A Hacker's Guide to GNUPG ================================ (Some notes on GNUPG internals.) ===> Under construction <======= CVS Access ========== Anonymous read-only CVS access is available: cvs -z3 -d :pserver:anoncvs@cvs.gnupg.org:/cvs/gnupg login use the password "anoncvs". To check out the the complete archive use: cvs -z3 -d :pserver:anoncvs@cvs.gnupg.org:/cvs/gnupg \ checkout -R STABLE-BRANCH-1-0 gnupg This service is provided to help you in hunting bugs and not to deliver stable snapshots; it may happen that it even does not compile, so please don't complain. CVS may put a high load on a server, so please don't poll poll for new updates but wait for an announcement; to receive this you may want to subscribe to: gnupg-commit-watchers@gnupg.org by sending a mail with subject "subscribe" to gnupg-commit-watchers-request@gnupg.org You must run scripts/autogen.sh before doing the ./configure, as this creates some needed while which are not in the CVS. autogen.sh should checks that you have all required tools installed. RSYNC access ============ The FTP archive is also available by anonymous rsync. A daily snapshot of the CVS head revision is also available. See rsync(1) and try "rsync ftp.gnupg.org::" to see available resources. Special Tools ============= Documentation is based on the docbook DTD. Actually we have only the man page for now. To build a man page you need the docbook-to-man tool and all the other thinks needed for SGML processing. Debian comes with the docbook tools and you only need this docbook-to-man script which is comes with gtk-doc or download it from ftp.openit.de:/pub/devel/sgml. If you don't have it everything should still work fine but you will have only a dummy man page. RFCs ==== 1423 Privacy Enhancement for Internet Electronic Mail: Part III: Algorithms, Modes, and Identifiers. 1489 Registration of a Cyrillic Character Set. 1750 Randomness Recommendations for Security. 1991 PGP Message Exchange Formats. 2015 MIME Security with Pretty Good Privacy (PGP). 2144 The CAST-128 Encryption Algorithm. 2279 UTF-8, a transformation format of ISO 10646. 2440 OpenPGP. Debug Flags ----------- Use the option "--debug n" to output debug information. This option can be used multiple times, all values are ORed; n maybe prefixed with 0x to use hex-values. value used for ----- ---------------------------------------------- 1 packet reading/writing 2 MPI details 4 ciphers and primes (may reveal sensitive data) 8 iobuf filter functions 16 iobuf stuff 32 memory allocation stuff 64 caching 128 show memory statistics at exit 256 trust verification stuff Directory Layout ---------------- ./ Readme, configure ./scripts Scripts needed by configure and others ./doc Documentation ./util General purpose utility function ./mpi Multi precision integer library ./cipher Cryptographic functions ./g10 GnuPG application ./tools Some helper and demo programs ./keybox The keybox library (under construction) ./gcrypt Stuff needed to build libgcrypt (under construction) Detailed Roadmap ---------------- g10/g10.c Main module with option parsing and all the stuff you have to do on startup. Also has the exout handler and some helper functions. g10/sign.c Create signature and optionally encrypt g10/parse-packet.c g10/build-packet.c g10/free-packet.c Parsing and creating of OpenPGP message packets. g10/getkey.c Key selection code g10/pkclist.c Build a list of public keys g10/skclist.c Build a list of secret keys g10/ringedit.c Keyring I/O g10/keydb.h g10/keyid.c Helper functions to get the keyid, fingerprint etc. g10/trustdb.c g10/trustdb.h g10/tdbdump.c Management of the trustdb.gpg g10/compress.c Filter to handle compression g10/filter.h Declarations for all filter functions g10/delkey.c Delete a key g10/kbnode.c Helper for the KBNODE linked list g10/main.h Prototypes and some constants g10/mainproc.c Message processing g10/armor.c Ascii armor filter g10/mdfilter.c Filter to calculate hashs g10/textfilter.c Filter to handle CR/LF and trailing white space g10/cipher.c En-/Decryption filter g10/misc.c Utlity functions g10/options.h Structure with all the command line options and related constants g10/openfile.c Create/Open Files g10/tdbio.c I/O handling for the trustdb.gpg g10/tdbio.h g10/hkp.h Keyserver access g10/hkp.c g10/packet.h Defintion of OpenPGP structures. g10/passphrase.c Passphrase handling code g10/pubkey-enc.c g10/seckey-cert.c g10/seskey.c g10/import.c g10/export.c g10/comment.c g10/status.c g10/status.h g10/sign.c g10/plaintext.c g10/encr-data.c g10/encode.c g10/revoke.c g10/keylist.c g10/sig-check.c g10/signal.c g10/helptext.c g10/verify.c g10/decrypt.c g10/keyedit.c g10/dearmor.c g10/keygen.c Memory