-256 algorithm which yields a message digest of
2946 32 bytes. See RFC 7693 for the specification.
2948 'GCRY_MD_BLAKE2S_224'
2949 This is the BLAKE2s-224 algorithm which yields a message digest of
2950 28 bytes. See RFC 7693 for the specification.
2952 'GCRY_MD_BLAKE2S_160'
2953 This is the BLAKE2s-160 algorithm which yields a message digest of
2954 20 bytes. See RFC 7693 for the specification.
2956 'GCRY_MD_BLAKE2S_128'
2957 This is the BLAKE2s-128 algorithm which yields a message digest of
2958 16 bytes. See RFC 7693 for the specification.
2961 File: gcrypt.info, Node: Working with hash algorithms, Prev: Available hash algorithms, Up: Hashing
2963 7.2 Working with hash algorithms
2964 ================================
2966 To use most of these function it is necessary to create a context; this
2969 -- Function: gcry_error_t gcry_md_open (gcry_md_hd_t *HD, int ALGO,
2972 Create a message digest object for algorithm ALGO. FLAGS may be
2973 given as an bitwise OR of constants described below. ALGO may be
2974 given as '0' if the algorithms to use are later set using
2975 'gcry_md_enable'. HD is guaranteed to either receive a valid
2978 For a list of supported algorithms, see *note Available hash
2981 The flags allowed for MODE are:
2983 'GCRY_MD_FLAG_SECURE'
2984 Allocate all buffers and the resulting digest in "secure
2985 memory". Use this is the hashed data is highly confidential.
2988 Turn the algorithm into a HMAC message authentication
2989 algorithm. This only works if just one algorithm is enabled
2990 for the handle and that algorithm is not an extendable-output
2991 function. Note that the function 'gcry_md_setkey' must be
2992 used to set the MAC key. The size of the MAC is equal to the
2993 message digest of the underlying hash algorithm. If you want
2994 CBC message authentication codes based on a cipher, see *note
2995 Working with cipher handles::.
2997 'GCRY_MD_FLAG_BUGEMU1'
2998 Versions of Libgcrypt before 1.6.0 had a bug in the Whirlpool
2999 code which led to a wrong result for certain input sizes and
3000 write patterns. Using this flag emulates that bug. This may
3001 for example be useful for applications which use Whirlpool as
3002 part of their key generation. It is strongly suggested to use
3003 this flag only if really needed and if possible to the data
3004 should be re-processed using the regular Whirlpool algorithm.
3006 Note that this flag works for the entire hash context. If
3007 needed arises it may be used to enable bug emulation for other
3008 hash algorithms. Thus you should not use this flag for a
3009 multi-algorithm hash context.
3011 You may use the function 'gcry_md_is_enabled' to later check
3012 whether an algorithm has been enabled.
3014 If you want to calculate several hash algorithms at the same time,
3015 you have to use the following function right after the 'gcry_md_open':
3017 -- Function: gcry_error_t gcry_md_enable (gcry_md_hd_t H, int ALGO)
3019 Add the message digest algorithm ALGO to the digest object
3020 described by handle H. Duplicated enabling of algorithms is
3021 detected and ignored.
3023 If the flag 'GCRY_MD_FLAG_HMAC' was used, the key for the MAC must be
3024 set using the function:
3026 -- Function: gcry_error_t gcry_md_setkey (gcry_md_hd_t H, const void
3027 *KEY, size_t KEYLEN)
3029 For use with the HMAC feature or BLAKE2 keyed hash, set the MAC key
3030 to the value of KEY of length KEYLEN bytes. For HMAC, there is no
3031 restriction on the length of the key. For keyed BLAKE2b hash,
3032 length of the key must be 64 bytes or less. For keyed BLAKE2s
3033 hash, length of the key must be 32 bytes or less.
3035 After you are done with the hash calculation, you should release the
3038 -- Function: void gcry_md_close (gcry_md_hd_t H)
3040 Release all resources of hash context H. H should not be used
3041 after a call to this function. A 'NULL' passed as H is ignored.
3042 The function also zeroises all sensitive information associated
3045 Often you have to do several hash operations using the same
3046 algorithm. To avoid the overhead of creating and releasing context, a
3047 reset function is provided:
3049 -- Function: void gcry_md_reset (gcry_md_hd_t H)
3051 Reset the current context to its initial state. This is
3052 effectively identical to a close followed by an open and enabling
3053 all currently active algorithms.
3055 Often it is necessary to start hashing some data and then continue to
3056 hash different data. To avoid hashing the same data several times
3057 (which might not even be possible if the data is received from a pipe),
3058 a snapshot of the current hash context can be taken and turned into a
3061 -- Function: gcry_error_t gcry_md_copy (gcry_md_hd_t *HANDLE_DST,
3062 gcry_md_hd_t HANDLE_SRC)
3064 Create a new digest object as an exact copy of the object described
3065 by handle HANDLE_SRC and store it in HANDLE_DST. The context is
3066 not reset and you can continue to hash data using this context and
3067 independently using the original context.
3069 Now that we have prepared everything to calculate hashes, it is time
3070 to see how it is actually done. There are two ways for this, one to
3071 update the hash with a block of memory and one macro to update the hash
3072 by just one character. Both methods can be used on the same hash
3075 -- Function: void gcry_md_write (gcry_md_hd_t H, const void *BUFFER,
3078 Pass LENGTH bytes of the data in BUFFER to the digest object with
3079 handle H to update the digest values. This function should be used
3080 for large blocks of data. If this function is used after the
3081 context has been finalized, it will keep on pushing the data
3082 through the algorithm specific transform function and change the
3083 context; however the results are not meaningful and this feature is
3084 only available to mitigate timing attacks.
3086 -- Function: void gcry_md_putc (gcry_md_hd_t H, int C)
3088 Pass the byte in C to the digest object with handle H to update the
3089 digest value. This is an efficient function, implemented as a
3090 macro to buffer the data before an actual update.
3092 The semantics of the hash functions do not provide for reading out
3093 intermediate message digests because the calculation must be finalized
3094 first. This finalization may for example include the number of bytes
3095 hashed in the message digest or some padding.
3097 -- Function: void gcry_md_final (gcry_md_hd_t H)
3099 Finalize the message digest calculation. This is not really needed
3100 because 'gcry_md_read' and 'gcry_md_extract' do this implicitly.
3101 After this has been done no further updates (by means of
3102 'gcry_md_write' or 'gcry_md_putc' should be done; However, to
3103 mitigate timing attacks it is sometimes useful to keep on updating
3104 the context after having stored away the actual digest. Only the
3105 first call to this function has an effect. It is implemented as a
3108 The way to read out the calculated message digest is by using the
3111 -- Function: unsigned char * gcry_md_read (gcry_md_hd_t H, int ALGO)
3113 'gcry_md_read' returns the message digest after finalizing the
3114 calculation. This function may be used as often as required but it
3115 will always return the same value for one handle. The returned
3116 message digest is allocated within the message context and
3117 therefore valid until the handle is released or reset-ed (using
3118 'gcry_md_close' or 'gcry_md_reset' or it has been updated as a
3119 mitigation measure against timing attacks. ALGO may be given as 0
3120 to return the only enabled message digest or it may specify one of
3121 the enabled algorithms. The function does return 'NULL' if the
3122 requested algorithm has not been enabled.
3124 The way to read output of extendable-output function is by using the
3127 -- Function: gpg_err_code_t gcry_md_extract (gcry_md_hd_t H, int ALGO,
3128 void *BUFFER, size_t LENGTH)
3130 'gcry_mac_read' returns output from extendable-output function.
3131 This function may be used as often as required to generate more
3132 output byte stream from the algorithm. Function extracts the new
3133 output bytes to BUFFER of the length LENGTH. Buffer will be fully
3134 populated with new output. ALGO may be given as 0 to return the
3135 only enabled message digest or it may specify one of the enabled
3136 algorithms. The function does return non-zero value if the
3137 requested algorithm has not been enabled.
3139 Because it is often necessary to get the message digest of blocks of
3140 memory, two fast convenience function are available for this task:
3142 -- Function: gpg_err_code_t gcry_md_hash_buffers ( int ALGO,
3143 unsigned int FLAGS, void *DIGEST, const gcry_buffer_t *IOV,
3146 'gcry_md_hash_buffers' is a shortcut function to calculate a
3147 message digest from several buffers. This function does not
3148 require a context and immediately returns the message digest of the
3149 data described by IOV and IOVCNT. DIGEST must be allocated by the
3150 caller, large enough to hold the message digest yielded by the the
3151 specified algorithm ALGO. This required size may be obtained by
3152 using the function 'gcry_md_get_algo_dlen'.
3154 IOV is an array of buffer descriptions with IOVCNT items. The
3155 caller should zero out the structures in this array and for each
3156 array item set the fields '.data' to the address of the data to be
3157 hashed, '.len' to number of bytes to be hashed. If .OFF is also
3158 set, the data is taken starting at .OFF bytes from the begin of the
3159 buffer. The field '.size' is not used.
3161 The only supported flag value for FLAGS is GCRY_MD_FLAG_HMAC which
3162 turns this function into a HMAC function; the first item in IOV is
3163 then used as the key.
3165 On success the function returns 0 and stores the resulting hash or
3168 -- Function: void gcry_md_hash_buffer (int ALGO, void *DIGEST, const
3169 void *BUFFER, size_t LENGTH);
3171 'gcry_md_hash_buffer' is a shortcut function to calculate a message
3172 digest of a buffer. This function does not require a context and
3173 immediately returns the message digest of the LENGTH bytes at
3174 BUFFER. DIGEST must be allocated by the caller, large enough to
3175 hold the message digest yielded by the the specified algorithm
3176 ALGO. This required size may be obtained by using the function
3177 'gcry_md_get_algo_dlen'.
3179 Note that in contrast to 'gcry_md_hash_buffers' this function will
3180 abort the process if an unavailable algorithm is used.
3182 Hash algorithms are identified by internal algorithm numbers (see
3183 'gcry_md_open' for a list). However, in most applications they are used
3184 by names, so two functions are available to map between string
3185 representations and hash algorithm identifiers.
3187 -- Function: const char * gcry_md_algo_name (int ALGO)
3189 Map the digest algorithm id ALGO to a string representation of the
3190 algorithm name. For unknown algorithms this function returns the
3191 string '"?"'. This function should not be used to test for the
3192 availability of an algorithm.
3194 -- Function: int gcry_md_map_name (const char *NAME)
3196 Map the algorithm with NAME to a digest algorithm identifier.
3197 Returns 0 if the algorithm name is not known. Names representing
3198 ASN.1 object identifiers are recognized if the IETF dotted format
3199 is used and the OID is prefixed with either "'oid.'" or "'OID.'".
3200 For a list of supported OIDs, see the source code at 'cipher/md.c'.
3201 This function should not be used to test for the availability of an
3204 -- Function: gcry_error_t gcry_md_get_asnoid (int ALGO, void *BUFFER,
3207 Return an DER encoded ASN.1 OID for the algorithm ALGO in the user
3208 allocated BUFFER. LENGTH must point to variable with the available
3209 size of BUFFER and receives after return the actual size of the
3210 returned OID. The returned error code may be 'GPG_ERR_TOO_SHORT' if
3211 the provided buffer is to short to receive the OID; it is possible
3212 to call the function with 'NULL' for BUFFER to have it only return
3213 the required size. The function returns 0 on success.
3215 To test whether an algorithm is actually available for use, the
3216 following macro should be used:
3218 -- Function: gcry_error_t gcry_md_test_algo (int ALGO)
3220 The macro returns 0 if the algorithm ALGO is available for use.
3222 If the length of a message digest is not known, it can be retrieved
3223 using the following function:
3225 -- Function: unsigned int gcry_md_get_algo_dlen (int ALGO)
3227 Retrieve the length in bytes of the digest yielded by algorithm
3228 ALGO. This is often used prior to 'gcry_md_read' to allocate
3229 sufficient memory for the digest.
3231 In some situations it might be hard to remember the algorithm used
3232 for the ongoing hashing. The following function might be used to get
3235 -- Function: int gcry_md_get_algo (gcry_md_hd_t H)
3237 Retrieve the algorithm used with the handle H. Note that this does
3238 not work reliable if more than one algorithm is enabled in H.
3240 The following macro might also be useful:
3242 -- Function: int gcry_md_is_secure (gcry_md_hd_t H)
3244 This function returns true when the digest object H is allocated in
3245 "secure memory"; i.e. H was created with the
3246 'GCRY_MD_FLAG_SECURE'.
3248 -- Function: int gcry_md_is_enabled (gcry_md_hd_t H, int ALGO)
3250 This function returns true when the algorithm ALGO has been enabled
3251 for the digest object H.
3253 Tracking bugs related to hashing is often a cumbersome task which
3254 requires to add a lot of printf statements into the code. Libgcrypt
3255 provides an easy way to avoid this. The actual data hashed can be
3256 written to files on request.
3258 -- Function: void gcry_md_debug (gcry_md_hd_t H, const char *SUFFIX)
3260 Enable debugging for the digest object with handle H. This creates
3261 files named 'dbgmd-<n>.<string>' while doing the actual hashing.
3262 SUFFIX is the string part in the filename. The number is a counter
3263 incremented for each new hashing. The data in the file is the raw
3264 data as passed to 'gcry_md_write' or 'gcry_md_putc'. If 'NULL' is
3265 used for SUFFIX, the debugging is stopped and the file closed.
3266 This is only rarely required because 'gcry_md_close' implicitly
3270 File: gcrypt.info, Node: Message Authentication Codes, Next: Key Derivation, Prev: Hashing, Up: Top
3272 8 Message Authentication Codes
3273 ******************************
3275 Libgcrypt provides an easy and consistent to use interface for
3276 generating Message Authentication Codes (MAC). MAC generation is
3277 buffered and interface similar to the one used with hash algorithms.
3278 The programming model follows an open/process/close paradigm and is in
3279 that similar to other building blocks provided by Libgcrypt.
3283 * Available MAC algorithms:: List of MAC algorithms supported by the library.
3284 * Working with MAC algorithms:: List of functions related to MAC algorithms.
3287 File: gcrypt.info, Node: Available MAC algorithms, Next: Working with MAC algorithms, Up: Message Authentication Codes
3289 8.1 Available MAC algorithms
3290 ============================
3293 This is not a real algorithm but used by some functions as an error
3294 return value. This constant is guaranteed to have the value '0'.
3296 'GCRY_MAC_HMAC_SHA256'
3297 This is keyed-hash message authentication code (HMAC) message
3298 authentication algorithm based on the SHA-256 hash algorithm.
3300 'GCRY_MAC_HMAC_SHA224'
3301 This is HMAC message authentication algorithm based on the SHA-224
3304 'GCRY_MAC_HMAC_SHA512'
3305 This is HMAC message authentication algorithm based on the SHA-512
3308 'GCRY_MAC_HMAC_SHA384'
3309 This is HMAC message authentication algorithm based on the SHA-384
3312 'GCRY_MAC_HMAC_SHA3_256'
3313 This is HMAC message authentication algorithm based on the SHA3-384
3316 'GCRY_MAC_HMAC_SHA3_224'
3317 This is HMAC message authentication algorithm based on the SHA3-224
3320 'GCRY_MAC_HMAC_SHA3_512'
3321 This is HMAC message authentication algorithm based on the SHA3-512
3324 'GCRY_MAC_HMAC_SHA3_384'
3325 This is HMAC message authentication algorithm based on the SHA3-384
3328 'GCRY_MAC_HMAC_SHA1'
3329 This is HMAC message authentication algorithm based on the SHA-1
3333 This is HMAC message authentication algorithm based on the MD5 hash
3337 This is HMAC message authentication algorithm based on the MD4 hash
3340 'GCRY_MAC_HMAC_RMD160'
3341 This is HMAC message authentication algorithm based on the
3342 RIPE-MD-160 hash algorithm.
3344 'GCRY_MAC_HMAC_WHIRLPOOL'
3345 This is HMAC message authentication algorithm based on the
3346 WHIRLPOOL hash algorithm.
3348 'GCRY_MAC_HMAC_GOSTR3411_94'
3349 This is HMAC message authentication algorithm based on the GOST R
3350 34.11-94 hash algorithm.
3352 'GCRY_MAC_HMAC_STRIBOG256'
3353 This is HMAC message authentication algorithm based on the 256-bit
3354 hash algorithm described in GOST R 34.11-2012.
3356 'GCRY_MAC_HMAC_STRIBOG512'
3357 This is HMAC message authentication algorithm based on the 512-bit
3358 hash algorithm described in GOST R 34.11-2012.
3361 This is CMAC (Cipher-based MAC) message authentication algorithm
3362 based on the AES block cipher algorithm.
3364 'GCRY_MAC_CMAC_3DES'
3365 This is CMAC message authentication algorithm based on the
3366 three-key EDE Triple-DES block cipher algorithm.
3368 'GCRY_MAC_CMAC_CAMELLIA'
3369 This is CMAC message authentication algorithm based on the Camellia
3370 block cipher algorithm.
3372 'GCRY_MAC_CMAC_CAST5'
3373 This is CMAC message authentication algorithm based on the
3374 CAST128-5 block cipher algorithm.
3376 'GCRY_MAC_CMAC_BLOWFISH'
3377 This is CMAC message authentication algorithm based on the Blowfish
3378 block cipher algorithm.
3380 'GCRY_MAC_CMAC_TWOFISH'
3381 This is CMAC message authentication algorithm based on the Twofish
3382 block cipher algorithm.
3384 'GCRY_MAC_CMAC_SERPENT'
3385 This is CMAC message authentication algorithm based on the Serpent
3386 block cipher algorithm.
3388 'GCRY_MAC_CMAC_SEED'
3389 This is CMAC message authentication algorithm based on the SEED
3390 block cipher algorithm.
3392 'GCRY_MAC_CMAC_RFC2268'
3393 This is CMAC message authentication algorithm based on the Ron's
3394 Cipher 2 block cipher algorithm.
3396 'GCRY_MAC_CMAC_IDEA'
3397 This is CMAC message authentication algorithm based on the IDEA
3398 block cipher algorithm.
3400 'GCRY_MAC_CMAC_GOST28147'
3401 This is CMAC message authentication algorithm based on the GOST
3402 28147-89 block cipher algorithm.
3405 This is GMAC (GCM mode based MAC) message authentication algorithm
3406 based on the AES block cipher algorithm.
3408 'GCRY_MAC_GMAC_CAMELLIA'
3409 This is GMAC message authentication algorithm based on the Camellia
3410 block cipher algorithm.
3412 'GCRY_MAC_GMAC_TWOFISH'
3413 This is GMAC message authentication algorithm based on the Twofish
3414 block cipher algorithm.
3416 'GCRY_MAC_GMAC_SERPENT'
3417 This is GMAC message authentication algorithm based on the Serpent
3418 block cipher algorithm.
3420 'GCRY_MAC_GMAC_SEED'
3421 This is GMAC message authentication algorithm based on the SEED
3422 block cipher algorithm.
3425 This is plain Poly1305 message authentication algorithm, used with
3428 'GCRY_MAC_POLY1305_AES'
3429 This is Poly1305-AES message authentication algorithm, used with
3430 key and one-time nonce.
3432 'GCRY_MAC_POLY1305_CAMELLIA'
3433 This is Poly1305-Camellia message authentication algorithm, used
3434 with key and one-time nonce.
3436 'GCRY_MAC_POLY1305_TWOFISH'
3437 This is Poly1305-Twofish message authentication algorithm, used
3438 with key and one-time nonce.
3440 'GCRY_MAC_POLY1305_SERPENT'
3441 This is Poly1305-Serpent message authentication algorithm, used
3442 with key and one-time nonce.