allocation ----------------- Use only the functions: m_alloc() m_alloc_clear() m_strdup() m_free() If you want to store a passphrase or some other sensitive data you may want to use m_alloc_secure() instead of m_alloc(), as this puts the data into a memory region which is protected from swapping (on some platforms). m_free() works for both. This functions will not return if there is not enough memory available. Logging ------- Option parsing --------------- GNUPG does not use getopt or GNU getopt but functions of it's own. See util/argparse.c for details. The advantage of these functions is that it is more easy to display and maintain the help texts for the options. The same option table is also used to parse resource files. What is an IOBUF ---------------- This is the data structure used for most I/O of gnupg. It is similar to System V Streams but much simpler. Because OpenPGP messages are nested in different ways; the use of such a system has big advantages. Here is an example, how it works: If the parser sees a packet header with a partial length, it pushes the block_filter onto the IOBUF to handle these partial length packets: from now on you don't have to worry about this. When it sees a compressed packet it pushes the uncompress filter and the next read byte is one which has already been uncompressed by this filter. Same goes for enciphered packet, plaintext packets and so on. The file g10/encode.c might be a good staring point to see how it is used - actually this is the other way: constructing messages using pushed filters but it may be easier to understand. How to use the message digest functions --------------------------------------- cipher/md.c implements an interface to hash (message digest functions). a) If you have a common part of data and some variable parts and you need to hash of the concatenated parts, you can use this: md = md_open(...) md_write( md, common_part ) md1 = md_copy( md ) md_write(md1, part1) md_final(md1); digest1 = md_read(md1) md2 = md_copy( md ) md_write(md2, part2) md_final(md2); digest2 = md_read(md2) An example are key signatures; the key packet is the common part and the user-id packets are the variable parts. b) If you need a running digest you should use this: md = md_open(...) md_write( md, part1 ) digest_of_part1 = md_digest( md ); md_write( md, part2 ) digest_of_part1_cat_part2 = md_digest( md ); .... Both methods may be combined. [Please see the source for the real syntax] How to use the cipher functions ------------------------------- cipher/cipher.c implements the interface to symmetric encryption functions. As usual you have a function to open a cipher (which returns a handle to be used with all other functions), some functions to set the key and other stuff and a encrypt and decrypt function which does the real work. You probably know how to work with files - so it should really be easy to work with these functions. Here is an example: CIPHER_HANDLE hd; hd = cipher_open( CIPHER_ALGO_TWOFISH, CIPHER_MODE_CFB, 0 ); if( !hd ) oops( use other function to check for the real error ); rc = cipher_setkey( hd, key256bit, 32 ) ) if( rc ) oops( weak key or something like this ); cipher_setiv( hd, some_IV_or_NULL_for_all_zeroes ); cipher_encrypt( hd, plain, cipher, size ); cipher_close( hd ); How to use the public key functions ----------------------------------- cipher/pubkey.c implements the interface to asymmetric encryption and signature functions. This is basically the same as with the symmetric counterparts, but due to their nature it is a little bit more complicated. [Give an example] Format of colon listings ======================== First an example: $ gpg --fixed-list-mode --with-colons --list-keys \ --with-fingerprint --with-fingerprint wk@gnupg.org pub:f:1024:17:6C7EE1B8621CC013:899817715:1055898235::m:::scESC: fpr:::::::::ECAF7590EB3443B5C7CF3ACB6C7EE1B8621CC013: uid:f::::::::Werner Koch : uid:f::::::::Werner Koch : sub:f:1536:16:06AD222CADF6A6E1:919537416:1036177416:::::e: fpr:::::::::CF8BCC4B18DE08FCD8A1615906AD222CADF6A6E1: sub:r:1536:20:5CE086B5B5A18FF4:899817788:1025961788:::::esc: fpr:::::::::AB059359A3B81F410FCFF97F5CE086B5B5A18FF4: The double --with-fingerprint prints the fingerprint for the subkeys too, --fixed-list-mode is themodern listing way printing dates in seconds since Epoch and does not merge the first userID with the pub record. 