3444 'GCRY_MAC_POLY1305_SEED'
3445 This is Poly1305-SEED message authentication algorithm, used with
3446 key and one-time nonce.
3449 File: gcrypt.info, Node: Working with MAC algorithms, Prev: Available MAC algorithms, Up: Message Authentication Codes
3451 8.2 Working with MAC algorithms
3452 ===============================
3454 To use most of these function it is necessary to create a context; this
3457 -- Function: gcry_error_t gcry_mac_open (gcry_mac_hd_t *HD, int ALGO,
3458 unsigned int FLAGS, gcry_ctx_t CTX)
3460 Create a MAC object for algorithm ALGO. FLAGS may be given as an
3461 bitwise OR of constants described below. HD is guaranteed to
3462 either receive a valid handle or NULL. CTX is context object to
3463 associate MAC object with. CTX maybe set to NULL.
3465 For a list of supported algorithms, see *note Available MAC
3468 The flags allowed for MODE are:
3470 'GCRY_MAC_FLAG_SECURE'
3471 Allocate all buffers and the resulting MAC in "secure memory".
3472 Use this if the MAC data is highly confidential.
3474 In order to use a handle for performing MAC algorithm operations, a
3475 'key' has to be set first:
3477 -- Function: gcry_error_t gcry_mac_setkey (gcry_mac_hd_t H, const void
3478 *KEY, size_t KEYLEN)
3480 Set the MAC key to the value of KEY of length KEYLEN bytes. With
3481 HMAC algorithms, there is no restriction on the length of the key.
3482 With CMAC algorithms, the length of the key is restricted to those
3483 supported by the underlying block cipher.
3485 GMAC algorithms and Poly1305-with-cipher algorithms need
3486 initialization vector to be set, which can be performed with function:
3488 -- Function: gcry_error_t gcry_mac_setiv (gcry_mac_hd_t H, const void
3491 Set the IV to the value of IV of length IVLEN bytes.
3493 After you are done with the MAC calculation, you should release the
3496 -- Function: void gcry_mac_close (gcry_mac_hd_t H)
3498 Release all resources of MAC context H. H should not be used after
3499 a call to this function. A 'NULL' passed as H is ignored. The
3500 function also clears all sensitive information associated with this
3503 Often you have to do several MAC operations using the same algorithm.
3504 To avoid the overhead of creating and releasing context, a reset
3505 function is provided:
3507 -- Function: gcry_error_t gcry_mac_reset (gcry_mac_hd_t H)
3509 Reset the current context to its initial state. This is
3510 effectively identical to a close followed by an open and setting
3513 Note that gcry_mac_reset is implemented as a macro.
3515 Now that we have prepared everything to calculate MAC, it is time to
3516 see how it is actually done.
3518 -- Function: gcry_error_t gcry_mac_write (gcry_mac_hd_t H, const void
3519 *BUFFER, size_t LENGTH)
3521 Pass LENGTH bytes of the data in BUFFER to the MAC object with
3522 handle H to update the MAC values. If this function is used after
3523 the context has been finalized, it will keep on pushing the data
3524 through the algorithm specific transform function and thereby
3525 change the context; however the results are not meaningful and this
3526 feature is only available to mitigate timing attacks.
3528 The way to read out the calculated MAC is by using the function:
3530 -- Function: gcry_error_t gcry_mac_read (gcry_mac_hd_t H, void *BUFFER,
3533 'gcry_mac_read' returns the MAC after finalizing the calculation.
3534 Function copies the resulting MAC value to BUFFER of the length
3535 LENGTH. If LENGTH is larger than length of resulting MAC value,
3536 then length of MAC is returned through LENGTH.
3538 To compare existing MAC value with recalculated MAC, one is to use
3541 -- Function: gcry_error_t gcry_mac_verify (gcry_mac_hd_t H, void
3542 *BUFFER, size_t LENGTH)
3544 'gcry_mac_verify' finalizes MAC calculation and compares result
3545 with LENGTH bytes of data in BUFFER. Error code 'GPG_ERR_CHECKSUM'
3546 is returned if the MAC value in the buffer BUFFER does not match
3547 the MAC calculated in object H.
3549 In some situations it might be hard to remember the algorithm used
3550 for the MAC calculation. The following function might be used to get
3553 -- Function: int gcry_mac_get_algo (gcry_mac_hd_t H)
3555 Retrieve the algorithm used with the handle H.
3557 MAC algorithms are identified by internal algorithm numbers (see
3558 'gcry_mac_open' for a list). However, in most applications they are
3559 used by names, so two functions are available to map between string
3560 representations and MAC algorithm identifiers.
3562 -- Function: const char * gcry_mac_algo_name (int ALGO)
3564 Map the MAC algorithm id ALGO to a string representation of the
3565 algorithm name. For unknown algorithms this function returns the
3566 string '"?"'. This function should not be used to test for the
3567 availability of an algorithm.
3569 -- Function: int gcry_mac_map_name (const char *NAME)
3571 Map the algorithm with NAME to a MAC algorithm identifier. Returns
3572 0 if the algorithm name is not known. This function should not be
3573 used to test for the availability of an algorithm.
3575 To test whether an algorithm is actually available for use, the
3576 following macro should be used:
3578 -- Function: gcry_error_t gcry_mac_test_algo (int ALGO)
3580 The macro returns 0 if the MAC algorithm ALGO is available for use.
3582 If the length of a message digest is not known, it can be retrieved
3583 using the following function:
3585 -- Function: unsigned int gcry_mac_get_algo_maclen (int ALGO)
3587 Retrieve the length in bytes of the MAC yielded by algorithm ALGO.
3588 This is often used prior to 'gcry_mac_read' to allocate sufficient
3589 memory for the MAC value. On error '0' is returned.
3591 -- Function: unsigned int gcry_mac_get_algo_keylen (ALGO)
3593 This function returns length of the key for MAC algorithm ALGO. If
3594 the algorithm supports multiple key lengths, the default supported
3595 key length is returned. On error '0' is returned. The key length
3596 is returned as number of octets.
3599 File: gcrypt.info, Node: Key Derivation, Next: Random Numbers, Prev: Message Authentication Codes, Up: Top
3604 Libgcypt provides a general purpose function to derive keys from
3607 -- Function: gpg_error_t gcry_kdf_derive ( const void *PASSPHRASE,
3608 size_t PASSPHRASELEN, int ALGO, int SUBALGO, const void *SALT,
3609 size_t SALTLEN, unsigned long ITERATIONS, size_t KEYSIZE,
3612 Derive a key from a passphrase. KEYSIZE gives the requested size
3613 of the keys in octets. KEYBUFFER is a caller provided buffer
3614 filled on success with the derived key. The input passphrase is
3615 taken from PASSPHRASE which is an arbitrary memory buffer of
3616 PASSPHRASELEN octets. ALGO specifies the KDF algorithm to use; see
3617 below. SUBALGO specifies an algorithm used internally by the KDF
3618 algorithms; this is usually a hash algorithm but certain KDF
3619 algorithms may use it differently. SALT is a salt of length
3620 SALTLEN octets, as needed by most KDF algorithms. ITERATIONS is a
3621 positive integer parameter to most KDFs.
3623 On success 0 is returned; on failure an error code.
3625 Currently supported KDFs (parameter ALGO):
3627 'GCRY_KDF_SIMPLE_S2K'
3628 The OpenPGP simple S2K algorithm (cf. RFC4880). Its use is
3629 strongly deprecated. SALT and ITERATIONS are not needed and
3630 may be passed as 'NULL'/'0'.
3632 'GCRY_KDF_SALTED_S2K'
3633 The OpenPGP salted S2K algorithm (cf. RFC4880). Usually not
3634 used. ITERATIONS is not needed and may be passed as '0'.
3635 SALTLEN must be given as 8.
3637 'GCRY_KDF_ITERSALTED_S2K'
3638 The OpenPGP iterated+salted S2K algorithm (cf. RFC4880).
3639 This is the default for most OpenPGP applications. SALTLEN
3640 must be given as 8. Note that OpenPGP defines a special
3641 encoding of the ITERATIONS; however this function takes the
3642 plain decoded iteration count.
3645 The PKCS#5 Passphrase Based Key Derivation Function number 2.
3648 The SCRYPT Key Derivation Function. The subalgorithm is used
3649 to specify the CPU/memory cost parameter N, and the number of
3650 iterations is used for the parallelization parameter p. The
3651 block size is fixed at 8 in the current implementation.
3654 File: gcrypt.info, Node: Random Numbers, Next: S-expressions, Prev: Key Derivation, Up: Top
3661 * Quality of random numbers:: Libgcrypt uses different quality levels.
3662 * Retrieving random numbers:: How to retrieve random numbers.
3665 File: gcrypt.info, Node: Quality of random numbers, Next: Retrieving random numbers, Up: Random Numbers
3667 10.1 Quality of random numbers
3668 ==============================
3670 Libgcypt offers random numbers of different quality levels:
3672 -- Data type: gcry_random_level_t
3673 The constants for the random quality levels are of this enum type.
3676 For all functions, except for 'gcry_mpi_randomize', this level maps
3677 to GCRY_STRONG_RANDOM. If you do not want this, consider using
3678 'gcry_create_nonce'.
3679 'GCRY_STRONG_RANDOM'
3680 Use this level for session keys and similar purposes.
3681 'GCRY_VERY_STRONG_RANDOM'
3682 Use this level for long term key material.
3685 File: gcrypt.info, Node: Retrieving random numbers, Prev: Quality of random numbers, Up: Random Numbers
3687 10.2 Retrieving random numbers
3688 ==============================
3690 -- Function: void gcry_randomize (unsigned char *BUFFER, size_t LENGTH,
3691 enum gcry_random_level LEVEL)
3693 Fill BUFFER with LENGTH random bytes using a random quality as
3696 -- Function: void * gcry_random_bytes (size_t NBYTES, enum
3697 gcry_random_level LEVEL)
3699 Convenience function to allocate a memory block consisting of
3700 NBYTES fresh random bytes using a random quality as defined by
3703 -- Function: void * gcry_random_bytes_secure (size_t NBYTES, enum
3704 gcry_random_level LEVEL)
3706 Convenience function to allocate a memory block consisting of
3707 NBYTES fresh random bytes using a random quality as defined by
3708 LEVEL. This function differs from 'gcry_random_bytes' in that the
3709 returned buffer is allocated in a "secure" area of the memory.
3711 -- Function: void gcry_create_nonce (unsigned char *BUFFER, size_t
3714 Fill BUFFER with LENGTH unpredictable bytes. This is commonly
3715 called a nonce and may also be used for initialization vectors and
3716 padding. This is an extra function nearly independent of the other
3717 random function for 3 reasons: It better protects the regular
3718 random generator's internal state, provides better performance and
3719 does not drain the precious entropy pool.
3722 File: gcrypt.info, Node: S-expressions, Next: MPI library, Prev: Random Numbers, Up: Top
3727 S-expressions are used by the public key functions to pass complex data
3728 structures around. These LISP like objects are used by some
3729 cryptographic protocols (cf. RFC-2692) and Libgcrypt provides functions
3730 to parse and construct them. For detailed information, see 'Ron Rivest,
3731 code and description of S-expressions,
3732 <http://theory.lcs.mit.edu/~rivest/sexp.html>'.
3736 * Data types for S-expressions:: Data types related with S-expressions.
3737 * Working with S-expressions:: How to work with S-expressions.
3740 File: gcrypt.info, Node: Data types for S-expressions, Next: Working with S-expressions, Up: S-expressions
3742 11.1 Data types for S-expressions
3743 =================================
3745 -- Data type: gcry_sexp_t
3746 The 'gcry_sexp_t' type describes an object with the Libgcrypt
3747 internal representation of an S-expression.
3750 File: gcrypt.info, Node: Working with S-expressions, Prev: Data types for S-expressions, Up: S-expressions
3752 11.2 Working with S-expressions
3753 ===============================
3755 There are several functions to create an Libgcrypt S-expression object
3756 from its external representation or from a string template. There is
3757 also a function to convert the internal representation back into one of
3758 the external formats:
3760 -- Function: gcry_error_t gcry_sexp_new (gcry_sexp_t *R_SEXP,
3761 const void *BUFFER, size_t LENGTH, int AUTODETECT)
3763 This is the generic function to create an new S-expression object
3764 from its external representation in BUFFER of LENGTH bytes. On
3765 success the result is stored at the address given by R_SEXP. With
3766 AUTODETECT set to 0, the data in BUFFER is expected to be in
3767 canonized format, with AUTODETECT set to 1 the parses any of the
3768 defined external formats. If BUFFER does not hold a valid
3769 S-expression an error code is returned and R_SEXP set to 'NULL'.
3770 Note that the caller is responsible for releasing the newly
3771 allocated S-expression using 'gcry_sexp_release'.
3773 -- Function: gcry_error_t gcry_sexp_create (gcry_sexp_t *R_SEXP,
3774 void *BUFFER, size_t LENGTH, int AUTODETECT,
3775 void (*FREEFNC)(void*))
3777 This function is identical to 'gcry_sexp_new' but has an extra
3778 argument FREEFNC, which, when not set to 'NULL', is expected to be
3779 a function to release the BUFFER; most likely the standard 'free'
3780 function is used for this argument. This has the effect of
3781 transferring the ownership of BUFFER to the created object in
3782 R_SEXP. The advantage of using this function is that Libgcrypt
3783 might decide to directly use the provided buffer and thus avoid
3786 -- Function: gcry_error_t gcry_sexp_sscan (gcry_sexp_t *R_SEXP,
3787 size_t *ERROFF, const char *BUFFER, size_t LENGTH)
3789 This is another variant of the above functions. It behaves nearly
3790 identical but provides an ERROFF argument which will receive the
3791 offset into the buffer where the parsing stopped on error.
3793 -- Function: gcry_error_t gcry_sexp_build (gcry_sexp_t *R_SEXP,
3794 size_t *ERROFF, const char *FORMAT, ...)
3796 This function creates an internal S-expression from the string
3797 template FORMAT and stores it at the address of R_SEXP. If there
3798 is a parsing error, the function returns an appropriate error code
3799 and stores the offset into FORMAT where the parsing stopped in
3800 ERROFF. The function supports a couple of printf-like formatting
3801 characters and expects arguments for some of these escape sequences
3802 right after FORMAT. The following format characters are defined:
3805 The next argument is expected to be of type 'gcry_mpi_t' and a
3806 copy of its value is inserted into the resulting S-expression.
3807 The MPI is stored as a signed integer.
3809 The next argument is expected to be of type 'gcry_mpi_t' and a
3810 copy of its value is inserted into the resulting S-expression.
3811 The MPI is stored as an unsigned integer.
3813 The next argument is expected to be of type 'char *' and that
3814 string is inserted into the resulting S-expression.
3816 The next argument is expected to be of type 'int' and its
3817 value is inserted into the resulting S-expression.
3819 The next argument is expected to be of type 'unsigned int' and
3820 its value is inserted into the resulting S-expression.
3822 The next argument is expected to be of type 'int' directly
3823 followed by an argument of type 'char *'. This represents a
3824 buffer of given length to be inserted into the resulting
3827 The next argument is expected to be of type 'gcry_sexp_t' and
3828 a copy of that S-expression is embedded in the resulting
3829 S-expression. The argument needs to be a regular
3830 S-expression, starting with a parenthesis.
3832 No other format characters are defined and would return an error.
3833 Note that the format character '%%' does not exists, because a
3834 percent sign is not a valid character in an S-expression.
3836 -- Function: void gcry_sexp_release (gcry_sexp_t SEXP)
3838 Release the S-expression object SEXP. If the S-expression is
3839 stored in secure memory it explicitly zeroises that memory; note
3840 that this is done in addition to the zeroisation always done when
3841 freeing secure memory.
3843 The next 2 functions are used to convert the internal representation
3844 back into a regular external S-expression format and to show the
3845 structure for debugging.
3847 -- Function: size_t gcry_sexp_sprint (gcry_sexp_t SEXP, int MODE,
3848 char *BUFFER, size_t MAXLENGTH)
3850 Copies the S-expression object SEXP into BUFFER using the format
3851 specified in MODE. MAXLENGTH must be set to the allocated length
3852 of BUFFER. The function returns the actual length of valid bytes
3853 put into BUFFER or 0 if the provided buffer is too short. Passing
3854 'NULL' for BUFFER returns the required length for BUFFER. For
3855 convenience reasons an extra byte with value 0 is appended to the
3858 The following formats are supported:
3860 'GCRYSEXP_FMT_DEFAULT'
3861 Returns a convenient external S-expression representation.
3863 'GCRYSEXP_FMT_CANON'
3864 Return the S-expression in canonical format.
3866 'GCRYSEXP_FMT_BASE64'
3867 Not currently supported.
3869 'GCRYSEXP_FMT_ADVANCED'
3870 Returns the S-expression in advanced format.
3872 -- Function: void gcry_sexp_dump (gcry_sexp_t SEXP)
3874 Dumps SEXP in a format suitable for debugging to Libgcrypt's
3877 Often canonical encoding is used in the external representation. The
3878 following function can be used to check for valid encoding and to learn
3879 the length of the S-expression.
3881 -- Function: size_t gcry_sexp_canon_len (const unsigned char *BUFFER,
3882 size_t LENGTH, size_t *ERROFF, int *ERRCODE)
3884 Scan the canonical encoded BUFFER with implicit length values and
3885 return the actual length this S-expression uses. For a valid
3886 S-expression it should never return 0. If LENGTH is not 0, the
3887 maximum length to scan is given; this can be used for syntax checks
3888 of data passed from outside. ERRCODE and ERROFF may both be passed
3891 There are functions to parse S-expressions and retrieve elements:
3893 -- Function: gcry_sexp_t gcry_sexp_find_token (const gcry_sexp_t LIST,
3894 const char *TOKEN, size_t TOKLEN)
3896 Scan the S-expression for a sublist with a type (the car of the
3897 list) matching the string TOKEN. If TOKLEN is not 0, the token is
3898 assumed to be raw memory of this length. The function returns a
3899 newly allocated S-expression consisting of the found sublist or
3900 'NULL' when not found.
3902 -- Function: int gcry_sexp_length (const gcry_sexp_t LIST)
3904 Return the length of the LIST. For a valid S-expression this
3905 should be at least 1.
3907 -- Function: gcry_sexp_t gcry_sexp_nth (const gcry_sexp_t LIST,
3910 Create and return a new S-expression from the element with index
3911 NUMBER in LIST. Note that the first element has the index 0. If
3912 there is no such element, 'NULL' is returned.
3914 -- Function: gcry_sexp_t gcry_sexp_car (const gcry_sexp_t LIST)
3916 Create and return a new S-expression from the first element in
3917 LIST; this is called the "type" and should always exist per
3918 S-expression specification and in general be a string. 'NULL' is
3919 returned in case of a problem.
3921 -- Function: gcry_sexp_t gcry_sexp_cdr (const gcry_sexp_t LIST)
3923 Create and return a new list form all elements except for the first
3924 one. Note that this function may return an invalid S-expression
3925 because it is not guaranteed, that the type exists and is a string.
3926 However, for parsing a complex S-expression it might be useful for
3927 intermediate lists. Returns 'NULL' on error.
3929 -- Function: const char * gcry_sexp_nth_data (const gcry_sexp_t LIST,
3930 int NUMBER, size_t *DATALEN)
3932 This function is used to get data from a LIST. A pointer to the
3933 actual data with index NUMBER is returned and the length of this
3934 data will be stored to DATALEN. If there is no data at the given
3935 index or the index represents another list, 'NULL' is returned.
3936 *Caution:* The returned pointer is valid as long as LIST is not
3937 modified or released.