1. Field: Type of record pub = public key crt = X.509 certificate crs = X.509 certificate and private key available sub = subkey (secondary key) sec = secret key ssb = secret subkey (secondary key) uid = user id (only field 10 is used). uat = user attribute (same as user id except for field 10). sig = signature rev = revocation signature fpr = fingerprint: (fingerprint is in field 10) pkd = public key data (special field format, see below) grp = reserved for gpgsm rvk = revocation key 2. Field: A letter describing the calculated trust. This is a single letter, but be prepared that additional information may follow in some future versions. (not used for secret keys) o = Unknown (this key is new to the system) i = The key is invalid (e.g. due to a missing self-signature) d = The key has been disabled r = The key has been revoked e = The key has expired - = Unknown trust (i.e. no value assigned) q = Undefined trust '-' and 'q' may safely be treated as the same value for most purposes n = Don't trust this key at all m = There is marginal trust in this key f = The key is full trusted. u = The key is ultimately trusted; this is only used for keys for which the secret key is also available. 3. Field: length of key in bits. 4. Field: Algorithm: 1 = RSA 16 = ElGamal (encrypt only) 17 = DSA (sometimes called DH, sign only) 20 = ElGamal (sign and encrypt) (for other id's see include/cipher.h) 5. Field: KeyID either of 6. Field: Creation Date (in UTC) 7. Field: Key expiration date or empty if none. 8. Field: Used for serial number in crt records (used to be the Local-ID) 9. Field: Ownertrust (primary public keys only) This is a single letter, but be prepared that additional information may follow in some future versions. 10. Field: User-ID. The value is quoted like a C string to avoid control characters (the colon is quoted "\x3a"). This is not used with --fixed-list-mode in gpg. A UAT record puts the attribute subpacket count here, a space, and then the total attribute subpacket size. In gpgsm the issuer name comes here An FPR record stores the fingerprint here. The fingerprint of an revocation key is stored here. 11. Field: Signature class. This is a 2 digit hexnumber followed by either the letter 'x' for an exportable signature or the letter 'l' for a local-only signature. The class byte of an revocation key is also given here, 'x' and 'l' ist used the same way. 12. Field: Key capabilities: e = encrypt s = sign c = certify A key may have any combination of them. The primary key has in addition to these letters, uppercase version of the letter to denote the _usable_ capabilities of the entire key. 13. Field: Used in FPR records for S/MIME keys to store the fingerprint of the issuer certificate. This is useful to build the certificate path based on certificates stored in the local keyDB; it is only filled if the issue certificate is available. The advantage of using this value is that it is guaranteed to have been been build by the same lookup algorithm as gpgsm uses. For "uid" recods this lists the preferences n the sameway the -edit menu does. 14. Field Flag field used in the --edit menu output: All dates are displayed in the format yyyy-mm-dd unless you use the option --fixed-list-mode in which case they are displayed as seconds since Epoch. More fields may be added later, so parsers should be prepared for this. When parsing a number the parser should stop at the first non-number character so that additional information can later be added. If field 1 has the tag "pkd", a listing looks like this: pkd:0:1024:B665B1435F4C2 .... FF26ABB: ! ! !-- the value ! !------ for information number of bits in the value !--------- index (eg. DSA goes from 0 to 3: p,q,g,y) Format of the "--status-fd" output ================================== Every line is prefixed with "[GNUPG:] ", followed by a keyword with the type of the status line and a some arguments depending on the type (maybe none); an application should always be prepared to see more arguments in future versions. GOODSIG The signature with the keyid is good. For each signature only one of the three codes GOODSIG, BADSIG or ERRSIG will be emitted and they may be used as a marker for a new signature. The username is the primary one encoded in UTF-8 and %XX escaped. EXPSIG The signature with the keyid is good, but the signature is expired. The username is the primary one encoded in UTF-8 and %XX escaped. EXPKEYSIG The signature with the keyid is good, but the signature was made by an expired key. The username is the primary one encoded in UTF-8 and %XX escaped. BADSIG The signature with the keyid has not been verified okay. The username is the primary one encoded in UTF-8 and %XX escaped. ERRSIG \ It was not possible to check the signature. This may be caused by a missing public key or an unsupported algorithm. A RC of 4 indicates unknown algorithm, a 9 indicates a missing public key. The other fields give more information about this signature. sig_class is a 2 byte