3939 Here is an example on how to extract and print the surname (Meier)
3940 from the S-expression '(Name Otto Meier (address Burgplatz 3))':
3945 name = gcry_sexp_nth_data (list, 2, &len);
3946 printf ("my name is %.*s\n", (int)len, name);
3948 -- Function: void * gcry_sexp_nth_buffer (const gcry_sexp_t LIST,
3949 int NUMBER, size_t *RLENGTH)
3951 This function is used to get data from a LIST. A malloced buffer
3952 with the actual data at list index NUMBER is returned and the
3953 length of this buffer will be stored to RLENGTH. If there is no
3954 data at the given index or the index represents another list,
3955 'NULL' is returned. The caller must release the result using
3958 Here is an example on how to extract and print the CRC value from
3959 the S-expression '(hash crc32 #23ed00d7)':
3964 value = gcry_sexp_nth_buffer (list, 2, &len);
3966 fwrite (value, len, 1, stdout);
3969 -- Function: char * gcry_sexp_nth_string (gcry_sexp_t LIST, int NUMBER)
3971 This function is used to get and convert data from a LIST. The
3972 data is assumed to be a Nul terminated string. The caller must
3973 release this returned value using 'gcry_free'. If there is no data
3974 at the given index, the index represents a list or the value can't
3975 be converted to a string, 'NULL' is returned.
3977 -- Function: gcry_mpi_t gcry_sexp_nth_mpi (gcry_sexp_t LIST,
3978 int NUMBER, int MPIFMT)
3980 This function is used to get and convert data from a LIST. This
3981 data is assumed to be an MPI stored in the format described by
3982 MPIFMT and returned as a standard Libgcrypt MPI. The caller must
3983 release this returned value using 'gcry_mpi_release'. If there is
3984 no data at the given index, the index represents a list or the
3985 value can't be converted to an MPI, 'NULL' is returned. If you use
3986 this function to parse results of a public key function, you most
3987 likely want to use 'GCRYMPI_FMT_USG'.
3989 -- Function: gpg_error_t gcry_sexp_extract_param ( gcry_sexp_t SEXP,
3990 const char *PATH, const char *LIST, ...)
3992 Extract parameters from an S-expression using a list of parameter
3993 names. The names of these parameters are specified in LIST. White
3994 space between the parameter names are ignored. Some special
3995 characters may be given to control the conversion:
3998 Switch to unsigned integer format (GCRYMPI_FMT_USG). This is
4001 Switch to standard signed format (GCRYMPI_FMT_STD).
4003 Switch to opaque MPI format. The resulting MPIs may not be
4004 used for computations; see 'gcry_mpi_get_opaque' for details.
4006 Switch to buffer descriptor mode. See below for details.
4008 If immediately following a parameter letter (no white space
4009 allowed), that parameter is considered optional.
4011 In general parameter names are single letters. To use a string for
4012 a parameter name, enclose the name in single quotes.
4014 Unless in buffer descriptor mode for each parameter name a pointer
4015 to an 'gcry_mpi_t' variable is expected that must be set to 'NULL'
4016 prior to invoking this function, and finally a 'NULL' is expected.
4019 gcry_sexp_extract_param (key, NULL, "n/x+e d-'foo'",
4020 &mpi_n, &mpi_x, &mpi_e, &mpi_d, &mpi_foo, NULL)
4022 stores the parameter 'n' from KEY as an unsigned MPI into MPI_N,
4023 the parameter 'x' as an opaque MPI into MPI_X, the parameters 'e'
4024 and 'd' again as an unsigned MPI into MPI_E and MPI_D and finally
4025 the parameter 'foo' as a signed MPI into MPI_FOO.
4027 PATH is an optional string used to locate a token. The exclamation
4028 mark separated tokens are used via 'gcry_sexp_find_token' to find a
4029 start point inside the S-expression.
4031 In buffer descriptor mode a pointer to a 'gcry_buffer_t' descriptor
4032 is expected instead of a pointer to an MPI. The caller may use two
4033 different operation modes here: If the DATA field of the provided
4034 descriptor is 'NULL', the function allocates a new buffer and
4035 stores it at DATA; the other fields are set accordingly with OFF
4036 set to 0. If DATA is not 'NULL', the function assumes that the
4037 DATA, SIZE, and OFF fields specify a buffer where to but the value
4038 of the respective parameter; on return the LEN field receives the
4039 number of bytes copied to that buffer; in case the buffer is too
4040 small, the function immediately returns with an error code (and LEN
4043 The function returns 0 on success. On error an error code is
4044 returned, all passed MPIs that might have been allocated up to this
4045 point are deallocated and set to 'NULL', and all passed buffers are
4046 either truncated if the caller supplied the buffer, or deallocated
4047 if the function allocated the buffer.
4050 File: gcrypt.info, Node: MPI library, Next: Prime numbers, Prev: S-expressions, Up: Top
4057 * Data types:: MPI related data types.
4058 * Basic functions:: First steps with MPI numbers.
4059 * MPI formats:: External representation of MPIs.
4060 * Calculations:: Performing MPI calculations.
4061 * Comparisons:: How to compare MPI values.
4062 * Bit manipulations:: How to access single bits of MPI values.
4063 * EC functions:: Elliptic curve related functions.
4064 * Miscellaneous:: Miscellaneous MPI functions.
4066 Public key cryptography is based on mathematics with large numbers. To
4067 implement the public key functions, a library for handling these large
4068 numbers is required. Because of the general usefulness of such a
4069 library, its interface is exposed by Libgcrypt. In the context of
4070 Libgcrypt and in most other applications, these large numbers are called
4071 MPIs (multi-precision-integers).
4074 File: gcrypt.info, Node: Data types, Next: Basic functions, Up: MPI library
4079 -- Data type: gcry_mpi_t
4080 This type represents an object to hold an MPI.
4082 -- Data type: gcry_mpi_point_t
4083 This type represents an object to hold a point for elliptic curve
4087 File: gcrypt.info, Node: Basic functions, Next: MPI formats, Prev: Data types, Up: MPI library
4089 12.2 Basic functions
4090 ====================
4092 To work with MPIs, storage must be allocated and released for the
4093 numbers. This can be done with one of these functions:
4095 -- Function: gcry_mpi_t gcry_mpi_new (unsigned int NBITS)
4097 Allocate a new MPI object, initialize it to 0 and initially
4098 allocate enough memory for a number of at least NBITS. This
4099 pre-allocation is only a small performance issue and not actually
4100 necessary because Libgcrypt automatically re-allocates the required
4103 -- Function: gcry_mpi_t gcry_mpi_snew (unsigned int NBITS)
4105 This is identical to 'gcry_mpi_new' but allocates the MPI in the so
4106 called "secure memory" which in turn will take care that all
4107 derived values will also be stored in this "secure memory". Use
4108 this for highly confidential data like private key parameters.
4110 -- Function: gcry_mpi_t gcry_mpi_copy (const gcry_mpi_t A)
4112 Create a new MPI as the exact copy of A but with the constant and
4113 immutable flags cleared.
4115 -- Function: void gcry_mpi_release (gcry_mpi_t A)
4117 Release the MPI A and free all associated resources. Passing
4118 'NULL' is allowed and ignored. When a MPI stored in the "secure
4119 memory" is released, that memory gets wiped out immediately.
4121 The simplest operations are used to assign a new value to an MPI:
4123 -- Function: gcry_mpi_t gcry_mpi_set (gcry_mpi_t W, const gcry_mpi_t U)
4125 Assign the value of U to W and return W. If 'NULL' is passed for
4126 W, a new MPI is allocated, set to the value of U and returned.
4128 -- Function: gcry_mpi_t gcry_mpi_set_ui (gcry_mpi_t W, unsigned long U)
4130 Assign the value of U to W and return W. If 'NULL' is passed for
4131 W, a new MPI is allocated, set to the value of U and returned.
4132 This function takes an 'unsigned int' as type for U and thus it is
4133 only possible to set W to small values (usually up to the word size
4136 -- Function: void gcry_mpi_swap (gcry_mpi_t A, gcry_mpi_t B)
4138 Swap the values of A and B.
4140 -- Function: void gcry_mpi_snatch (gcry_mpi_t W, const gcry_mpi_t U)
4142 Set U into W and release U. If W is 'NULL' only U will be
4145 -- Function: void gcry_mpi_neg (gcry_mpi_t W, gcry_mpi_t U)
4147 Set the sign of W to the negative of U.
4149 -- Function: void gcry_mpi_abs (gcry_mpi_t W)
4151 Clear the sign of W.
4154 File: gcrypt.info, Node: MPI formats, Next: Calculations, Prev: Basic functions, Up: MPI library
4159 The following functions are used to convert between an external
4160 representation of an MPI and the internal one of Libgcrypt.
4162 -- Function: gcry_error_t gcry_mpi_scan (gcry_mpi_t *R_MPI,
4163 enum gcry_mpi_format FORMAT, const unsigned char *BUFFER,
4164 size_t BUFLEN, size_t *NSCANNED)
4166 Convert the external representation of an integer stored in BUFFER
4167 with a length of BUFLEN into a newly created MPI returned which
4168 will be stored at the address of R_MPI. For certain formats the
4169 length argument is not required and should be passed as '0'. A
4170 BUFLEN larger than 16 MiByte will be rejected. After a successful
4171 operation the variable NSCANNED receives the number of bytes
4172 actually scanned unless NSCANNED was given as 'NULL'. FORMAT
4173 describes the format of the MPI as stored in BUFFER:
4176 2-complement stored without a length header. Note that
4177 'gcry_mpi_print' stores a '0' as a string of zero length.
4180 As used by OpenPGP (only defined as unsigned). This is
4181 basically 'GCRYMPI_FMT_STD' with a 2 byte big endian length
4182 header. A length header indicating a length of more than
4183 16384 is not allowed.
4186 As used in the Secure Shell protocol. This is
4187 'GCRYMPI_FMT_STD' with a 4 byte big endian header.
4190 Stored as a string with each byte of the MPI encoded as 2 hex
4191 digits. Negative numbers are prefix with a minus sign and in
4192 addition the high bit is always zero to make clear that an
4193 explicit sign ist used. When using this format, BUFLEN must
4197 Simple unsigned integer.
4199 Note that all of the above formats store the integer in big-endian
4202 -- Function: gcry_error_t gcry_mpi_print (enum gcry_mpi_format FORMAT,
4203 unsigned char *BUFFER, size_t BUFLEN, size_t *NWRITTEN,
4206 Convert the MPI A into an external representation described by
4207 FORMAT (see above) and store it in the provided BUFFER which has a
4208 usable length of at least the BUFLEN bytes. If NWRITTEN is not
4209 NULL, it will receive the number of bytes actually stored in BUFFER
4210 after a successful operation.
4212 -- Function: gcry_error_t gcry_mpi_aprint (enum gcry_mpi_format FORMAT,
4213 unsigned char **BUFFER, size_t *NBYTES, const gcry_mpi_t A)
4215 Convert the MPI A into an external representation described by
4216 FORMAT (see above) and store it in a newly allocated buffer which
4217 address will be stored in the variable BUFFER points to. The
4218 number of bytes stored in this buffer will be stored in the
4219 variable NBYTES points to, unless NBYTES is 'NULL'.
4221 Even if NBYTES is zero, the function allocates at least one byte
4222 and store a zero there. Thus with formats 'GCRYMPI_FMT_STD' and
4223 'GCRYMPI_FMT_USG' the caller may safely set a returned length of 0
4224 to 1 to represent a zero as a 1 byte string.
4226 -- Function: void gcry_mpi_dump (const gcry_mpi_t A)
4228 Dump the value of A in a format suitable for debugging to
4229 Libgcrypt's logging stream. Note that one leading space but no
4230 trailing space or linefeed will be printed. It is okay to pass
4234 File: gcrypt.info, Node: Calculations, Next: Comparisons, Prev: MPI formats, Up: MPI library
4239 Basic arithmetic operations:
4241 -- Function: void gcry_mpi_add (gcry_mpi_t W, gcry_mpi_t U,
4246 -- Function: void gcry_mpi_add_ui (gcry_mpi_t W, gcry_mpi_t U,
4249 W = U + V. Note that V is an unsigned integer.
4251 -- Function: void gcry_mpi_addm (gcry_mpi_t W, gcry_mpi_t U,
4252 gcry_mpi_t V, gcry_mpi_t M)
4256 -- Function: void gcry_mpi_sub (gcry_mpi_t W, gcry_mpi_t U,
4261 -- Function: void gcry_mpi_sub_ui (gcry_mpi_t W, gcry_mpi_t U,
4264 W = U - V. V is an unsigned integer.
4266 -- Function: void gcry_mpi_subm (gcry_mpi_t W, gcry_mpi_t U,
4267 gcry_mpi_t V, gcry_mpi_t M)
4271 -- Function: void gcry_mpi_mul (gcry_mpi_t W, gcry_mpi_t U,
4276 -- Function: void gcry_mpi_mul_ui (gcry_mpi_t W, gcry_mpi_t U,
4279 W = U * V. V is an unsigned integer.
4281 -- Function: void gcry_mpi_mulm (gcry_mpi_t W, gcry_mpi_t U,
4282 gcry_mpi_t V, gcry_mpi_t M)
4286 -- Function: void gcry_mpi_mul_2exp (gcry_mpi_t W, gcry_mpi_t U,
4291 -- Function: void gcry_mpi_div (gcry_mpi_t Q, gcry_mpi_t R,
4292 gcry_mpi_t DIVIDEND, gcry_mpi_t DIVISOR, int ROUND)
4294 Q = DIVIDEND / DIVISOR, R = DIVIDEND \bmod DIVISOR. Q and R may be
4295 passed as 'NULL'. ROUND should be negative or 0.
4297 -- Function: void gcry_mpi_mod (gcry_mpi_t R, gcry_mpi_t DIVIDEND,
4300 R = DIVIDEND \bmod DIVISOR.
4302 -- Function: void gcry_mpi_powm (gcry_mpi_t W, const gcry_mpi_t B,
4303 const gcry_mpi_t E, const gcry_mpi_t M)
4307 -- Function: int gcry_mpi_gcd (gcry_mpi_t G, gcry_mpi_t A,
4310 Set G to the greatest common divisor of A and B. Return true if
4313 -- Function: int gcry_mpi_invm (gcry_mpi_t X, gcry_mpi_t A,
4316 Set X to the multiplicative inverse of A \bmod M. Return true if
4320 File: gcrypt.info, Node: Comparisons, Next: Bit manipulations, Prev: Calculations, Up: MPI library
4325 The next 2 functions are used to compare MPIs:
4327 -- Function: int gcry_mpi_cmp (const gcry_mpi_t U, const gcry_mpi_t V)
4329 Compare the multi-precision-integers number U and V returning 0 for
4330 equality, a positive value for U > V and a negative for U < V. If
4331 both numbers are opaque values (cf, gcry_mpi_set_opaque) the
4332 comparison is done by checking the bit sizes using memcmp. If only
4333 one number is an opaque value, the opaque value is less than the
4336 -- Function: int gcry_mpi_cmp_ui (const gcry_mpi_t U, unsigned long V)
4338 Compare the multi-precision-integers number U with the unsigned
4339 integer V returning 0 for equality, a positive value for U > V and
4340 a negative for U < V.
4342 -- Function: int gcry_mpi_is_neg (const gcry_mpi_t A)
4344 Return 1 if A is less than zero; return 0 if zero or positive.
4347 File: gcrypt.info, Node: Bit manipulations, Next: EC functions, Prev: Comparisons, Up: MPI library
4349 12.6 Bit manipulations
4350 ======================
4352 There are a couple of functions to get information on arbitrary bits in
4353 an MPI and to set or clear them:
4355 -- Function: unsigned int gcry_mpi_get_nbits (gcry_mpi_t A)
4357 Return the number of bits required to represent A.
4359 -- Function: int gcry_mpi_test_bit (gcry_mpi_t A, unsigned int N)
4361 Return true if bit number N (counting from 0) is set in A.
4363 -- Function: void gcry_mpi_set_bit (gcry_mpi_t A, unsigned int N)
4365 Set bit number N in A.
4367 -- Function: void gcry_mpi_clear_bit (gcry_mpi_t A, unsigned int N)
4369 Clear bit number N in A.
4371 -- Function: void gcry_mpi_set_highbit (gcry_mpi_t A, unsigned int N)
4373 Set bit number N in A and clear all bits greater than N.
4375 -- Function: void gcry_mpi_clear_highbit (gcry_mpi_t A, unsigned int N)
4377 Clear bit number N in A and all bits greater than N.
4379 -- Function: void gcry_mpi_rshift (gcry_mpi_t X, gcry_mpi_t A,
4382 Shift the value of A by N bits to the right and store the result in
4385 -- Function: void gcry_mpi_lshift (gcry_mpi_t X, gcry_mpi_t A,
4388 Shift the value of A by N bits to the left and store the result in
4392 File: gcrypt.info, Node: EC functions, Next: Miscellaneous, Prev: Bit manipulations, Up: MPI library
4397 Libgcrypt provides an API to access low level functions used by its
4398 elliptic curve implementation. These functions allow to implement
4399 elliptic curve methods for which no explicit support is available.
4401 -- Function: gcry_mpi_point_t gcry_mpi_point_new (unsigned int NBITS)
4403 Allocate a new point object, initialize it to 0, and allocate
4404 enough memory for a points of at least NBITS. This pre-allocation
4405 yields only a small performance win and is not really necessary
4406 because Libgcrypt automatically re-allocates the required memory.
4407 Using 0 for NBITS is usually the right thing to do.
4409 -- Function: void gcry_mpi_point_release (gcry_mpi_point_t POINT)
4411 Release POINT and free all associated resources. Passing 'NULL' is
4412 allowed and ignored.
4414 -- Function: gcry_mpi_point_t gcry_mpi_point_copy
4415 (gcry_mpi_point_t POINT)
4417 Allocate and return a new point object and initialize it with
4418 POINT. If POINT is NULL the function is identical to
4419 'gcry_mpi_point_new(0)'.
4421 -- Function: void gcry_mpi_point_get (gcry_mpi_t X, gcry_mpi_t Y,
4422 gcry_mpi_t Z, gcry_mpi_point_t POINT)
4424 Store the projective coordinates from POINT into the MPIs X, Y, and
4425 Z. If a coordinate is not required, 'NULL' may be used for X, Y,
4428 -- Function: void gcry_mpi_point_snatch_get (gcry_mpi_t X,
4429 gcry_mpi_t Y, gcry_mpi_t Z, gcry_mpi_point_t POINT)
4431 Store the projective coordinates from POINT into the MPIs X, Y, and
4432 Z. If a coordinate is not required, 'NULL' may be used for X, Y,
4433 or Z. The object POINT is then released. Using this function
4434 instead of 'gcry_mpi_point_get' and 'gcry_mpi_point_release' has
4435 the advantage of avoiding some extra memory allocations and copies.
4437 -- Function: gcry_mpi_point_t gcry_mpi_point_set (
4438 gcry_mpi_point_t POINT, gcry_mpi_t X, gcry_mpi_t Y,
4441 Store the projective coordinates from X, Y, and Z into POINT. If a
4442 coordinate is given as 'NULL', the value 0 is used. If 'NULL' is
4443 used for POINT a new point object is allocated and returned.
4444 Returns POINT or the newly allocated point object.
4446 -- Function: gcry_mpi_point_t gcry_mpi_point_snatch_set (
4447 gcry_mpi_point_t POINT, gcry_mpi_t X, gcry_mpi_t Y,
4450 Store the projective coordinates from X, Y, and Z into POINT. If a
4451 coordinate is given as 'NULL', the value 0 is used. If 'NULL' is
4452 used for POINT a new point object is allocated and returned. The
4453 MPIs X, Y, and Z are released. Using this function instead of
4454 'gcry_mpi_point_set' and 3 calls to 'gcry_mpi_release' has the
4455 advantage of avoiding some extra memory allocations and copies.
4456 Returns POINT or the newly allocated point object.
4458 -- Function: gpg_error_t gcry_mpi_ec_new (gcry_ctx_t *R_CTX,
4459 gcry_sexp_t KEYPARAM, const char *CURVENAME)
4461 Allocate a new context for elliptic curve operations. If KEYPARAM
4462 is given it specifies the parameters of the curve (*note
4463 ecc_keyparam::). If CURVENAME is given in addition to KEYPARAM and
4464 the key parameters do not include a named curve reference, the
4465 string CURVENAME is used to fill in missing parameters. If only
4466 CURVENAME is given, the context is initialized for this named
4469 If a parameter specifying a point (e.g. 'g' or 'q') is not found,
4470 the parser looks for a non-encoded point by appending '.x', '.y',
4471 and '.z' to the parameter name and looking them all up to create a
4472 point. A parameter with the suffix '.z' is optional and defaults
4475 On success the function returns 0 and stores the new context object
4476 at R_CTX; this object eventually needs to be released (*note
4477 gcry_ctx_release::). On error the function stores 'NULL' at R_CTX
4478 and returns an error code.
4480 -- Function: gcry_mpi_t gcry_mpi_ec_get_mpi ( const char *NAME,
4481 gcry_ctx_t CTX, int COPY)
4483 Return the MPI with NAME from the context CTX. If not found 'NULL'
4484 is returned. If the returned MPI may later be modified, it is
4485 suggested to pass '1' to COPY, so that the function guarantees that
4486 a modifiable copy of the MPI is returned. If '0' is used for COPY,
4487 this function may return a constant flagged MPI. In any case
4488 'gcry_mpi_release' needs to be called to release the result. For
4489 valid names *note ecc_keyparam::. If the public key 'q' is
4490 requested but only the private key 'd' is available, 'q' will be
4491 recomputed on the fly. If a point parameter is requested it is
4492 returned as an uncompressed encoded point unless these special
4495 Return an EdDSA style compressed point. This is only
4496 supported for Twisted Edwards curves.
4498 -- Function: gcry_mpi_point_t gcry_mpi_ec_get_point ( const char *NAME,
4499 gcry_ctx_t CTX, int COPY)
4501 Return the point with NAME from the context CTX. If not found
4502 'NULL' is returned. If the returned MPI may later be modified, it
4503 is suggested to pass '1' to COPY, so that the function guarantees
4504 that a modifiable copy of the MPI is returned. If '0' is used for
4505 COPY, this function may return a constant flagged point. In any
4506 case 'gcry_mpi_point_release' needs to be called to release the
4507 result. If the public key 'q' is requested but only the private
4508 key 'd' is available, 'q' will be recomputed on the fly.
4510 -- Function: gpg_error_t gcry_mpi_ec_set_mpi ( const char *NAME,
4511 gcry_mpi_t NEWVALUE, gcry_ctx_t CTX)
4513 Store the MPI NEWVALUE at NAME into the context CTX. On success
4514 '0' is returned; on error an error code. Valid names are the MPI
4515 parameters of an elliptic curve (*note ecc_keyparam::).
4517 -- Function: gpg_error_t gcry_mpi_ec_set_point ( const char *NAME,
4518 gcry_mpi_point_t NEWVALUE, gcry_ctx_t CTX)
4520 Store the point NEWVALUE at NAME into the context CTX. On success
4521 '0' is returned; on error an error code. Valid names are the point
4522 parameters of an elliptic curve (*note ecc_keyparam::).
4524 -- Function: gpg_err_code_t gcry_mpi_ec_decode_point (
4525 mpi_point_t RESULT, gcry_mpi_t VALUE, gcry_ctx_t CTX)
4527 Decode the point given as an MPI in VALUE and store at RESULT. To
4528 decide which encoding is used the function takes a context CTX
4529 which can be created with 'gcry_mpi_ec_new'. If 'NULL' is given
4530 for the context the function assumes a 0x04 prefixed uncompressed
4531 encoding. On error an error code is returned and RESULT might be
4534 -- Function: int gcry_mpi_ec_get_affine ( gcry_mpi_t X, gcry_mpi_t Y,
4535 gcry_mpi_point_t POINT, gcry_ctx_t CTX)
4537 Compute the affine coordinates from the projective coordinates in
4538 POINT and store them into X and Y. If one coordinate is not
4539 required, 'NULL' may be passed to X or Y. CTX is the context
4540 object which has been created using 'gcry_mpi_ec_new'. Returns 0
4541 on success or not 0 if POINT is at infinity.
4543 Note that you can use 'gcry_mpi_ec_set_point' with the value
4544 'GCRYMPI_CONST_ONE' for Z to convert affine coordinates back into
4545 projective coordinates.
4547 -- Function: void gcry_mpi_ec_dup ( gcry_mpi_point_t W,
4548 gcry_mpi_point_t U, gcry_ctx_t CTX)
4550 Double the point U of the elliptic curve described by CTX and store
4553 -- Function: void gcry_mpi_ec_add ( gcry_mpi_point_t W,
4554 gcry_mpi_point_t U, gcry_mpi_point_t V, gcry_ctx_t CTX)
4556 Add the points U and V of the elliptic curve described by CTX and
4557 store the result into W.
4559 -- Function: void gcry_mpi_ec_sub ( gcry_mpi_point_t W,
4560 gcry_mpi_point_t U, gcry_mpi_point_t V, gcry_ctx_t CTX)
4562 Subtracts the point V from the point U of the elliptic curve
4563 described by CTX and store the result into W. Only Twisted Edwards
4564 curves are supported for now.
4566 -- Function: void gcry_mpi_ec_mul ( gcry_mpi_point_t W, gcry_mpi_t N,
4567 gcry_mpi_point_t U, gcry_ctx_t CTX)
4569 Multiply the point U of the elliptic curve described by CTX by N
4570 and store the result into W.
4572 -- Function: int gcry_mpi_ec_curve_point ( gcry_mpi_point_t POINT,
4575 Return true if POINT is on the elliptic curve described by CTX.
4578 File: gcrypt.info, Node: Miscellaneous, Prev: EC functions, Up: MPI library
4583 An MPI data type is allowed to be "misused" to store an arbitrary value.
4584 Two functions implement this kludge:
4586 -- Function: gcry_mpi_t gcry_mpi_set_opaque (gcry_mpi_t A, void *P,
4589 Store NBITS of the value P points to in A and mark A as an opaque
4590 value (i.e. an value that can't be used for any math calculation
4591 and is only used to store an arbitrary bit pattern in A).
4592 Ownership of P is taken by this function and thus the user may not
4593 use dereference the passed value anymore. It is required that them
4594 memory referenced by P has been allocated in a way that 'gcry_free'
4595 is able to release it.
4597 WARNING: Never use an opaque MPI for actual math operations. The
4598 only valid functions are gcry_mpi_get_opaque and gcry_mpi_release.
4599 Use gcry_mpi_scan to convert a string of arbitrary bytes into an
4602 -- Function: gcry_mpi_t gcry_mpi_set_opaque_copy (gcry_mpi_t A,
4603 const void *P, unsigned int NBITS)
4605 Same as 'gcry_mpi_set_opaque' but ownership of P is not taken
4606 instead a copy of P is used.
4608 -- Function: void * gcry_mpi_get_opaque (gcry_mpi_t A,
4609 unsigned int *NBITS)
4611 Return a pointer to an opaque value stored in A and return its size
4612 in NBITS. Note that the returned pointer is still owned by A and
4613 that the function should never be used for an non-opaque MPI.
4615 Each MPI has an associated set of flags for special purposes. The
4616 currently defined flags are:
4618 'GCRYMPI_FLAG_SECURE'
4619 Setting this flag converts A into an MPI stored in "secure memory".
4620 Clearing this flag is not allowed.
4621 'GCRYMPI_FLAG_OPAQUE'
4622 This is an interanl flag, indicating the an opaque valuue and not
4623 an integer is stored. This is an read-only flag; it may not be set
4625 'GCRYMPI_FLAG_IMMUTABLE'
4626 If this flag is set, the MPI is marked as immutable. Setting or
4627 changing the value of that MPI is ignored and an error message is
4628 logged. The flag is sometimes useful for debugging.
4629 'GCRYMPI_FLAG_CONST'
4630 If this flag is set, the MPI is marked as a constant and as
4631 immutable Setting or changing the value of that MPI is ignored and
4632 an error message is logged. Such an MPI will never be deallocated
4633 and may thus be used without copying. Note that using
4634 gcry_mpi_copy will return a copy of that constant with this and the
4635 immutable flag cleared. A few commonly used constants are
4636 pre-defined and accessible using the macros 'GCRYMPI_CONST_ONE',
4637 'GCRYMPI_CONST_TWO', 'GCRYMPI_CONST_THREE', 'GCRYMPI_CONST_FOUR',
4638 and 'GCRYMPI_CONST_EIGHT'.
4639 'GCRYMPI_FLAG_USER1'
4640 'GCRYMPI_FLAG_USER2'
4641 'GCRYMPI_FLAG_USER3'
4642 'GCRYMPI_FLAG_USER4'
4643 These flags are reserved for use by the application.
4645 -- Function: void gcry_mpi_set_flag (gcry_mpi_t A,
4646 enum gcry_mpi_flag FLAG)
4648 Set the FLAG for the MPI A. The only allowed flags are
4649 'GCRYMPI_FLAG_SECURE', 'GCRYMPI_FLAG_IMMUTABLE', and
4650 'GCRYMPI_FLAG_CONST'.
4652 -- Function: void gcry_mpi_clear_flag (gcry_mpi_t A,
4653 enum gcry_mpi_flag FLAG)
4655 Clear FLAG for the multi-precision-integers A. The only allowed
4656 flag is 'GCRYMPI_FLAG_IMMUTABLE' but only if 'GCRYMPI_FLAG_CONST'
4657 is not set. If 'GCRYMPI_FLAG_CONST' is set, clearing
4658 'GCRYMPI_FLAG_IMMUTABLE' will simply be ignored.
4660 -- Function: int gcry_mpi_get_flag (gcry_mpi_t A,
4661 enum gcry_mpi_flag FLAG)
4663 Return true if FLAG is set for A.
4665 To put a random value into an MPI, the following convenience function
4668 -- Function: void gcry_mpi_randomize (gcry_mpi_t W, unsigned int NBITS,
4669 enum gcry_random_level LEVEL)
4671 Set the multi-precision-integers W to a random non-negative number
4672 of NBITS, using random data quality of level LEVEL. In case NBITS
4673 is not a multiple of a byte, NBITS is rounded up to the next byte
4674 boundary. When using a LEVEL of 'GCRY_WEAK_RANDOM' this function
4675 makes use of 'gcry_create_nonce'.
4678 File: gcrypt.info, Node: Prime numbers, Next: Utilities, Prev: MPI library, Up: Top
4685 * Generation:: Generation of new prime numbers.
4686 * Checking:: Checking if a given number is prime.
4689 File: gcrypt.info, Node: Generation, Next: Checking, Up: Prime numbers
4694 -- Function: gcry_error_t gcry_prime_generate (gcry_mpi_t
4695 *PRIME,unsigned int PRIME_BITS, unsigned int FACTOR_BITS,
4696 gcry_mpi_t **FACTORS, gcry_prime_check_func_t CB_FUNC, void
4697 *CB_ARG, gcry_random_level_t RANDOM_LEVEL, unsigned int FLAGS)
4699 Generate a new prime number of PRIME_BITS bits and store it in
4700 PRIME. If FACTOR_BITS is non-zero, one of the prime factors of
4701 (PRIME - 1) / 2 must be FACTOR_BITS bits long. If FACTORS is
4702 non-zero, allocate a new, 'NULL'-terminated array holding the prime
4703 factors and store it in FACTORS. FLAGS might be used to influence
4704 the prime number generation process.
4706 -- Function: gcry_error_t gcry_prime_group_generator (gcry_mpi_t *R_G,
4707 gcry_mpi_t PRIME, gcry_mpi_t *FACTORS, gcry_mpi_t START_G)
4709 Find a generator for PRIME where the factorization of (PRIME-1) is
4710 in the 'NULL' terminated array FACTORS. Return the generator as a
4711 newly allocated MPI in R_G. If START_G is not NULL, use this as
4712 the start for the search.
4714 -- Function: void gcry_prime_release_factors (gcry_mpi_t *FACTORS)
4716 Convenience function to release the FACTORS array.
4719 File: gcrypt.info, Node: Checking, Prev: Generation, Up: Prime numbers
4724 -- Function: gcry_error_t gcry_prime_check (gcry_mpi_t P, unsigned int
4727 Check whether the number P is prime. Returns zero in case P is
4728 indeed a prime, returns 'GPG_ERR_NO_PRIME' in case P is not a prime
4729 and a different error code in case something went horribly wrong.
4732 File: gcrypt.info, Node: Utilities, Next: Tools, Prev: Prime numbers, Up: Top
4739 * Memory allocation:: Functions related with memory allocation.
4740 * Context management:: Functions related with context management.
4741 * Buffer description:: A data type to describe buffers.
4742 * Config reporting:: How to return Libgcrypt's configuration.
4745 File: gcrypt.info, Node: Memory allocation, Next: Context management, Up: Utilities
4747 14.1 Memory allocation
4748 ======================
4750 -- Function: void * gcry_malloc (size_t N)
4752 This function tries to allocate N bytes of memory. On success it
4753 returns a pointer to the memory area, in an out-of-core condition,
4756 -- Function: void * gcry_malloc_secure (size_t N)
4757 Like 'gcry_malloc', but uses secure memory.
4759 -- Function: void * gcry_calloc (size_t N, size_t M)
4761 This function allocates a cleared block of memory (i.e.
4762 initialized with zero bytes) long enough to contain a vector of N
4763 elements, each of size M bytes. On success it returns a pointer to
4764 the memory block; in an out-of-core condition, it returns NULL.
4766 -- Function: void * gcry_calloc_secure (size_t N, size_t M)
4767 Like 'gcry_calloc', but uses secure memory.
4769 -- Function: void * gcry_realloc (void *P, size_t N)
4771 This function tries to resize the memory area pointed to by P to N
4772 bytes. On success it returns a pointer to the new memory area, in
4773 an out-of-core condition, it returns NULL. Depending on whether the
4774 memory pointed to by P is secure memory or not, gcry_realloc tries
4775 to use secure memory as well.
4777 -- Function: void gcry_free (void *P)
4778 Release the memory area pointed to by P.
4781 File: gcrypt.info, Node: Context management, Next: Buffer description, Prev: Memory allocation, Up: Utilities
4783 14.2 Context management
4784 =======================
4786 Some function make use of a context object. As of now there are only a
4787 few math functions. However, future versions of Libgcrypt may make more
4788 use of this context object.
4790 -- Data type: gcry_ctx_t
4791 This type is used to refer to the general purpose context object.
4793 -- Function: void gcry_ctx_release (gcry_ctx_t CTX)
4794 Release the context object CTX and all associated resources. A
4795 'NULL' passed as CTX is ignored.
4798 File: gcrypt.info, Node: Buffer description, Next: Config reporting, Prev: Context management, Up: Utilities
4800 14.3 Buffer description
4801 =======================
4803 To help hashing non-contiguous areas of memory a general purpose data
4806 -- Data type: gcry_buffer_t
4807 This type is a structure to describe a buffer. The user should
4808 make sure that this structure is initialized to zero. The
4809 available fields of this structure are:
4812 This is either 0 for no information available or indicates the
4813 allocated length of the buffer.
4815 This is the offset into the buffer.
4817 This is the valid length of the buffer starting at '.off'.
4819 This is the address of the buffer.
4822 File: gcrypt.info, Node: Config reporting, Prev: Buffer description, Up: Utilities
4824 14.4 How to return Libgcrypt's configuration.
4825 =============================================
4827 Although 'GCRYCTL_PRINT_CONFIG' can be used to print configuration
4828 options, it is sometimes necessary to check them in a program. This can
4829 be accomplished by using this function:
4831 -- Function: char * gcry_get_config (int MODE, const char *WHAT)
4833 This function returns a malloced string with colon delimited
4834 configure options. With a value of 0 for MODE this string
4835 resembles the output of 'GCRYCTL_PRINT_CONFIG'. However, if WHAT
4836 is not NULL, only the line where the first field (e.g. "cpu-arch")
4837 matches WHAT is returned.
4839 Other values than 0 for MODE are not defined. The caller shall
4840 free the string using 'gcry_free'. On error NULL is returned and
4841 ERRNO is set; if a value for WHAT is unknow ERRNO will be set to 0.
4844 File: gcrypt.info, Node: Tools, Next: Configuration, Prev: Utilities, Up: Top
4851 * hmac256:: A standalone HMAC-SHA-256 implementation
4854 File: gcrypt.info, Node: hmac256, Up: Tools
4856 15.1 A HMAC-SHA-256 tool
4857 ========================
4859 This is a standalone HMAC-SHA-256 implementation used to compute an
4860 HMAC-SHA-256 message authentication code. The tool has originally been
4861 developed as a second implementation for Libgcrypt to allow comparing
4862 against the primary implementation and to be used for internal
4863 consistency checks. It should not be used for sensitive data because no
4864 mechanisms to clear the stack etc are used.
4866 The code has been written in a highly portable manner and requires
4867 only a few standard definitions to be provided in a config.h file.
4869 'hmac256' is commonly invoked as
4871 hmac256 "This is my key" foo.txt
4873 This compute the MAC on the file 'foo.txt' using the key given on the
4876 'hmac256' understands these options:
4879 Print the MAC as a binary string. The default is to print the MAC
4880 encoded has lower case hex digits.
4883 Print version of the program and exit.
4886 File: gcrypt.info, Node: Configuration, Next: Architecture, Prev: Tools, Up: Top
4888 16 Configuration files and environment variables
4889 ************************************************
4891 This chapter describes which files and environment variables can be used
4892 to change the behaviour of Libgcrypt.
4894 The environment variables considered by Libgcrypt are:
4897 By setting this variable to any value a different algorithm for
4898 modular reduction is used for ECC.
4900 'GCRYPT_RNDUNIX_DBG'
4901 'GCRYPT_RNDUNIX_DBGALL'
4902 These two environment variables are used to enable debug output for
4903 the rndunix entropy gatherer, which is used on systems lacking a
4904 /dev/random device. The value of 'GCRYPT_RNDUNIX_DBG' is a file
4905 name or '-' for stdout. Debug output is the written to this file.
4906 By setting 'GCRYPT_RNDUNIX_DBGALL' to any value the debug output
4907 will be more verbose.
4909 'GCRYPT_RNDW32_NOPERF'
4910 Setting this environment variable on Windows to any value disables
4911 the use of performance data ('HKEY_PERFORMANCE_DATA') as source for
4912 entropy. On some older Windows systems this could help to speed up
4913 the creation of random numbers but also decreases the amount of
4914 data used to init the random number generator.
4917 Setting the value of this variable to a positive integer logs
4918 information about the Windows entropy gatherer using the standard
4922 This is used to locate the socket to connect to the EGD random
4923 daemon. The EGD can be used on system without a /dev/random to
4924 speed up the random number generator. It is not needed on the
4925 majority of today's operating systems and support for EGD requires
4926 the use of a configure option at build time.
4928 The files which Libgcrypt uses to retrieve system information and the
4929 files which can be created by the user to modify Libgcrypt's behavior
4932 '/etc/gcrypt/hwf.deny'
4933 This file can be used to disable the use of hardware based
4934 optimizations, *note hardware features::.