hex-value. VALIDSIG The signature with the keyid is good. This is the same as GOODSIG but has the fingerprint as the argument. Both status lines are emitted for a good signature. sig-timestamp is the signature creation time in seconds after the epoch. expire-timestamp is the signature expiration time in seconds after the epoch (zero means "does not expire"). SIG_ID This is emitted only for signatures of class 0 or 1 which have been verified okay. The string is a signature id and may be used in applications to detect replay attacks of signed messages. Note that only DLP algorithms give unique ids - others may yield duplicated ones when they have been created in the same second. ENC_TO The message is encrypted to this keyid. keytype is the numerical value of the public key algorithm, keylength is the length of the key or 0 if it is not known (which is currently always the case). NODATA No data has been found. Codes for what are: 1 - No armored data. 2 - Expected a packet but did not found one. 3 - Invalid packet found, this may indicate a non OpenPGP message. You may see more than one of these status lines. UNEXPECTED Unexpected data has been encountered 0 - not further specified 1 TRUST_UNDEFINED TRUST_NEVER TRUST_MARGINAL TRUST_FULLY TRUST_ULTIMATE For good signatures one of these status lines are emitted to indicate how trustworthy the signature is. The error token values are currently only emiited by gpgsm. SIGEXPIRED This is deprecated in favor of KEYEXPIRED. KEYEXPIRED The key has expired. expire-timestamp is the expiration time in seconds after the epoch. KEYREVOKED The used key has been revoked by its owner. No arguments yet. BADARMOR The ASCII armor is corrupted. No arguments yet. RSA_OR_IDEA The IDEA algorithms has been used in the data. A program might want to fallback to another program to handle the data if GnuPG failed. This status message used to be emitted also for RSA but this has been dropped after the RSA patent expired. However we can't change the name of the message. SHM_INFO SHM_GET SHM_GET_BOOL SHM_GET_HIDDEN GET_BOOL GET_LINE GET_HIDDEN GOT_IT NEED_PASSPHRASE Issued whenever a passphrase is needed. keytype is the numerical value of the public key algorithm or 0 if this is not applicable, keylength is the length of the key or 0 if it is not known (this is currently always the case). NEED_PASSPHRASE_SYM Issued whenever a passphrase for symmetric encryption is needed. MISSING_PASSPHRASE No passphrase was supplied. An application which encounters this message may want to stop parsing immediately because the next message will probably be a BAD_PASSPHRASE. However, if the application is a wrapper around the key edit menu functionality it might not make sense to stop parsing but simply ignoring the following BAD_PASSPHRASE. BAD_PASSPHRASE The supplied passphrase was wrong or not given. In the latter case you may have seen a MISSING_PASSPHRASE. GOOD_PASSPHRASE The supplied passphrase was good and the secret key material is therefore usable. DECRYPTION_FAILED The symmetric decryption failed - one reason could be a wrong passphrase for a symmetrical encrypted message. DECRYPTION_OKAY The decryption process succeeded. This means, that either the correct secret key has been used or the correct passphrase for a conventional encrypted message was given. The program itself may return an errorcode because it may not be possible to verify a signature for some reasons. NO_PUBKEY NO_SECKEY The key is not available IMPORTED The keyid and name of the signature just imported IMPORT_OK [] The key with the primary key's FINGERPRINT has been imported. Reason flags: 0 := Not actually changed 1 := Entirely new key. 2 := New user IDs 4 := New signatures 8 := New subkeys 16 := Contains private key. The flags may be ORed. IMPORT_PROBLEM [] Issued for each import failure. Reason codes are: 0 := "No specific reason given". 1 := "Invalid Certificate". 2 := "Issuer Certificate missing". 3 := "Certificate Chain too long". 4 := "Error storing certificate". IMPORT_RES Final statistics on import process (this is one long line) FILE_START Start processing a file . indicates the performed operation: 1 - verify 2 - encrypt 3 - decrypt FILE_DONE Marks the end of a file processing which has been started by FILE_START. BEGIN_DECRYPTION END_DECRYPTION Mark the start and end of the actual decryption process. These are also emitted when in --list-only mode. BEGIN_ENCRYPTION END_ENCRYPTION Mark the start and end of the actual encryption process. DELETE_PROBLEM reason_code Deleting a key failed. Reason codes are: 1 - No such key 2 - Must delete secret key first 3 - Ambigious specification PROGRESS what char cur total Used by the primegen and Public key functions to indicate progress. "char" is the character displayed with no --status-fd enabled, with the linefeed replaced by an 'X'. "cur" is the current amount done and "total" is amount to be done; a "total" of 0 indicates that the total amount is not known. 