4936 '/etc/gcrypt/random.conf'
4937 This file can be used to globally change parameters of the random
4938 generator. The file is a simple text file where empty lines and
4939 lines with the first non white-space character being '#' are
4940 ignored. Supported options are
4943 Disable the use of the jitter based entropy generator.
4946 Always use the non-blocking /dev/urandom or the respective
4947 system call instead of the blocking /dev/random. If Libgcrypt
4948 is used early in the boot process of the system, this option
4949 should only be used if the system also supports the getrandom
4952 '/etc/gcrypt/fips_enabled'
4953 '/proc/sys/crypto/fips_enabled'
4954 On Linux these files are used to enable FIPS mode, *note enabling
4959 On Linux running on the ARM architecture, these files are used to
4960 read hardware capabilities of the CPU.
4963 File: gcrypt.info, Node: Architecture, Next: Self-Tests, Prev: Configuration, Up: Top
4968 This chapter describes the internal architecture of Libgcrypt.
4970 Libgcrypt is a function library written in ISO C-90. Any compliant
4971 compiler should be able to build Libgcrypt as long as the target is
4972 either a POSIX platform or compatible to the API used by Windows NT.
4973 Provisions have been take so that the library can be directly used from
4974 C++ applications; however building with a C++ compiler is not supported.
4976 Building Libgcrypt is done by using the common './configure && make'
4977 approach. The configure command is included in the source distribution
4978 and as a portable shell script it works on any Unix-alike system. The
4979 result of running the configure script are a C header file ('config.h'),
4980 customized Makefiles, the setup of symbolic links and a few other
4981 things. After that the make tool builds and optionally installs the
4982 library and the documentation. See the files 'INSTALL' and 'README' in
4983 the source distribution on how to do this.
4985 Libgcrypt is developed using a Subversion(1) repository. Although
4986 all released versions are tagged in this repository, they should not be
4987 used to build production versions of Libgcrypt. Instead released
4988 tarballs should be used. These tarballs are available from several
4989 places with the master copy at 'ftp://ftp.gnupg.org/gcrypt/libgcrypt/'.
4990 Announcements of new releases are posted to the
4991 'gnupg-announce@gnupg.org' mailing list(2).
4993 \0\b[image src="libgcrypt-modules.png" alt="Libgcrypt subsystems"
\0\b]
4995 Figure 17.1: Libgcrypt subsystems
4997 Libgcrypt consists of several subsystems (*note Figure 17.1:
4998 fig:subsystems.) and all these subsystems provide a public API; this
4999 includes the helper subsystems like the one for S-expressions. The API
5000 style depends on the subsystem; in general an open-use-close approach is
5001 implemented. The open returns a handle to a context used for all
5002 further operations on this handle, several functions may then be used on
5003 this handle and a final close function releases all resources associated
5008 * Public-Key Subsystem Architecture:: About public keys.
5009 * Symmetric Encryption Subsystem Architecture:: About standard ciphers.
5010 * Hashing and MACing Subsystem Architecture:: About hashing.
5011 * Multi-Precision-Integer Subsystem Architecture:: About big integers.
5012 * Prime-Number-Generator Subsystem Architecture:: About prime numbers.
5013 * Random-Number Subsystem Architecture:: About random stuff.
5015 ---------- Footnotes ----------
5017 (1) A version control system available for many platforms
5019 (2) See <http://www.gnupg.org/documentation/mailing-lists.en.html>
5023 File: gcrypt.info, Node: Public-Key Subsystem Architecture, Next: Symmetric Encryption Subsystem Architecture, Up: Architecture
5025 17.1 Public-Key Architecture
5026 ============================
5028 Because public key cryptography is almost always used to process small
5029 amounts of data (hash values or session keys), the interface is not
5030 implemented using the open-use-close paradigm, but with single
5031 self-contained functions. Due to the wide variety of parameters
5032 required by different algorithms S-expressions, as flexible way to
5033 convey these parameters, are used. There is a set of helper functions
5034 to work with these S-expressions.
5036 Aside of functions to register new algorithms, map algorithms names
5037 to algorithms identifiers and to lookup properties of a key, the
5038 following main functions are available:
5041 Encrypt data using a public key.
5044 Decrypt data using a private key.
5047 Sign data using a private key.
5050 Verify that a signature matches the data.
5053 Perform a consistency over a public or private key.
5056 Create a new public/private key pair.
5058 All these functions lookup the module implementing the algorithm and
5059 pass the actual work to that module. The parsing of the S-expression
5060 input and the construction of S-expression for the return values is done
5061 by the high level code ('cipher/pubkey.c'). Thus the internal interface
5062 between the algorithm modules and the high level functions passes data
5065 By default Libgcrypt uses a blinding technique for RSA decryption to
5066 mitigate real world timing attacks over a network: Instead of using the
5067 RSA decryption directly, a blinded value y = x r^{e} \bmod n is
5068 decrypted and the unblinded value x' = y' r^{-1} \bmod n returned. The
5069 blinding value r is a random value with the size of the modulus n and
5070 generated with 'GCRY_WEAK_RANDOM' random level.
5072 The algorithm used for RSA and DSA key generation depends on whether
5073 Libgcrypt is operated in standard or in FIPS mode. In standard mode an
5074 algorithm based on the Lim-Lee prime number generator is used. In FIPS
5075 mode RSA keys are generated as specified in ANSI X9.31 (1998) and DSA
5076 keys as specified in FIPS 186-2.
5079 File: gcrypt.info, Node: Symmetric Encryption Subsystem Architecture, Next: Hashing and MACing Subsystem Architecture, Prev: Public-Key Subsystem Architecture, Up: Architecture
5081 17.2 Symmetric Encryption Subsystem Architecture
5082 ================================================
5084 The interface to work with symmetric encryption algorithms is made up of
5085 functions from the 'gcry_cipher_' name space. The implementation
5086 follows the open-use-close paradigm and uses registered algorithm
5087 modules for the actual work. Unless a module implements optimized
5088 cipher mode implementations, the high level code ('cipher/cipher.c')
5089 implements the modes and calls the core algorithm functions to process
5092 The most important functions are:
5095 Create a new instance to encrypt or decrypt using a specified
5099 Release an instance.
5101 'gcry_cipher_setkey'
5102 Set a key to be used for encryption or decryption.
5105 Set an initialization vector to be used for encryption or
5108 'gcry_cipher_encrypt'
5109 'gcry_cipher_decrypt'
5110 Encrypt or decrypt data. These functions may be called with
5111 arbitrary amounts of data and as often as needed to encrypt or
5114 There are also functions to query properties of algorithms or
5115 context, like block length, key length, map names or to enable features
5116 like padding methods.
5119 File: gcrypt.info, Node: Hashing and MACing Subsystem Architecture, Next: Multi-Precision-Integer Subsystem Architecture, Prev: Symmetric Encryption Subsystem Architecture, Up: Architecture
5121 17.3 Hashing and MACing Subsystem Architecture
5122 ==============================================
5124 The interface to work with message digests and CRC algorithms is made up
5125 of functions from the 'gcry_md_' name space. The implementation follows
5126 the open-use-close paradigm and uses registered algorithm modules for
5127 the actual work. Although CRC algorithms are not considered
5128 cryptographic hash algorithms, they share enough properties so that it
5129 makes sense to handle them in the same way. It is possible to use
5130 several algorithms at once with one context and thus compute them all on
5133 The most important functions are:
5136 Create a new message digest instance and optionally enable one
5137 algorithm. A flag may be used to turn the message digest algorithm
5138 into a HMAC algorithm.
5141 Enable an additional algorithm for the instance.
5144 Set the key for the MAC.
5147 Pass more data for computing the message digest to an instance.
5150 Buffered version of 'gcry_md_write' implemented as a macro.
5153 Finalize the computation of the message digest or HMAC and return
5159 'gcry_md_hash_buffer'
5160 Convenience function to directly compute a message digest over a
5161 memory buffer without the need to create an instance first.
5163 There are also functions to query properties of algorithms or the
5164 instance, like enabled algorithms, digest length, map algorithm names.
5165 it is also possible to reset an instance or to copy the current state of
5166 an instance at any time. Debug functions to write the hashed data to
5167 files are available as well.
5170 File: gcrypt.info, Node: Multi-Precision-Integer Subsystem Architecture, Next: Prime-Number-Generator Subsystem Architecture, Prev: Hashing and MACing Subsystem Architecture, Up: Architecture
5172 17.4 Multi-Precision-Integer Subsystem Architecture
5173 ===================================================
5175 The implementation of Libgcrypt's big integer computation code is based
5176 on an old release of GNU Multi-Precision Library (GMP). The decision not
5177 to use the GMP library directly was due to stalled development at that
5178 time and due to security requirements which could not be provided by the
5179 code in GMP. As GMP does, Libgcrypt provides high performance assembler
5180 implementations of low level code for several CPUS to gain much better
5181 performance than with a generic C implementation.
5183 Major features of Libgcrypt's multi-precision-integer code compared to
5186 * Avoidance of stack based allocations to allow protection against
5187 swapping out of sensitive data and for easy zeroing of sensitive
5188 intermediate results.
5190 * Optional use of secure memory and tracking of its use so that
5191 results are also put into secure memory.
5193 * MPIs are identified by a handle (implemented as a pointer) to give
5194 better control over allocations and to augment them with extra
5195 properties like opaque data.
5197 * Removal of unnecessary code to reduce complexity.
5199 * Functions specialized for public key cryptography.
5202 File: gcrypt.info, Node: Prime-Number-Generator Subsystem Architecture, Next: Random-Number Subsystem Architecture, Prev: Multi-Precision-Integer Subsystem Architecture, Up: Architecture
5204 17.5 Prime-Number-Generator Subsystem Architecture
5205 ==================================================
5207 Libgcrypt provides an interface to its prime number generator. These
5208 functions make use of the internal prime number generator which is
5209 required for the generation for public key key pairs. The plain prime
5210 checking function is exported as well.
5212 The generation of random prime numbers is based on the Lim and Lee
5213 algorithm to create practically save primes.(1) This algorithm creates
5214 a pool of smaller primes, select a few of them to create candidate
5215 primes of the form 2 * p_0 * p_1 * ... * p_n + 1, tests the candidate
5216 for primality and permutates the pool until a prime has been found. It
5217 is possible to clamp one of the small primes to a certain size to help
5218 DSA style algorithms. Because most of the small primes in the pool are
5219 not used for the resulting prime number, they are saved for later use
5220 (see 'save_pool_prime' and 'get_pool_prime' in 'cipher/primegen.c').
5221 The prime generator optionally supports the finding of an appropriate
5224 The primality test works in three steps:
5226 1. The standard sieve algorithm using the primes up to 4999 is used as
5227 a quick first check.
5229 2. A Fermat test filters out almost all non-primes.
5231 3. A 5 round Rabin-Miller test is finally used. The first round uses
5232 a witness of 2, whereas the next rounds use a random witness.
5234 To support the generation of RSA and DSA keys in FIPS mode according
5235 to X9.31 and FIPS 186-2, Libgcrypt implements two additional prime
5236 generation functions: '_gcry_derive_x931_prime' and
5237 '_gcry_generate_fips186_2_prime'. These functions are internal and not
5238 available through the public API.
5240 ---------- Footnotes ----------
5242 (1) Chae Hoon Lim and Pil Joong Lee. A key recovery attack on
5243 discrete log-based schemes using a prime order subgroup. In Burton S.
5244 Kaliski Jr., editor, Advances in Cryptology: Crypto '97, pages 249Â-263,
5245 Berlin / Heidelberg / New York, 1997. Springer-Verlag. Described on
5249 File: gcrypt.info, Node: Random-Number Subsystem Architecture, Prev: Prime-Number-Generator Subsystem Architecture, Up: Architecture
5251 17.6 Random-Number Subsystem Architecture
5252 =========================================
5254 Libgcrypt provides 3 levels or random quality: The level
5255 'GCRY_VERY_STRONG_RANDOM' usually used for key generation, the level
5256 'GCRY_STRONG_RANDOM' for all other strong random requirements and the
5257 function 'gcry_create_nonce' which is used for weaker usages like
5258 nonces. There is also a level 'GCRY_WEAK_RANDOM' which in general maps
5259 to 'GCRY_STRONG_RANDOM' except when used with the function
5260 'gcry_mpi_randomize', where it randomizes an multi-precision-integer
5261 using the 'gcry_create_nonce' function.
5263 There are two distinct random generators available:
5265 * The Continuously Seeded Pseudo Random Number Generator (CSPRNG),
5266 which is based on the classic GnuPG derived big pool
5267 implementation. Implemented in 'random/random-csprng.c' and used
5269 * A FIPS approved ANSI X9.31 PRNG using AES with a 128 bit key.
5270 Implemented in 'random/random-fips.c' and used if Libgcrypt is in
5273 Both generators make use of so-called entropy gathering modules:
5276 Uses the operating system provided '/dev/random' and '/dev/urandom'
5277 devices. The '/dev/gcrypt/random.conf' config option
5278 'only-urandom' can be used to inhibit the use of the blocking
5279 '/dev/random' device.
5282 Runs several operating system commands to collect entropy from
5283 sources like virtual machine and process statistics. It is a kind
5284 of poor-man's '/dev/random' implementation. It is not available in
5288 Uses the operating system provided Entropy Gathering Daemon (EGD).
5289 The EGD basically uses the same algorithms as rndunix does.
5290 However as a system daemon it keeps on running and thus can serve
5291 several processes requiring entropy input and does not waste
5292 collected entropy if the application does not need all the
5293 collected entropy. It is not available in FIPS mode.
5296 Targeted for the Microsoft Windows OS. It uses certain properties
5297 of that system and is the only gathering module available for that
5301 Extra module to collect additional entropy by utilizing a hardware
5302 random number generator. As of now the supported hardware RNG is
5303 the Padlock engine of VIA (Centaur) CPUs and x86 CPUs with the
5304 RDRAND instruction. It is not available in FIPS mode.
5307 Extra module to collect additional entropy using a CPU jitter based
5308 approach. This is only used on X86 hardware where the RDTSC opcode
5309 is available. The '/dev/gcrypt/random.conf' config option
5310 'disable-jent' can be used to inhibit the use of this module.
5314 * CSPRNG Description:: Description of the CSPRNG.
5315 * FIPS PRNG Description:: Description of the FIPS X9.31 PRNG.
5318 File: gcrypt.info, Node: CSPRNG Description, Next: FIPS PRNG Description, Up: Random-Number Subsystem Architecture
5320 17.6.1 Description of the CSPRNG
5321 --------------------------------
5323 This random number generator is loosely modelled after the one described
5324 in Peter Gutmann's paper: "Software Generation of Practically Strong
5327 A pool of 600 bytes is used and mixed using the core SHA-1 hash
5328 transform function. Several extra features are used to make the robust
5329 against a wide variety of attacks and to protect against failures of
5330 subsystems. The state of the generator may be saved to a file and
5331 initially seed form a file.
5333 Depending on how Libgcrypt was build the generator is able to select
5334 the best working entropy gathering module. It makes use of the slow and
5335 fast collection methods and requires the pool to initially seeded form
5336 the slow gatherer or a seed file. An entropy estimation is used to mix
5337 in enough data from the gather modules before returning the actual
5338 random output. Process fork detection and protection is implemented.
5340 The implementation of the nonce generator (for 'gcry_create_nonce')
5341 is a straightforward repeated hash design: A 28 byte buffer is initially
5342 seeded with the PID and the time in seconds in the first 20 bytes and
5343 with 8 bytes of random taken from the 'GCRY_STRONG_RANDOM' generator.
5344 Random numbers are then created by hashing all the 28 bytes with SHA-1
5345 and saving that again in the first 20 bytes. The hash is also returned
5348 ---------- Footnotes ----------
5350 (1) Also described in chapter 6 of his book "Cryptographic Security
5351 Architecture", New York, 2004, ISBN 0-387-95387-6.
5354 File: gcrypt.info, Node: FIPS PRNG Description, Prev: CSPRNG Description, Up: Random-Number Subsystem Architecture
5356 17.6.2 Description of the FIPS X9.31 PRNG
5357 -----------------------------------------
5359 The core of this deterministic random number generator is implemented
5360 according to the document "NIST-Recommended Random Number Generator
5361 Based on ANSI X9.31 Appendix A.2.4 Using the 3-Key Triple DES and AES
5362 Algorithms", dated 2005-01-31. This implementation uses the AES
5365 The generator is based on contexts to utilize the same core functions
5366 for all random levels as required by the high-level interface. All
5367 random generators return their data in 128 bit blocks. If the caller
5368 requests less bits, the extra bits are not used. The key for each
5369 generator is only set once at the first time a generator context is
5370 used. The seed value is set along with the key and again after 1000
5373 On Unix like systems the 'GCRY_VERY_STRONG_RANDOM' and
5374 'GCRY_STRONG_RANDOM' generators are keyed and seeded using the rndlinux
5375 module with the '/dev/random' device. Thus these generators may block
5376 until the OS kernel has collected enough entropy. When used with
5377 Microsoft Windows the rndw32 module is used instead.
5379 The generator used for 'gcry_create_nonce' is keyed and seeded from
5380 the 'GCRY_STRONG_RANDOM' generator. Thus is may also block if the
5381 'GCRY_STRONG_RANDOM' generator has not yet been used before and thus
5382 gets initialized on the first use by 'gcry_create_nonce'. This special
5383 treatment is justified by the weaker requirements for a nonce generator
5384 and to save precious kernel entropy for use by the "real" random
5387 A self-test facility uses a separate context to check the
5388 functionality of the core X9.31 functions using a known answers test.
5389 During runtime each output block is compared to the previous one to
5390 detect a stuck generator.
5392 The DT value for the generator is made up of the current time down to
5393 microseconds (if available) and a free running 64 bit counter. When
5394 used with the test context the DT value is taken from the context and
5395 incremented on each use.
5398 File: gcrypt.info, Node: Self-Tests, Next: FIPS Mode, Prev: Architecture, Up: Top
5400 Appendix A Description of the Self-Tests
5401 ****************************************
5403 In addition to the build time regression test suite, Libgcrypt
5404 implements self-tests to be performed at runtime. Which self-tests are
5405 actually used depends on the mode Libgcrypt is used in. In standard
5406 mode a limited set of self-tests is run at the time an algorithm is
5407 first used. Note that not all algorithms feature a self-test in
5408 standard mode. The 'GCRYCTL_SELFTEST' control command may be used to
5409 run all implemented self-tests at any time; this will even run more
5410 tests than those run in FIPS mode.
5412 If any of the self-tests fails, the library immediately returns an
5413 error code to the caller. If Libgcrypt is in FIPS mode the self-tests
5414 will be performed within the "Self-Test" state and any failure puts the
5415 library into the "Error" state.
5420 Power-up tests are only performed if Libgcrypt is in FIPS mode.
5422 A.1.1 Symmetric Cipher Algorithm Power-Up Tests
5423 -----------------------------------------------
5425 The following symmetric encryption algorithm tests are run during
5429 To test the 3DES 3-key EDE encryption in ECB mode these tests are
5431 1. A known answer test is run on a 64 bit test vector processed
5432 by 64 rounds of Single-DES block encryption and decryption
5433 using a key changed with each round.
5434 2. A known answer test is run on a 64 bit test vector processed
5435 by 16 rounds of 2-key and 3-key Triple-DES block encryption
5436 and decryptions using a key changed with each round.
5437 3. 10 known answer tests using 3-key Triple-DES EDE encryption,
5438 comparing the ciphertext to the known value, then running a
5439 decryption and comparing it to the initial plaintext.