100/100 may be used to detect the end of operation. SIG_CREATED A signature has been created using these parameters. type: 'D' = detached 'C' = cleartext 'S' = standard (only the first character should be checked) class: 2 hex digits with the signature class KEY_CREATED A key has been created type: 'B' = primary and subkey 'P' = primary 'S' = subkey The fingerprint is one of the primary key for type B and P and the one of the subkey for S. SESSION_KEY : The session key used to decrypt the message. This message will only be emmited when the special option --show-session-key is used. The format is suitable to be passed to the option --override-session-key NOTATION_NAME NOTATION_DATA name and string are %XX escaped; the data may be splitted among several notation_data lines. USERID_HINT Give a hint about the user ID for a certain keyID. POLICY_URL string is %XX escaped BEGIN_STREAM END_STREAM Issued by pipemode. INV_RECP Issued for each unusable recipient. The reasons codes currently in use are: 0 := "No specific reason given". 1 := "Not Found" 2 := "Ambigious specification" 3 := "Wrong key usage" 4 := "Key revoked" 5 := "Key expired" 6 := "No CRL known" 7 := "CRL too old" 8 := "Policy mismatch" 9 := "Not a secret key" 10 := "Key not trusted" Note that this status is also used for gpgsm's SIGNER command where it relates to signer's of course. NO_RECP Issued when no recipients are usable. ALREADY_SIGNED Warning: This is experimental and might be removed at any time. TRUNCATED The output was truncated to MAXNO items. This status code is issued for certain external requests ERROR This is a generic error status message, it might be followed by error location specific data. and should not contain a space. ATTRIBUTE This is one long line issued for each attribute subpacket when an attribute packet is seen during key listing. is the fingerprint of the key. is the length of the attribute subpacket. is the attribute type (1==image). / indicates that this is the Nth indexed subpacket of count total subpackets in this attribute packet. and are from the self-signature on the attribute packet. If the attribute packet does not have a valid self-signature, then the timestamp is 0. are a bitwise OR of: 0x01 = this attribute packet is a primary uid 0x02 = this attribute packet is revoked 0x04 = this attribute packet is expired Key generation ============== Key generation shows progress by printing different characters to stderr: "." Last 10 Miller-Rabin tests failed "+" Miller-Rabin test succeeded "!" Reloading the pool with fresh prime numbers "^" Checking a new value for the generator "<" Size of one factor decreased ">" Size of one factor increased The prime number for ElGamal is generated this way: 1) Make a prime number q of 160, 200, 240 bits (depending on the keysize) 2) Select the length of the other prime factors to be at least the size of q and calculate the number of prime factors needed 3) Make a pool of prime numbers, each of the length determined in step 2 4) Get a new permutation out of the pool or continue with step 3 if we have tested all permutations. 5) Calculate a candidate prime p = 2 * q * p[1] * ... * p[n] + 1 6) Check that this prime has the correct length (this may change q if it seems not to be possible to make a prime of the desired length) 7) Check whether this is a prime using trial divisions and the Miller-Rabin test. 8) Continue with step 4 if we did not find a prime in step 7. 9) Find a generator for that prime. This algorithm is based on Lim and Lee's suggestion from the Crypto '97 proceedings p. 260. Unattended key generation ========================= This feature allows unattended generation of keys controlled by a parameter file. To use this feature, you use --gen-key together with --batch and feed the parameters either from stdin or from a file given on the commandline. The format of this file is as follows: o Text only, line length is limited to about 1000 chars. o You must use UTF-8 encoding to specify non-ascii characters. o Empty lines are ignored. o Leading and trailing spaces are ignored. o A hash sign as the first non white space character indicates a comment line. o Control statements are indicated by a leading percent sign, the arguments are separated by white space from the keyword. o Parameters are specified by a keyword, followed by a colon. Arguments are separated by white space. o The first