5440 ('cipher/des.c:selftest')
5443 A known answer tests is run using one test vector and one test key
5444 with AES in ECB mode. ('cipher/rijndael.c:selftest_basic_128')
5447 A known answer tests is run using one test vector and one test key
5448 with AES in ECB mode. ('cipher/rijndael.c:selftest_basic_192')
5451 A known answer tests is run using one test vector and one test key
5452 with AES in ECB mode. ('cipher/rijndael.c:selftest_basic_256')
5454 A.1.2 Hash Algorithm Power-Up Tests
5455 -----------------------------------
5457 The following hash algorithm tests are run during power-up:
5460 A known answer test using the string '"abc"' is run.
5461 ('cipher/sha1.c:selftests_sha1')
5463 A known answer test using the string '"abc"' is run.
5464 ('cipher/sha256.c:selftests_sha224')
5466 A known answer test using the string '"abc"' is run.
5467 ('cipher/sha256.c:selftests_sha256')
5469 A known answer test using the string '"abc"' is run.
5470 ('cipher/sha512.c:selftests_sha384')
5472 A known answer test using the string '"abc"' is run.
5473 ('cipher/sha512.c:selftests_sha512')
5475 A.1.3 MAC Algorithm Power-Up Tests
5476 ----------------------------------
5478 The following MAC algorithm tests are run during power-up:
5481 A known answer test using 9 byte of data and a 64 byte key is run.
5482 ('cipher/hmac-tests.c:selftests_sha1')
5484 A known answer test using 28 byte of data and a 4 byte key is run.
5485 ('cipher/hmac-tests.c:selftests_sha224')
5487 A known answer test using 28 byte of data and a 4 byte key is run.
5488 ('cipher/hmac-tests.c:selftests_sha256')
5490 A known answer test using 28 byte of data and a 4 byte key is run.
5491 ('cipher/hmac-tests.c:selftests_sha384')
5493 A known answer test using 28 byte of data and a 4 byte key is run.
5494 ('cipher/hmac-tests.c:selftests_sha512')
5496 A.1.4 Random Number Power-Up Test
5497 ---------------------------------
5499 The DRNG is tested during power-up this way:
5501 1. Requesting one block of random using the public interface to check
5502 general working and the duplicated block detection.
5503 2. 3 know answer tests using pre-defined keys, seed and initial DT
5504 values. For each test 3 blocks of 16 bytes are requested and
5505 compared to the expected result. The DT value is incremented for
5508 A.1.5 Public Key Algorithm Power-Up Tests
5509 -----------------------------------------
5511 The public key algorithms are tested during power-up:
5514 A pre-defined 1024 bit RSA key is used and these tests are run in
5516 1. Conversion of S-expression to internal format.
5517 ('cipher/rsa.c:selftests_rsa')
5518 2. Private key consistency check. ('cipher/rsa.c:selftests_rsa')
5519 3. A pre-defined 20 byte value is signed with PKCS#1 padding for
5520 SHA-1. The result is verified using the public key against
5521 the original data and against modified data.
5522 ('cipher/rsa.c:selftest_sign_1024')
5523 4. A 1000 bit random value is encrypted and checked that it does
5524 not match the original random value. The encrypted result is
5525 then decrypted and checked that it matches the original random
5526 value. ('cipher/rsa.c:selftest_encr_1024')
5529 A pre-defined 1024 bit DSA key is used and these tests are run in
5531 1. Conversion of S-expression to internal format.
5532 ('cipher/dsa.c:selftests_dsa')
5533 2. Private key consistency check. ('cipher/dsa.c:selftests_dsa')
5534 3. A pre-defined 20 byte value is signed with PKCS#1 padding for
5535 SHA-1. The result is verified using the public key against
5536 the original data and against modified data.
5537 ('cipher/dsa.c:selftest_sign_1024')
5539 A.1.6 Integrity Power-Up Tests
5540 ------------------------------
5542 The integrity of the Libgcrypt is tested during power-up but only if
5543 checking has been enabled at build time. The check works by computing a
5544 HMAC SHA-256 checksum over the file used to load Libgcrypt into memory.
5545 That checksum is compared against a checksum stored in a file of the
5546 same name but with a single dot as a prefix and a suffix of '.hmac'.
5548 A.1.7 Critical Functions Power-Up Tests
5549 ---------------------------------------
5551 The 3DES weak key detection is tested during power-up by calling the
5552 detection function with keys taken from a table listening all weak keys.
5553 The table itself is protected using a SHA-1 hash.
5554 ('cipher/des.c:selftest')
5556 A.2 Conditional Tests
5557 =====================
5559 The conditional tests are performed if a certain condition is met. This
5560 may occur at any time; the library does not necessary enter the
5561 "Self-Test" state to run these tests but will transit to the "Error"
5562 state if a test failed.
5564 A.2.1 Key-Pair Generation Tests
5565 -------------------------------
5567 After an asymmetric key-pair has been generated, Libgcrypt runs a
5568 pair-wise consistency tests on the generated key. On failure the
5569 generated key is not used, an error code is returned and, if in FIPS
5570 mode, the library is put into the "Error" state.
5573 The test uses a random number 64 bits less the size of the modulus
5574 as plaintext and runs an encryption and decryption operation in
5575 turn. The encrypted value is checked to not match the plaintext
5576 and the result of the decryption is checked to match the plaintext.
5578 A new random number of the same size is generated, signed and
5579 verified to test the correctness of the signing operation. As a
5580 second signing test, the signature is modified by incrementing its
5581 value and then verified with the expected result that the
5582 verification fails. ('cipher/rsa.c:test_keys')
5584 The test uses a random number of the size of the Q parameter to
5585 create a signature and then checks that the signature verifies. As
5586 a second signing test, the data is modified by incrementing its
5587 value and then verified against the signature with the expected
5588 result that the verification fails. ('cipher/dsa.c:test_keys')
5590 A.2.2 Software Load Tests
5591 -------------------------
5593 No code is loaded at runtime.
5595 A.2.3 Manual Key Entry Tests
5596 ----------------------------
5598 A manual key entry feature is not implemented in Libgcrypt.
5600 A.2.4 Continuous RNG Tests
5601 --------------------------
5603 The continuous random number test is only used in FIPS mode. The RNG
5604 generates blocks of 128 bit size; the first block generated per context
5605 is saved in the context and another block is generated to be returned to
5606 the caller. Each block is compared against the saved block and then
5607 stored in the context. If a duplicated block is detected an error is
5608 signaled and the library is put into the "Fatal-Error" state.
5609 ('random/random-fips.c:x931_aes_driver')
5611 A.3 Application Requested Tests
5612 ===============================
5614 The application may requests tests at any time by means of the
5615 'GCRYCTL_SELFTEST' control command. Note that using these tests is not
5616 FIPS conform: Although Libgcrypt rejects all application requests for
5617 services while running self-tests, it does not ensure that no other
5618 operations of Libgcrypt are still being executed. Thus, in FIPS mode an
5619 application requesting self-tests needs to power-cycle Libgcrypt
5622 When self-tests are requested, Libgcrypt runs all the tests it does
5623 during power-up as well as a few extra checks as described below.
5625 A.3.1 Symmetric Cipher Algorithm Tests
5626 --------------------------------------
5628 The following symmetric encryption algorithm tests are run in addition
5629 to the power-up tests:
5632 A known answer tests with test vectors taken from NIST SP800-38a
5633 and using the high level functions is run for block modes CFB and
5636 A.3.2 Hash Algorithm Tests
5637 --------------------------
5639 The following hash algorithm tests are run in addition to the power-up
5645 1. A known answer test using a 56 byte string is run.
5646 2. A known answer test using a string of one million letters "a"
5648 ('cipher/sha1.c:selftests_sha1',
5649 'cipher/sha256.c:selftests_sha224',
5650 'cipher/sha256.c:selftests_sha256')
5653 1. A known answer test using a 112 byte string is run.
5654 2. A known answer test using a string of one million letters "a"
5656 ('cipher/sha512.c:selftests_sha384',
5657 'cipher/sha512.c:selftests_sha512')
5659 A.3.3 MAC Algorithm Tests
5660 -------------------------
5662 The following MAC algorithm tests are run in addition to the power-up
5666 1. A known answer test using 9 byte of data and a 20 byte key is
5668 2. A known answer test using 9 byte of data and a 100 byte key is
5670 3. A known answer test using 9 byte of data and a 49 byte key is
5672 ('cipher/hmac-tests.c:selftests_sha1')
5677 1. A known answer test using 9 byte of data and a 20 byte key is
5679 2. A known answer test using 50 byte of data and a 20 byte key is
5681 3. A known answer test using 50 byte of data and a 26 byte key is
5683 4. A known answer test using 54 byte of data and a 131 byte key
5685 5. A known answer test using 152 byte of data and a 131 byte key
5687 ('cipher/hmac-tests.c:selftests_sha224',
5688 'cipher/hmac-tests.c:selftests_sha256',
5689 'cipher/hmac-tests.c:selftests_sha384',
5690 'cipher/hmac-tests.c:selftests_sha512')
5693 File: gcrypt.info, Node: FIPS Mode, Next: Library Copying, Prev: Self-Tests, Up: Top
5695 Appendix B Description of the FIPS Mode
5696 ***************************************
5698 This appendix gives detailed information pertaining to the FIPS mode.
5699 In particular, the changes to the standard mode and the finite state
5700 machine are described. The self-tests required in this mode are
5701 described in the appendix on self-tests.
5703 B.1 Restrictions in FIPS Mode
5704 =============================
5706 If Libgcrypt is used in FIPS mode these restrictions are effective:
5708 * The cryptographic algorithms are restricted to this list:
5711 3 key EDE Triple-DES symmetric encryption.
5713 AES 128 bit symmetric encryption.
5715 AES 192 bit symmetric encryption.
5717 AES 256 bit symmetric encryption.
5719 SHA-1 message digest.
5721 SHA-224 message digest.
5723 SHA-256 message digest.
5725 SHA-384 message digest.
5727 SHA-512 message digest.
5728 GCRY_MD_SHA1,GCRY_MD_FLAG_HMAC
5729 HMAC using a SHA-1 message digest.
5730 GCRY_MD_SHA224,GCRY_MD_FLAG_HMAC
5731 HMAC using a SHA-224 message digest.
5732 GCRY_MD_SHA256,GCRY_MD_FLAG_HMAC
5733 HMAC using a SHA-256 message digest.
5734 GCRY_MD_SHA384,GCRY_MD_FLAG_HMAC
5735 HMAC using a SHA-384 message digest.
5736 GCRY_MD_SHA512,GCRY_MD_FLAG_HMAC
5737 HMAC using a SHA-512 message digest.
5739 RSA encryption and signing.
5743 Note that the CRC algorithms are not considered cryptographic
5744 algorithms and thus are in addition available.
5746 * RSA key generation refuses to create a key with a keysize of less
5749 * DSA key generation refuses to create a key with a keysize other
5752 * The 'transient-key' flag for RSA and DSA key generation is ignored.
5754 * Support for the VIA Padlock engine is disabled.
5756 * FIPS mode may only be used on systems with a /dev/random device.
5757 Switching into FIPS mode on other systems will fail at runtime.
5759 * Saving and loading a random seed file is ignored.
5761 * An X9.31 style random number generator is used in place of the
5762 large-pool-CSPRNG generator.
5764 * The command 'GCRYCTL_ENABLE_QUICK_RANDOM' is ignored.
5766 * Message digest debugging is disabled.
5768 * All debug output related to cryptographic data is suppressed.
5770 * On-the-fly self-tests are not performed, instead self-tests are run
5771 before entering operational state.
5773 * The function 'gcry_set_allocation_handler' may not be used. If it
5774 is used Libgcrypt disables FIPS mode unless Enforced FIPS mode is
5775 enabled, in which case Libgcrypt will enter the error state.
5777 * The digest algorithm MD5 may not be used. If it is used Libgcrypt
5778 disables FIPS mode unless Enforced FIPS mode is enabled, in which
5779 case Libgcrypt will enter the error state.
5781 * In Enforced FIPS mode the command 'GCRYCTL_DISABLE_SECMEM' is
5782 ignored. In standard FIPS mode it disables FIPS mode.
5784 * A handler set by 'gcry_set_outofcore_handler' is ignored.
5785 * A handler set by 'gcry_set_fatalerror_handler' is ignored.
5787 Note that when we speak about disabling FIPS mode, it merely means
5788 that the function 'gcry_fips_mode_active' returns false; it does not
5789 mean that any non FIPS algorithms are allowed.
5791 B.2 FIPS Finite State Machine
5792 =============================
5794 The FIPS mode of libgcrypt implements a finite state machine (FSM) using
5795 8 states (*note Table B.1: tbl:fips-states.) and checks at runtime that
5796 only valid transitions (*note Table B.2: tbl:fips-state-transitions.)
5799 \0\b[image src="fips-fsm.png" alt="FIPS FSM Diagram"
\0\b]
5801 Figure B.1: FIPS mode state diagram
5803 States used by the FIPS FSM:
5806 Libgcrypt is not runtime linked to another application. This
5807 usually means that the library is not loaded into main memory.
5808 This state is documentation only.
5811 Libgcrypt is loaded into memory and API calls may be made.
5812 Compiler introduced constructor functions may be run. Note that
5813 Libgcrypt does not implement any arbitrary constructor functions to
5814 be called by the operating system
5817 The Libgcrypt initialization functions are performed and the
5818 library has not yet run any self-test.
5821 Libgcrypt is performing self-tests.
5824 Libgcrypt is in the operational state and all interfaces may be
5828 Libgrypt is in the error state. When calling any FIPS relevant
5829 interfaces they either return an error ('GPG_ERR_NOT_OPERATIONAL')
5830 or put Libgcrypt into the Fatal-Error state and won't return.
5833 Libgcrypt is in a non-recoverable error state and will
5834 automatically transit into the Shutdown state.
5837 Libgcrypt is about to be terminated and removed from the memory.
5838 The application may at this point still running cleanup handlers.
5840 Table B.1: FIPS mode states
5842 The valid state transitions (*note Figure B.1: fig:fips-fsm.) are:
5844 Power-Off to Power-On is implicitly done by the OS loading
5845 Libgcrypt as a shared library and having it linked to an
5849 Power-On to Init is triggered by the application calling the
5850 Libgcrypt initialization function 'gcry_check_version'.
5853 Init to Self-Test is either triggered by a dedicated API call or
5854 implicit by invoking a libgrypt service controlled by the FSM.
5857 Self-Test to Operational is triggered after all self-tests passed
5861 Operational to Shutdown is an artificial state without any direct
5862 action in Libgcrypt. When reaching the Shutdown state the library
5863 is deinitialized and can't return to any other state again.
5866 Shutdown to Power-off is the process of removing Libgcrypt from the
5867 computer's memory. For obvious reasons the Power-Off state can't
5868 be represented within Libgcrypt and thus this transition is for
5872 Operational to Error is triggered if Libgcrypt detected an
5873 application error which can't be returned to the caller but still
5874 allows Libgcrypt to properly run. In the Error state all FIPS
5875 relevant interfaces return an error code.
5878 Error to Shutdown is similar to the Operational to Shutdown
5882 Error to Fatal-Error is triggered if Libgrypt detects an fatal
5883 error while already being in Error state.
5886 Fatal-Error to Shutdown is automatically entered by Libgcrypt after
5887 having reported the error.
5890 Power-On to Shutdown is an artificial state to document that
5891 Libgcrypt has not ye been initialized but the process is about to
5895 Power-On to Fatal-Error will be triggered if certain Libgcrypt
5896 functions are used without having reached the Init state.
5899 Self-Test to Fatal-Error is triggered by severe errors in Libgcrypt
5900 while running self-tests.
5903 Self-Test to Error is triggered by a failed self-test.
5906 Operational to Fatal-Error is triggered if Libcrypt encountered a
5907 non-recoverable error.
5910 Operational to Self-Test is triggered if the application requested
5911 to run the self-tests again.
5914 Error to Self-Test is triggered if the application has requested to
5915 run self-tests to get to get back into operational state after an
5919 Init to Error is triggered by errors in the initialization code.
5922 Init to Fatal-Error is triggered by non-recoverable errors in the
5923 initialization code.
5926 Error to Error is triggered by errors while already in the Error
5929 Table B.2: FIPS mode state transitions
5931 B.3 FIPS Miscellaneous Information
5932 ==================================
5934 Libgcrypt does not do any key management on itself; the application
5935 needs to care about it. Keys which are passed to Libgcrypt should be
5936 allocated in secure memory as available with the functions
5937 'gcry_malloc_secure' and 'gcry_calloc_secure'. By calling 'gcry_free'
5938 on this memory, the memory and thus the keys are overwritten with zero
5939 bytes before releasing the memory.
5941 For use with the random number generator, Libgcrypt generates 3
5942 internal keys which are stored in the encryption contexts used by the
5943 RNG. These keys are stored in secure memory for the lifetime of the
5944 process. Application are required to use 'GCRYCTL_TERM_SECMEM' before
5945 process termination. This will zero out the entire secure memory and
5946 thus also the encryption contexts with these keys.
5949 File: gcrypt.info, Node: Library Copying, Next: Copying, Prev: FIPS Mode, Up: Top
5951 GNU Lesser General Public License
5952 *********************************
5954 Version 2.1, February 1999
5956 Copyright (C) 1991, 1999 Free Software Foundation, Inc.
5957 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
5959 Everyone is permitted to copy and distribute verbatim copies
5960 of this license document, but changing it is not allowed.
5962 [This is the first released version of the Lesser GPL. It also counts
5963 as the successor of the GNU Library Public License, version 2, hence the
5964 version number 2.1.]
5969 The licenses for most software are designed to take away your freedom to
5970 share and change it. By contrast, the GNU General Public Licenses are
5971 intended to guarantee your freedom to share and change free software--to
5972 make sure the software is free for all its users.
5974 This license, the Lesser General Public License, applies to some
5975 specially designated software--typically libraries--of the Free Software
5976 Foundation and other authors who decide to use it. You can use it too,
5977 but we suggest you first think carefully about whether this license or
5978 the ordinary General Public License is the better strategy to use in any
5979 particular case, based on the explanations below.
5981 When we speak of free software, we are referring to freedom of use,
5982 not price. Our General Public Licenses are designed to make sure that
5983 you have the freedom to distribute copies of free software (and charge
5984 for this service if you wish); that you receive source code or can get
5985 it if you want it; that you can change the software and use pieces of it
5986 in new free programs; and that you are informed that you can do these
5989 To protect your rights, we need to make restrictions that forbid
5990 distributors to deny you these rights or to ask you to surrender these
5991 rights. These restrictions translate to certain responsibilities for
5992 you if you distribute copies of the library or if you modify it.
5994 For example, if you distribute copies of the library, whether gratis
5995 or for a fee, you must give the recipients all the rights that we gave
5996 you. You must make sure that they, too, receive or can get the source
5997 code. If you link other code with the library, you must provide
5998 complete object files to the recipients, so that they can relink them
5999 with the library after making changes to the library and recompiling it.
6000 And you must show them these terms so they know their rights.
6002 We protect your rights with a two-step method: (1) we copyright the
6003 library, and (2) we offer you this license, which gives you legal
6004 permission to copy, distribute and/or modify the library.
6006 To protect each distributor, we want to make it very clear that there
6007 is no warranty for the free library. Also, if the library is modified
6008 by someone else and passed on, the recipients should know that what they
6009 have is not the original version, so that the original author's
6010 reputation will not be affected by problems that might be introduced by
6013 Finally, software patents pose a constant threat to the existence of
6014 any free program. We wish to make sure that a company cannot
6015 effectively restrict the users of a free program by obtaining a
6016 restrictive license from a patent holder. Therefore, we insist that any
6017 patent license obtained for a version of the library must be consistent
6018 with the full freedom of use specified in this license.
6020 Most GNU software, including some libraries, is covered by the
6021 ordinary GNU General Public License. This license, the GNU Lesser
6022 General Public License, applies to certain designated libraries, and is
6023 quite different from the ordinary General Public License. We use this
6024 license for certain libraries in order to permit linking those libraries
6025 into non-free programs.