parameter must be "Key-Type", control statements may be placed anywhere. o Key generation takes place when either the end of the parameter file is reached, the next "Key-Type" parameter is encountered or at the control statement "%commit" o Control statements: %echo Print . %dry-run Suppress actual key generation (useful for syntax checking). %commit Perform the key generation. An implicit commit is done at the next "Key-Type" parameter. %pubring %secring Do not write the key to the default or commandline given keyring but to . This must be given before the first commit to take place, duplicate specification of the same filename is ignored, the last filename before a commit is used. The filename is used until a new filename is used (at commit points) and all keys are written to that file. If a new filename is given, this file is created (and overwrites an existing one). Both control statements must be given. o The order of the parameters does not matter except for "Key-Type" which must be the first parameter. The parameters are only for the generated keyblock and parameters from previous key generations are not used. Some syntactically checks may be performed. The currently defined parameters are: Key-Type: | Starts a new parameter block by giving the type of the primary key. The algorithm must be capable of signing. This is a required parameter. Key-Length: Length of the key in bits. Default is 1024. Key-Usage: Space or comma delimited list of key usage, allowed values are "encrypt" and "sign". This is used to generate the key flags. Please make sure that the algorithm is capable of this usage. Subkey-Type: | This generates a secondary key. Currently only one subkey can be handled. Subkey-Length: Length of the subkey in bits. Default is 1024. Subkey-Usage: Similar to Key-Usage. Passphrase: If you want to specify a passphrase for the secret key, enter it here. Default is not to use any passphrase. Name-Real: Name-Comment: Name-Email: The 3 parts of a key. Remember to use UTF-8 here. If you don't give any of them, no user ID is created. Expire-Date: |([d|w|m|y]) Set the expiration date for the key (and the subkey). It may either be entered in ISO date format (2000-08-15) or as number of days, weeks, month or years. Without a letter days are assumed. Preferences: Set the cipher, hash, and compression preference values for this key. This expects the same type of string as "setpref" in the --edit menu. Revoker: : [sensitive] Add a designated revoker to the generated key. Algo is the public key algorithm of the designated revoker (i.e. RSA=1, DSA=17, etc.) Fpr is the fingerprint of the designated revoker. The optional "sensitive" flag marks the designated revoker as sensitive information. Only v4 keys may be designated revokers. Here is an example: $ cat >foo < ssb 1024g/8F70E2C0 2000-03-09 Layout of the TrustDB ===================== The TrustDB is built from fixed length records, where the first byte describes the record type. All numeric values are stored in network byte order. The length of each record is 40 bytes. The first record of the DB is always of type 1 and this is the only record of this type. FIXME: The layout changed, document it here. Record type 0: -------------- Unused record, can be reused for any purpose. Record type 1: -------------- Version information for this TrustDB. This is always the first record of the DB and the only one with type 1. 1 byte value 1 3 bytes 'gpg' magic value 1 byte Version of the TrustDB (2) 1 byte marginals needed 1 byte completes needed 1 byte max_cert_depth The three items are used to check whether the cached validity value from the dir record can be used. 1 u32 locked flags 1 u32 timestamp of trustdb creation 1 u32 timestamp of last modification which may affect the validity of keys in the trustdb. This value is checked against the validity timestamp in the dir records. 1 u32 timestamp of last validation (Used to keep track of the time, when this TrustDB was checked against the pubring) 1 u32 record number of keyhashtable 1 u32 first free record 1 u32 record number of shadow directory hash table It does not make sense to combine this table with the key table because the keyid is not in every case a part of the fingerprint. 1 u32 record number of the trusthashtbale Record type 2: (directory record) -------------- Informations about a public key certificate. These are static values which are never changed without user interaction. 1 byte value 2 1 byte reserved 1 u32 LID . (This is simply the record number of this record.) 1 u32 List of key-records (the first one is the primary key) 1 u32 List of uid-records 1 u32 cache record 1 byte ownertrust 1 byte dirflag 1 byte maximum validity of all the user ids 1 u32 time of last validity check. 1 u32 Must check when this time has been reached. (0 = no check required) Record type 3: (key record) -------------- Informations about a primary public key. (This is mainly used to lookup a trust record) 1 byte value 3 1 byte reserved 1 u32 LID 1 u32 next - next key record 7 bytes reserved 1 byte keyflags 1 byte pubkey algorithm 1 byte length of the fingerprint (in bytes) 20 bytes fingerprint of the public key (This is the value we use to identify a key) Record type 4: (uid record) -------------- Informations about a userid We do not store the userid but the hash value of the userid because that is sufficient. 1 byte value 4 1 byte reserved 1 u32 LID points to the directory record. 1 u32 next next userid 1 u32 pointer to preference record 1 u32 siglist list of valid signatures 1 byte uidflags 1 byte validity of the key calculated over this user id 20 bytes ripemd160 hash of the username. Record type 5: (pref record) -------------- This record type is not anymore used. 1 byte value 5 1 byte reserved 1 u32 LID; points to the directory record (and not to the uid record!). (or 0 for standard preference record) 1 u32 next 30 byte preference data Record type 6 (sigrec) ------------- Used to keep track of key signatures. Self-signatures are not stored. If a public key is not in the DB, the signature points to a shadow dir record, which in turn has a list of records which might be interested in this key (and the signature record here is one). 1 byte value 6 1 byte reserved 1 u32 LID points back to the dir record 1 u32 next next sigrec of this uid or 0 to indicate the last sigrec. 6 times 1 u32 Local_id of signatures dir or shadow dir record 1 byte Flag: Bit 0 = checked: Bit 1 is valid (we have a real directory record for this) 1 = valid is set (but may be revoked) Record type 8: (shadow directory record) -------------- This record is used to reserve a LID for a public key. We need this to create the sig records of other keys, even if we do not yet have the public key of the signature. This record (the record number to be more precise) will be reused as the dir record when we import the real public key. 1 byte value 8 1 byte reserved 1 u32 LID (This is simply the record number of this record.) 2 u32 keyid 1 byte pubkey algorithm 3 byte reserved 1 u32 hintlist A list of records which have references to this key. This is used for fast access to signature records which are not yet checked. Note, that this is only a hint and the actual records may not anymore hold signature records for that key but that the code cares about this. 18 byte reserved Record Type 10 (hash table) -------------- Due to the fact that we use fingerprints to lookup keys, we can implement quick access by some simple hash methods, and avoid the overhead of gdbm. A property of fingerprints is that they can be used directly as hash values. (They can be considered as strong random numbers.) What we use is a dynamic multilevel architecture, which combines hashtables, record lists, and linked lists. This record is a hashtable of 256 entries; a special property is that all these records are stored consecutively to make one big table. The hash value is simple the 1st, 2nd, ... byte of the fingerprint (depending on the indirection level). When used to hash shadow directory records, a different table is used and indexed by the keyid. 1 byte value 10 1 byte reserved n u32 recnum; n depends on the record length: n = (reclen-2)/4 which yields 9 for the current record length of 40 bytes. the total number of such record which makes up the table is: m = (256+n-1) / n which is 29 for a record length of 40. To look up a key we use the first byte of the fingerprint to get the recnum from this hashtable and look up the addressed record: - If this record is another hashtable, we use 2nd byte to index this hash table and so on. - if this record is a hashlist, we walk all entries until we found one a matching one. - if this record is a key record, we compare the fingerprint and to decide whether it is the requested key; Record type 11 (hash list) -------------- see hash table for an explanation. This is also used for other purposes. 1 byte value 11 1 byte reserved 1 u32 next next hash list record n times n = (reclen-5)/5 1 u32 recnum For the current record length of 40, n is 7 Record type 254 (free record) --------------- All these records form a linked list of unused records. 