6027 When a program is linked with a library, whether statically or using
6028 a shared library, the combination of the two is legally speaking a
6029 combined work, a derivative of the original library. The ordinary
6030 General Public License therefore permits such linking only if the entire
6031 combination fits its criteria of freedom. The Lesser General Public
6032 License permits more lax criteria for linking other code with the
6035 We call this license the "Lesser" General Public License because it
6036 does _Less_ to protect the user's freedom than the ordinary General
6037 Public License. It also provides other free software developers Less of
6038 an advantage over competing non-free programs. These disadvantages are
6039 the reason we use the ordinary General Public License for many
6040 libraries. However, the Lesser license provides advantages in certain
6041 special circumstances.
6043 For example, on rare occasions, there may be a special need to
6044 encourage the widest possible use of a certain library, so that it
6045 becomes a de-facto standard. To achieve this, non-free programs must be
6046 allowed to use the library. A more frequent case is that a free library
6047 does the same job as widely used non-free libraries. In this case,
6048 there is little to gain by limiting the free library to free software
6049 only, so we use the Lesser General Public License.
6051 In other cases, permission to use a particular library in non-free
6052 programs enables a greater number of people to use a large body of free
6053 software. For example, permission to use the GNU C Library in non-free
6054 programs enables many more people to use the whole GNU operating system,
6055 as well as its variant, the GNU/Linux operating system.
6057 Although the Lesser General Public License is Less protective of the
6058 users' freedom, it does ensure that the user of a program that is linked
6059 with the Library has the freedom and the wherewithal to run that program
6060 using a modified version of the Library.
6062 The precise terms and conditions for copying, distribution and
6063 modification follow. Pay close attention to the difference between a
6064 "work based on the library" and a "work that uses the library". The
6065 former contains code derived from the library, whereas the latter must
6066 be combined with the library in order to run.
6068 GNU LESSER GENERAL PUBLIC LICENSE
6069 TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
6071 0. This License Agreement applies to any software library or other
6072 program which contains a notice placed by the copyright holder or
6073 other authorized party saying it may be distributed under the terms
6074 of this Lesser General Public License (also called "this License").
6075 Each licensee is addressed as "you".
6077 A "library" means a collection of software functions and/or data
6078 prepared so as to be conveniently linked with application programs
6079 (which use some of those functions and data) to form executables.
6081 The "Library", below, refers to any such software library or work
6082 which has been distributed under these terms. A "work based on the
6083 Library" means either the Library or any derivative work under
6084 copyright law: that is to say, a work containing the Library or a
6085 portion of it, either verbatim or with modifications and/or
6086 translated straightforwardly into another language. (Hereinafter,
6087 translation is included without limitation in the term
6090 "Source code" for a work means the preferred form of the work for
6091 making modifications to it. For a library, complete source code
6092 means all the source code for all modules it contains, plus any
6093 associated interface definition files, plus the scripts used to
6094 control compilation and installation of the library.
6096 Activities other than copying, distribution and modification are
6097 not covered by this License; they are outside its scope. The act
6098 of running a program using the Library is not restricted, and
6099 output from such a program is covered only if its contents
6100 constitute a work based on the Library (independent of the use of
6101 the Library in a tool for writing it). Whether that is true
6102 depends on what the Library does and what the program that uses the
6105 1. You may copy and distribute verbatim copies of the Library's
6106 complete source code as you receive it, in any medium, provided
6107 that you conspicuously and appropriately publish on each copy an
6108 appropriate copyright notice and disclaimer of warranty; keep
6109 intact all the notices that refer to this License and to the
6110 absence of any warranty; and distribute a copy of this License
6111 along with the Library.
6113 You may charge a fee for the physical act of transferring a copy,
6114 and you may at your option offer warranty protection in exchange
6117 2. You may modify your copy or copies of the Library or any portion of
6118 it, thus forming a work based on the Library, and copy and
6119 distribute such modifications or work under the terms of Section 1
6120 above, provided that you also meet all of these conditions:
6122 a. The modified work must itself be a software library.
6124 b. You must cause the files modified to carry prominent notices
6125 stating that you changed the files and the date of any change.
6127 c. You must cause the whole of the work to be licensed at no
6128 charge to all third parties under the terms of this License.
6130 d. If a facility in the modified Library refers to a function or
6131 a table of data to be supplied by an application program that
6132 uses the facility, other than as an argument passed when the
6133 facility is invoked, then you must make a good faith effort to
6134 ensure that, in the event an application does not supply such
6135 function or table, the facility still operates, and performs
6136 whatever part of its purpose remains meaningful.
6138 (For example, a function in a library to compute square roots
6139 has a purpose that is entirely well-defined independent of the
6140 application. Therefore, Subsection 2d requires that any
6141 application-supplied function or table used by this function
6142 must be optional: if the application does not supply it, the
6143 square root function must still compute square roots.)
6145 These requirements apply to the modified work as a whole. If
6146 identifiable sections of that work are not derived from the
6147 Library, and can be reasonably considered independent and separate
6148 works in themselves, then this License, and its terms, do not apply
6149 to those sections when you distribute them as separate works. But
6150 when you distribute the same sections as part of a whole which is a
6151 work based on the Library, the distribution of the whole must be on
6152 the terms of this License, whose permissions for other licensees
6153 extend to the entire whole, and thus to each and every part
6154 regardless of who wrote it.
6156 Thus, it is not the intent of this section to claim rights or
6157 contest your rights to work written entirely by you; rather, the
6158 intent is to exercise the right to control the distribution of
6159 derivative or collective works based on the Library.
6161 In addition, mere aggregation of another work not based on the
6162 Library with the Library (or with a work based on the Library) on a
6163 volume of a storage or distribution medium does not bring the other
6164 work under the scope of this License.
6166 3. You may opt to apply the terms of the ordinary GNU General Public
6167 License instead of this License to a given copy of the Library. To
6168 do this, you must alter all the notices that refer to this License,
6169 so that they refer to the ordinary GNU General Public License,
6170 version 2, instead of to this License. (If a newer version than
6171 version 2 of the ordinary GNU General Public License has appeared,
6172 then you can specify that version instead if you wish.) Do not
6173 make any other change in these notices.
6175 Once this change is made in a given copy, it is irreversible for
6176 that copy, so the ordinary GNU General Public License applies to
6177 all subsequent copies and derivative works made from that copy.
6179 This option is useful when you wish to copy part of the code of the
6180 Library into a program that is not a library.
6182 4. You may copy and distribute the Library (or a portion or derivative
6183 of it, under Section 2) in object code or executable form under the
6184 terms of Sections 1 and 2 above provided that you accompany it with
6185 the complete corresponding machine-readable source code, which must
6186 be distributed under the terms of Sections 1 and 2 above on a
6187 medium customarily used for software interchange.
6189 If distribution of object code is made by offering access to copy
6190 from a designated place, then offering equivalent access to copy
6191 the source code from the same place satisfies the requirement to
6192 distribute the source code, even though third parties are not
6193 compelled to copy the source along with the object code.
6195 5. A program that contains no derivative of any portion of the
6196 Library, but is designed to work with the Library by being compiled
6197 or linked with it, is called a "work that uses the Library". Such
6198 a work, in isolation, is not a derivative work of the Library, and
6199 therefore falls outside the scope of this License.
6201 However, linking a "work that uses the Library" with the Library
6202 creates an executable that is a derivative of the Library (because
6203 it contains portions of the Library), rather than a "work that uses
6204 the library". The executable is therefore covered by this License.
6205 Section 6 states terms for distribution of such executables.
6207 When a "work that uses the Library" uses material from a header
6208 file that is part of the Library, the object code for the work may
6209 be a derivative work of the Library even though the source code is
6210 not. Whether this is true is especially significant if the work
6211 can be linked without the Library, or if the work is itself a
6212 library. The threshold for this to be true is not precisely
6215 If such an object file uses only numerical parameters, data
6216 structure layouts and accessors, and small macros and small inline
6217 functions (ten lines or less in length), then the use of the object
6218 file is unrestricted, regardless of whether it is legally a
6219 derivative work. (Executables containing this object code plus
6220 portions of the Library will still fall under Section 6.)
6222 Otherwise, if the work is a derivative of the Library, you may
6223 distribute the object code for the work under the terms of Section
6224 6. Any executables containing that work also fall under Section 6,
6225 whether or not they are linked directly with the Library itself.
6227 6. As an exception to the Sections above, you may also combine or link
6228 a "work that uses the Library" with the Library to produce a work
6229 containing portions of the Library, and distribute that work under
6230 terms of your choice, provided that the terms permit modification
6231 of the work for the customer's own use and reverse engineering for
6232 debugging such modifications.
6234 You must give prominent notice with each copy of the work that the
6235 Library is used in it and that the Library and its use are covered
6236 by this License. You must supply a copy of this License. If the
6237 work during execution displays copyright notices, you must include
6238 the copyright notice for the Library among them, as well as a
6239 reference directing the user to the copy of this License. Also,
6240 you must do one of these things:
6242 a. Accompany the work with the complete corresponding
6243 machine-readable source code for the Library including
6244 whatever changes were used in the work (which must be
6245 distributed under Sections 1 and 2 above); and, if the work is
6246 an executable linked with the Library, with the complete
6247 machine-readable "work that uses the Library", as object code
6248 and/or source code, so that the user can modify the Library
6249 and then relink to produce a modified executable containing
6250 the modified Library. (It is understood that the user who
6251 changes the contents of definitions files in the Library will
6252 not necessarily be able to recompile the application to use
6253 the modified definitions.)
6255 b. Use a suitable shared library mechanism for linking with the
6256 Library. A suitable mechanism is one that (1) uses at run
6257 time a copy of the library already present on the user's
6258 computer system, rather than copying library functions into
6259 the executable, and (2) will operate properly with a modified
6260 version of the library, if the user installs one, as long as
6261 the modified version is interface-compatible with the version
6262 that the work was made with.
6264 c. Accompany the work with a written offer, valid for at least
6265 three years, to give the same user the materials specified in
6266 Subsection 6a, above, for a charge no more than the cost of
6267 performing this distribution.
6269 d. If distribution of the work is made by offering access to copy
6270 from a designated place, offer equivalent access to copy the
6271 above specified materials from the same place.
6273 e. Verify that the user has already received a copy of these
6274 materials or that you have already sent this user a copy.
6276 For an executable, the required form of the "work that uses the
6277 Library" must include any data and utility programs needed for
6278 reproducing the executable from it. However, as a special
6279 exception, the materials to be distributed need not include
6280 anything that is normally distributed (in either source or binary
6281 form) with the major components (compiler, kernel, and so on) of
6282 the operating system on which the executable runs, unless that
6283 component itself accompanies the executable.
6285 It may happen that this requirement contradicts the license
6286 restrictions of other proprietary libraries that do not normally
6287 accompany the operating system. Such a contradiction means you
6288 cannot use both them and the Library together in an executable that
6291 7. You may place library facilities that are a work based on the
6292 Library side-by-side in a single library together with other
6293 library facilities not covered by this License, and distribute such
6294 a combined library, provided that the separate distribution of the
6295 work based on the Library and of the other library facilities is
6296 otherwise permitted, and provided that you do these two things:
6298 a. Accompany the combined library with a copy of the same work
6299 based on the Library, uncombined with any other library
6300 facilities. This must be distributed under the terms of the
6303 b. Give prominent notice with the combined library of the fact
6304 that part of it is a work based on the Library, and explaining
6305 where to find the accompanying uncombined form of the same
6308 8. You may not copy, modify, sublicense, link with, or distribute the
6309 Library except as expressly provided under this License. Any
6310 attempt otherwise to copy, modify, sublicense, link with, or
6311 distribute the Library is void, and will automatically terminate
6312 your rights under this License. However, parties who have received
6313 copies, or rights, from you under this License will not have their
6314 licenses terminated so long as such parties remain in full
6317 9. You are not required to accept this License, since you have not
6318 signed it. However, nothing else grants you permission to modify
6319 or distribute the Library or its derivative works. These actions
6320 are prohibited by law if you do not accept this License.
6321 Therefore, by modifying or distributing the Library (or any work
6322 based on the Library), you indicate your acceptance of this License
6323 to do so, and all its terms and conditions for copying,
6324 distributing or modifying the Library or works based on it.
6326 10. Each time you redistribute the Library (or any work based on the
6327 Library), the recipient automatically receives a license from the
6328 original licensor to copy, distribute, link with or modify the
6329 Library subject to these terms and conditions. You may not impose
6330 any further restrictions on the recipients' exercise of the rights
6331 granted herein. You are not responsible for enforcing compliance
6332 by third parties with this License.
6334 11. If, as a consequence of a court judgment or allegation of patent
6335 infringement or for any other reason (not limited to patent
6336 issues), conditions are imposed on you (whether by court order,
6337 agreement or otherwise) that contradict the conditions of this
6338 License, they do not excuse you from the conditions of this
6339 License. If you cannot distribute so as to satisfy simultaneously
6340 your obligations under this License and any other pertinent
6341 obligations, then as a consequence you may not distribute the
6342 Library at all. For example, if a patent license would not permit
6343 royalty-free redistribution of the Library by all those who receive
6344 copies directly or indirectly through you, then the only way you
6345 could satisfy both it and this License would be to refrain entirely
6346 from distribution of the Library.
6348 If any portion of this section is held invalid or unenforceable
6349 under any particular circumstance, the balance of the section is
6350 intended to apply, and the section as a whole is intended to apply
6351 in other circumstances.
6353 It is not the purpose of this section to induce you to infringe any
6354 patents or other property right claims or to contest validity of
6355 any such claims; this section has the sole purpose of protecting
6356 the integrity of the free software distribution system which is
6357 implemented by public license practices. Many people have made
6358 generous contributions to the wide range of software distributed
6359 through that system in reliance on consistent application of that
6360 system; it is up to the author/donor to decide if he or she is
6361 willing to distribute software through any other system and a
6362 licensee cannot impose that choice.
6364 This section is intended to make thoroughly clear what is believed
6365 to be a consequence of the rest of this License.
6367 12. If the distribution and/or use of the Library is restricted in
6368 certain countries either by patents or by copyrighted interfaces,
6369 the original copyright holder who places the Library under this
6370 License may add an explicit geographical distribution limitation
6371 excluding those countries, so that distribution is permitted only
6372 in or among countries not thus excluded. In such case, this
6373 License incorporates the limitation as if written in the body of
6376 13. The Free Software Foundation may publish revised and/or new
6377 versions of the Lesser General Public License from time to time.
6378 Such new versions will be similar in spirit to the present version,
6379 but may differ in detail to address new problems or concerns.
6381 Each version is given a distinguishing version number. If the
6382 Library specifies a version number of this License which applies to
6383 it and "any later version", you have the option of following the
6384 terms and conditions either of that version or of any later version
6385 published by the Free Software Foundation. If the Library does not
6386 specify a license version number, you may choose any version ever
6387 published by the Free Software Foundation.
6389 14. If you wish to incorporate parts of the Library into other free
6390 programs whose distribution conditions are incompatible with these,
6391 write to the author to ask for permission. For software which is
6392 copyrighted by the Free Software Foundation, write to the Free
6393 Software Foundation; we sometimes make exceptions for this. Our
6394 decision will be guided by the two goals of preserving the free
6395 status of all derivatives of our free software and of promoting the
6396 sharing and reuse of software generally.
6400 15. BECAUSE THE LIBRARY IS LICENSED FREE OF CHARGE, THERE IS NO
6401 WARRANTY FOR THE LIBRARY, TO THE EXTENT PERMITTED BY APPLICABLE
6402 LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS
6403 AND/OR OTHER PARTIES PROVIDE THE LIBRARY "AS IS" WITHOUT WARRANTY
6404 OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT
6405 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
6406 FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND
6407 PERFORMANCE OF THE LIBRARY IS WITH YOU. SHOULD THE LIBRARY PROVE
6408 DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR
6411 16. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN
6412 WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY
6413 MODIFY AND/OR REDISTRIBUTE THE LIBRARY AS PERMITTED ABOVE, BE
6414 LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL,
6415 INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR
6416 INABILITY TO USE THE LIBRARY (INCLUDING BUT NOT LIMITED TO LOSS OF
6417 DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU
6418 OR THIRD PARTIES OR A FAILURE OF THE LIBRARY TO OPERATE WITH ANY
6419 OTHER SOFTWARE), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN
6420 ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
6422 END OF TERMS AND CONDITIONS
6424 How to Apply These Terms to Your New Libraries
6425 ==============================================
6427 If you develop a new library, and you want it to be of the greatest
6428 possible use to the public, we recommend making it free software that
6429 everyone can redistribute and change. You can do so by permitting
6430 redistribution under these terms (or, alternatively, under the terms of
6431 the ordinary General Public License).
6433 To apply these terms, attach the following notices to the library.
6434 It is safest to attach them to the start of each source file to most
6435 effectively convey the exclusion of warranty; and each file should have
6436 at least the "copyright" line and a pointer to where the full notice is
6439 ONE LINE TO GIVE THE LIBRARY'S NAME AND AN IDEA OF WHAT IT DOES.
6440 Copyright (C) YEAR NAME OF AUTHOR
6442 This library is free software; you can redistribute it and/or modify it
6443 under the terms of the GNU Lesser General Public License as published by
6444 the Free Software Foundation; either version 2.1 of the License, or (at
6445 your option) any later version.
6447 This library is distributed in the hope that it will be useful, but
6448 WITHOUT ANY WARRANTY; without even the implied warranty of
6449 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
6450 Lesser General Public License for more details.
6452 You should have received a copy of the GNU Lesser General Public
6453 License along with this library; if not, write to the Free Software
6454 Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307,
6457 Also add information on how to contact you by electronic and paper
6460 You should also get your employer (if you work as a programmer) or
6461 your school, if any, to sign a "copyright disclaimer" for the library,
6462 if necessary. Here is a sample; alter the names:
6464 Yoyodyne, Inc., hereby disclaims all copyright interest in the library
6465 `Frob' (a library for tweaking knobs) written by James Random Hacker.
6467 SIGNATURE OF TY COON, 1 April 1990
6468 Ty Coon, President of Vice
6470 That's all there is to it!
6473 File: gcrypt.info, Node: Copying, Next: Figures and Tables, Prev: Library Copying, Up: Top
6475 GNU General Public License
6476 **************************
6478 Version 2, June 1991
6480 Copyright (C) 1989, 1991 Free Software Foundation, Inc.
6481 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
6483 Everyone is permitted to copy and distribute verbatim copies
6484 of this license document, but changing it is not allowed.
6489 The licenses for most software are designed to take away your freedom to
6490 share and change it. By contrast, the GNU General Public License is
6491 intended to guarantee your freedom to share and change free software--to
6492 make sure the software is free for all its users. This General Public
6493 License applies to most of the Free Software Foundation's software and
6494 to any other program whose authors commit to using it. (Some other Free
6495 Software Foundation software is covered by the GNU Library General
6496 Public License instead.) You can apply it to your programs, too.
6498 When we speak of free software, we are referring to freedom, not
6499 price. Our General Public Licenses are designed to make sure that you
6500 have the freedom to distribute copies of free software (and charge for
6501 this service if you wish), that you receive source code or can get it if
6502 you want it, that you can change the software or use pieces of it in new
6503 free programs; and that you know you can do these things.
6505 To protect your rights, we need to make restrictions that forbid
6506 anyone to deny you these rights or to ask you to surrender the rights.
6507 These restrictions translate to certain responsibilities for you if you
6508 distribute copies of the software, or if you modify it.
6510 For example, if you distribute copies of such a program, whether
6511 gratis or for a fee, you must give the recipients all the rights that
6512 you have. You must make sure that they, too, receive or can get the
6513 source code. And you must show them these terms so they know their
6516 We protect your rights with two steps: (1) copyright the software,
6517 and (2) offer you this license which gives you legal permission to copy,
6518 distribute and/or modify the software.
6520 Also, for each author's protection and ours, we want to make certain
6521 that everyone understands that there is no warranty for this free
6522 software. If the software is modified by someone else and passed on, we
6523 want its recipients to know that what they have is not the original, so
6524 that any problems introduced by others will not reflect on the original
6525 authors' reputations.
6527 Finally, any free program is threatened constantly by software
6528 patents. We wish to avoid the danger that redistributors of a free
6529 program will individually obtain patent licenses, in effect making the
6530 program proprietary. To prevent this, we have made it clear that any
6531 patent must be licensed for everyone's free use or not licensed at all.
6533 The precise terms and conditions for copying, distribution and
6534 modification follow.
6536 TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
6538 1. This License applies to any program or other work which contains a
6539 notice placed by the copyright holder saying it may be distributed
6540 under the terms of this General Public License. The "Program",
6541 below, refers to any such program or work, and a "work based on the
6542 Program" means either the Program or any derivative work under
6543 copyright law: that is to say, a work containing the Program or a
6544 portion of it, either verbatim or with modifications and/or
6545 translated into another language. (Hereinafter, translation is
6546 included without limitation in the term "modification".) Each
6547 licensee is addressed as "you".
6549 Activities other than copying, distribution and modification are
6550 not covered by this License; they are outside its scope. The act
6551 of running the Program is not restricted, and the output from the
6552 Program is covered only if its contents constitute a work based on
6553 the Program (independent of having been made by running the
6554 Program). Whether that is true depends on what the Program does.
6556 2. You may copy and distribute verbatim copies of the Program's source
6557 code as you receive it, in any medium, provided that you
6558 conspicuously and appropriately publish on each copy an appropriate
6559 copyright notice and disclaimer of warranty; keep intact all the
6560 notices that refer to this License and to the absence of any
6561 warranty; and give any other recipients of the Program a copy of
6562 this License along with the Program.
6564 You may charge a fee for the physical act of transferring a copy,
6565 and you may at your option offer warranty protection in exchange
6568 3. You may modify your copy or copies of the Program or any portion of
6569 it, thus forming a work based on the Program, and copy and
6570 distribute such modifications or work under the terms of Section 1
6571 above, provided that you also meet all of these conditions:
6573 a. You must cause the modified files to carry prominent notices
6574 stating that you changed the files and the date of any change.
6576 b. You must cause any work that you distribute or publish, that
6577 in whole or in part contains or is derived from the Program or
6578 any part thereof, to be licensed as a whole at no charge to
6579 all third parties under the terms of this License.
6581 c. If the modified program normally reads commands interactively
6582 when run, you must cause it, when started running for such
6583 interactive use in the most ordinary way, to print or display
6584 an announcement including an appropriate copyright notice and
6585 a notice that there is no warranty (or else, saying that you
6586 provide a warranty) and that users may redistribute the
6587 program under these conditions, and telling the user how to
6588 view a copy of this License. (Exception: if the Program
6589 itself is interactive but does not normally print such an
6590 announcement, your work based on the Program is not required
6591 to print an announcement.)
6593 These requirements apply to the modified work as a whole. If
6594 identifiable sections of that work are not derived from the
6595 Program, and can be reasonably considered independent and separate
6596 works in themselves, then this License, and its terms, do not apply
6597 to those sections when you distribute them as separate works. But
6598 when you distribute the same sections as part of a whole which is a
6599 work based on the Program, the distribution of the whole must be on
6600 the terms of this License, whose permissions for other licensees
6601 extend to the entire whole, and thus to each and every part
6602 regardless of who wrote it.
6604 Thus, it is not the intent of this section to claim rights or
6605 contest your rights to work written entirely by you; rather, the
6606 intent is to exercise the right to control the distribution of
6607 derivative or collective works based on the Program.
6609 In addition, mere aggregation of another work not based on the
6610 Program with the Program (or with a work based on the Program) on a
6611 volume of a storage or distribution medium does not bring the other
6612 work under the scope of this License.
6614 4. You may copy and distribute the Program (or a work based on it,
6615 under Section 2) in object code or executable form under the terms
6616 of Sections 1 and 2 above provided that you also do one of the
6619 a. Accompany it with the complete corresponding machine-readable
6620 source code, which must be distributed under the terms of
6621 Sections 1 and 2 above on a medium customarily used for
6622 software interchange; or,
6624 b. Accompany it with a written offer, valid for at least three
6625 years, to give any third party, for a charge no more than your
6626 cost of physically performing source distribution, a complete
6627 machine-readable copy of the corresponding source code, to be
6628 distributed under the terms of Sections 1 and 2 above on a
6629 medium customarily used for software interchange; or,
6631 c. Accompany it with the information you received as to the offer
6632 to distribute corresponding source code. (This alternative is
6633 allowed only for noncommercial distribution and only if you
6634 received the program in object code or executable form with
6635 such an offer, in accord with Subsection b above.)
6637 The source code for a work means the preferred form of the work for
6638 making modifications to it. For an executable work, complete
6639 source code means all the source code for all modules it contains,
6640 plus any associated interface definition files, plus the scripts
6641 used to control compilation and installation of the executable.
6642 However, as a special exception, the source code distributed need
6643 not include anything that is normally distributed (in either source
6644 or binary form) with the major components (compiler, kernel, and so
6645 on) of the operating system on which the executable runs, unless
6646 that component itself accompanies the executable.
6648 If distribution of executable or object code is made by offering
6649 access to copy from a designated place, then offering equivalent
6650 access to copy the source code from the same place counts as
6651 distribution of the source code, even though third parties are not
6652 compelled to copy the source along with the object code.
6654 5. You may not copy, modify, sublicense, or distribute the Program
6655 except as expressly provided under this License. Any attempt
6656 otherwise to copy, modify, sublicense or distribute the Program is
6657 void, and will automatically terminate your rights under this
6658 License. However, parties who have received copies, or rights,
6659 from you under this License will not have their licenses terminated
6660 so long as such parties remain in full compliance.
6662 6. You are not required to accept this License, since you have not
6663 signed it. However, nothing else grants you permission to modify
6664 or distribute the Program or its derivative works. These actions
6665 are prohibited by law if you do not accept this License.
6666 Therefore, by modifying or distributing the Program (or any work
6667 based on the Program), you indicate your acceptance of this License
6668 to do so, and all its terms and conditions for copying,
6669 distributing or modifying the Program or works based on it.
6671 7. Each time you redistribute the Program (or any work based on the
6672 Program), the recipient automatically receives a license from the
6673 original licensor to copy, distribute or modify the Program subject
6674 to these terms and conditions. You may not impose any further
6675 restrictions on the recipients' exercise of the rights granted
6676 herein. You are not responsible for enforcing compliance by third
6677 parties to this License.
6679 8. If, as a consequence of a court judgment or allegation of patent
6680 infringement or for any other reason (not limited to patent
6681 issues), conditions are imposed on you (whether by court order,
6682 agreement or otherwise) that contradict the conditions of this
6683 License, they do not excuse you from the conditions of this
6684 License. If you cannot distribute so as to satisfy simultaneously
6685 your obligations under this License and any other pertinent
6686 obligations, then as a consequence you may not distribute the
6687 Program at all. For example, if a patent license would not permit
6688 royalty-free redistribution of the Program by all those who receive
6689 copies directly or indirectly through you, then the only way you
6690 could satisfy both it and this License would be to refrain entirely
6691 from distribution of the Program.
6693 If any portion of this section is held invalid or unenforceable
6694 under any particular circumstance, the balance of the section is
6695 intended to apply and the section as a whole is intended to apply
6696 in other circumstances.
6698 It is not the purpose of this section to induce you to infringe any
6699 patents or other property right claims or to contest validity of
6700 any such claims; this section has the sole purpose of protecting
6701 the integrity of the free software distribution system, which is
6702 implemented by public license practices. Many people have made
6703 generous contributions to the wide range of software distributed
6704 through that system in reliance on consistent application of that
6705 system; it is up to the author/donor to decide if he or she is
6706 willing to distribute software through any other system and a
6707 licensee cannot impose that choice.
6709 This section is intended to make thoroughly clear what is believed
6710 to be a consequence of the rest of this License.
6712 9. If the distribution and/or use of the Program is restricted in
6713 certain countries either by patents or by copyrighted interfaces,
6714 the original copyright holder who places the Program under this
6715 License may add an explicit geographical distribution limitation
6716 excluding those countries, so that distribution is permitted only
6717 in or among countries not thus excluded. In such case, this
6718 License incorporates the limitation as if written in the body of
6721 10. The Free Software Foundation may publish revised and/or new
6722 versions of the General Public License from time to time. Such new
6723 versions will be similar in spirit to the present version, but may
6724 differ in detail to address new problems or concerns.
6726 Each version is given a distinguishing version number. If the
6727 Program specifies a version number of this License which applies to
6728 it and "any later version", you have the option of following the
6729 terms and conditions either of that version or of any later version
6730 published by the Free Software Foundation. If the Program does not
6731 specify a version number of this License, you may choose any
6732 version ever published by the Free Software Foundation.
6734 11. If you wish to incorporate parts of the Program into other free
6735 programs whose distribution conditions are different, write to the
6736 author to ask for permission. For software which is copyrighted by
6737 the Free Software Foundation, write to the Free Software
6738 Foundation; we sometimes make exceptions for this. Our decision
6739 will be guided by the two goals of preserving the free status of
6740 all derivatives of our free software and of promoting the sharing
6741 and reuse of software generally.
6745 12. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO
6746 WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE
6747 LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS
6748 AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY
6749 OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT
6750 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
6751 FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND
6752 PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE
6753 DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR
6756 13. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN
6757 WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY
6758 MODIFY AND/OR REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE
6759 LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL,
6760 INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR
6761 INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF
6762 DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU
6763 OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY
6764 OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN
6765 ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
6767 END OF TERMS AND CONDITIONS
6769 How to Apply These Terms to Your New Programs
6770 =============================================
6772 If you develop a new program, and you want it to be of the greatest
6773 possible use to the public, the best way to achieve this is to make it
6774 free software which everyone can redistribute and change under these
6777 To do so, attach the following notices to the program. It is safest
6778 to attach them to the start of each source file to most effectively
6779 convey the exclusion of warranty; and each file should have at least the
6780 "copyright" line and a pointer to where the full notice is found.
6782 ONE LINE TO GIVE THE PROGRAM'S NAME AND AN IDEA OF WHAT IT DOES.
6783 Copyright (C) 19YY NAME OF AUTHOR
6785 This program is free software; you can redistribute it and/or
6786 modify it under the terms of the GNU General Public License
6787 as published by the Free Software Foundation; either version 2
6788 of the License, or (at your option) any later version.
6790 This program is distributed in the hope that it will be useful,
6791 but WITHOUT ANY WARRANTY; without even the implied warranty of
6792 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
6793 GNU General Public License for more details.
6795 You should have received a copy of the GNU General Public License along
6796 with this program; if not, write to the Free Software Foundation, Inc.,
6797 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA.
6799 Also add information on how to contact you by electronic and paper
6802 If the program is interactive, make it output a short notice like
6803 this when it starts in an interactive mode:
6805 Gnomovision version 69, Copyright (C) 19YY NAME OF AUTHOR
6806 Gnomovision comes with ABSOLUTELY NO WARRANTY; for details
6807 type `show w'. This is free software, and you are welcome
6808 to redistribute it under certain conditions; type `show c'
6811 The hypothetical commands 'show w' and 'show c' should show the
6812 appropriate parts of the General Public License. Of course, the
6813 commands you use may be called something other than 'show w' and 'show
6814 c'; they could even be mouse-clicks or menu items--whatever suits your
6817 You should also get your employer (if you work as a programmer) or
6818 your school, if any, to sign a "copyright disclaimer" for the program,
6819 if necessary. Here is a sample; alter the names:
6821 Yoyodyne, Inc., hereby disclaims all copyright
6822 interest in the program `Gnomovision'
6823 (which makes passes at compilers) written
6826 SIGNATURE OF TY COON, 1 April 1989
6827 Ty Coon, President of Vice
6829 This General Public License does not permit incorporating your
6830 program into proprietary programs. If your program is a subroutine
6831 library, you may consider it more useful to permit linking proprietary
6832 applications with the library. If this is what you want to do, use the
6833 GNU Library General Public License instead of this License.
6836 File: gcrypt.info, Node: Figures and Tables, Next: Concept Index, Prev: Copying, Up: Top
6838 List of Figures and Tables
6839 **************************
6843 * Figure 17.1: fig:subsystems. Libgcrypt subsystems
6844 * Figure B.1: fig:fips-fsm. FIPS mode state diagram
6848 * Table B.1: tbl:fips-states. FIPS mode states
6849 * Table B.2: tbl:fips-state-transitions. FIPS mode state transitions
6852 File: gcrypt.info, Node: Concept Index, Next: Function and Data Index, Prev: Figures and Tables, Up: Top
6860 * /etc/gcrypt/fips_enabled: Configuration. (line 69)
6861 * /etc/gcrypt/hwf.deny: Configuration. (line 48)
6862 * /etc/gcrypt/random.conf: Configuration. (line 52)
6863 * /proc/cpuinfo: Configuration. (line 74)
6864 * /proc/self/auxv: Configuration. (line 74)
6865 * 3DES: Available ciphers. (line 14)
6866 * Advanced Encryption Standard: Available ciphers. (line 35)
6867 * AES: Available ciphers. (line 35)
6868 * AES-Wrap mode: Available cipher modes.
6870 * Arcfour: Available ciphers. (line 52)
6871 * BLAKE2b-512, BLAKE2b-384, BLAKE2b-256, BLAKE2b-160: Available hash algorithms.
6873 * BLAKE2s-256, BLAKE2s-224, BLAKE2s-160, BLAKE2s-128: Available hash algorithms.
6875 * Blowfish: Available ciphers. (line 22)
6876 * bug emulation: Working with hash algorithms.
6878 * Camellia: Available ciphers. (line 77)
6879 * CAST5: Available ciphers. (line 19)
6880 * CBC, Cipher Block Chaining mode: Available cipher modes.
6882 * CBC-MAC: Working with cipher handles.
6884 * CCM, Counter with CBC-MAC mode: Available cipher modes.
6886 * CFB, Cipher Feedback mode: Available cipher modes.
6888 * ChaCha20: Available ciphers. (line 92)
6889 * cipher text stealing: Working with cipher handles.
6891 * comp: Cryptographic Functions.
6893 * CRC32: Available hash algorithms.
6895 * CTR, Counter mode: Available cipher modes.
6897 * DES: Available ciphers. (line 57)
6898 * DES-EDE: Available ciphers. (line 14)
6899 * Digital Encryption Standard: Available ciphers. (line 14)
6900 * disable-jent: Configuration. (line 58)
6901 * ECB, Electronic Codebook mode: Available cipher modes.
6903 * EdDSA: Cryptographic Functions.
6905 * Enforced FIPS mode: Enabling FIPS mode. (line 29)
6906 * error codes: Error Values. (line 6)
6907 * error codes, list of: Error Sources. (line 6)
6908 * error codes, list of <1>: Error Codes. (line 6)
6909 * error codes, printing of: Error Strings. (line 6)
6910 * error sources: Error Values. (line 6)
6911 * error sources, printing of: Error Strings. (line 6)
6912 * error strings: Error Strings. (line 6)
6913 * error values: Error Values. (line 6)
6914 * error values, printing of: Error Strings. (line 6)
6915 * FIPS 140: Enabling FIPS mode. (line 6)
6916 * FIPS 186: Cryptographic Functions.
6918 * FIPS 186 <1>: Public-Key Subsystem Architecture.
6920 * FIPS 186-2: Cryptographic Functions.
6922 * FIPS mode: Enabling FIPS mode. (line 6)
6923 * fips_enabled: Configuration. (line 69)
6924 * GCM, Galois/Counter Mode: Available cipher modes.
6926 * GCRYPT_BARRETT: Configuration. (line 12)
6927 * GCRYPT_RNDUNIX_DBG: Configuration. (line 17)
6928 * GCRYPT_RNDUNIX_DBGALL: Configuration. (line 17)
6929 * GCRYPT_RNDW32_DBG: Configuration. (line 32)
6930 * GCRYPT_RNDW32_NOPERF: Configuration. (line 25)
6931 * GOST 28147-89: Available ciphers. (line 88)
6932 * GPL, GNU General Public License: Copying. (line 6)
6933 * hardware features: Hardware features. (line 6)
6934 * HAVAL: Available hash algorithms.
6936 * HMAC: Working with hash algorithms.
6938 * HMAC-GOSTR-3411-94: Available MAC algorithms.
6940 * HMAC-MD2, HMAC-MD4, HMAC-MD5: Available MAC algorithms.
6942 * HMAC-RIPE-MD-160: Available MAC algorithms.
6944 * HMAC-SHA-1: Available MAC algorithms.
6946 * HMAC-SHA-224, HMAC-SHA-256, HMAC-SHA-384, HMAC-SHA-512: Available MAC algorithms.
6948 * HMAC-SHA3-224, HMAC-SHA3-256, HMAC-SHA3-384, HMAC-SHA3-512: Available MAC algorithms.
6950 * HMAC-Stribog-256, HMAC-Stribog-512: Available MAC algorithms.
6952 * HMAC-TIGER1: Available MAC algorithms.
6954 * HMAC-Whirlpool: Available MAC algorithms.
6956 * HOME: Configuration. (line 37)
6957 * IDEA: Available ciphers. (line 11)
6958 * LGPL, GNU Lesser General Public License: Library Copying. (line 6)
6959 * MD2, MD4, MD5: Available hash algorithms.
6961 * no-blinding: Cryptographic Functions.
6963 * no-keytest: Cryptographic Functions.
6965 * nocomp: Cryptographic Functions.
6967 * OAEP: Cryptographic Functions.
6969 * OCB, OCB3: Available cipher modes.
6971 * OFB, Output Feedback mode: Available cipher modes.
6973 * only-urandom: Configuration. (line 61)
6974 * param: Cryptographic Functions.
6976 * PKCS1: Cryptographic Functions.
6978 * Poly1305 based AEAD mode with ChaCha20: Available cipher modes.
6980 * PSS: Cryptographic Functions.
6982 * RC2: Available ciphers. (line 69)
6983 * RC4: Available ciphers. (line 52)
6984 * rfc-2268: Available ciphers. (line 69)
6985 * RFC6979: Cryptographic Functions.
6987 * Rijndael: Available ciphers. (line 35)
6988 * RIPE-MD-160: Available hash algorithms.
6990 * Salsa20: Available ciphers. (line 81)
6991 * Salsa20/12: Available ciphers. (line 84)
6992 * Seed (cipher): Available ciphers. (line 72)
6993 * Serpent: Available ciphers. (line 65)
6994 * SHA-1: Available hash algorithms.
6996 * SHA-224, SHA-256, SHA-384, SHA-512: Available hash algorithms.
6998 * SHA3-224, SHA3-256, SHA3-384, SHA3-512, SHAKE128, SHAKE256: Available hash algorithms.
7000 * sync mode (OpenPGP): Working with cipher handles.
7002 * TIGER, TIGER1, TIGER2: Available hash algorithms.
7004 * transient-key: Cryptographic Functions.
7006 * Triple-DES: Available ciphers. (line 14)
7007 * Twofish: Available ciphers. (line 46)
7008 * Whirlpool: Available hash algorithms.
7010 * X9.31: Cryptographic Functions.
7012 * X9.31 <1>: Public-Key Subsystem Architecture.
7014 * XTS, XTS mode: Available cipher modes.