1 byte value 254 1 byte reserved (0) 1 u32 next_free Packet Headers =============== GNUPG uses PGP 2 packet headers and also understands OpenPGP packet header. There is one enhancement used with the old style packet headers: CTB bits 10, the "packet-length length bits", have values listed in the following table: 00 - 1-byte packet-length field 01 - 2-byte packet-length field 10 - 4-byte packet-length field 11 - no packet length supplied, unknown packet length As indicated in this table, depending on the packet-length length bits, the remaining 1, 2, 4, or 0 bytes of the packet structure field are a "packet-length field". The packet-length field is a whole number field. The value of the packet-length field is defined to be the value of the whole number field. A value of 11 is currently used in one place: on compressed data. That is, a compressed data block currently looks like , where , binary 10 1000 11, is an indefinite-length packet. The proper interpretation is "until the end of the enclosing structure", although it should never appear outermost (where the enclosing structure is a file). + This will be changed with another version, where the new meaning of + the value 11 (see below) will also take place. + + A value of 11 for other packets enables a special length encoding, + which is used in case, where the length of the following packet can + not be determined prior to writing the packet; especially this will + be used if large amounts of data are processed in filter mode. + + It works like this: After the CTB (with a length field of 11) a + marker field is used, which gives the length of the following datablock. + This is a simple 2 byte field (MSB first) containing the amount of data + following this field, not including this length field. After this datablock + another length field follows, which gives the size of the next datablock. + A value of 0 indicates the end of the packet. The maximum size of a + data block is limited to 65534, thereby reserving a value of 0xffff for + future extensions. These length markers must be inserted into the data + stream just before writing the data out. + + This 2 byte field is large enough, because the application must buffer + this amount of data to prepend the length marker before writing it out. + Data block sizes larger than about 32k doesn't make any sense. Note + that this may also be used for compressed data streams, but we must use + another packet version to tell the application that it can not assume, + that this is the last packet. GNU extensions to the S2K algorithm =================================== S2K mode 101 is used to identify these extensions. After the hash algorithm the 3 bytes "GNU" are used to make clear that these are extensions for GNU, the next bytes gives the GNU protection mode - 1000. Defined modes are: 1001 - do not store the secret part at all Usage of gdbm files for keyrings ================================ The key to store the keyblock is its fingerprint, other records are used for secondary keys. Fingerprints are always 20 bytes where 16 bit fingerprints are appended with zero. The first byte of the key gives some information on the type of the key. 1 = key is a 20 bit fingerprint (16 bytes fpr are padded with zeroes) data is the keyblock 2 = key is the complete 8 byte keyid data is a list of 20 byte fingerprints 3 = key is the short 4 byte keyid data is a list of 20 byte fingerprints 4 = key is the email address data is a list of 20 byte fingerprints Data is prepended with a type byte: 1 = keyblock 2 = list of 20 byte padded fingerprints 3 = list of list fingerprints (but how to we key them?) Pipemode ======== This mode can be used to perform multiple operations with one call to gpg. It comes handy in cases where you have to verify a lot of signatures. Currently we support only detached signatures. This mode is a kludge to avoid running gpg n daemon mode and using Unix Domain Sockets to pass the data to it. There is no easy portable way to do this under Windows, so we use plain old pipes which do work well under Windows. Because there is no way to signal multiple EOFs in a pipe we have to embed control commands in the data stream: We distinguish between a data state and a control state. Initially the system is in data state but it won't accept any data. Instead it waits for transition to control state which is done by sending a single '@' character. While in control state the control command os expected and this command is just a single byte after which the system falls back to data state (but does not necesary accept data now). The simplest control command is a '@' which just inserts this character into the data stream. Here is the format we use for detached signatures: "@<" - Begin of new stream "@B" - Detached signature follows. This emits a control packet (1,'B') "@t" - Signed text follows. This emits the control packet (2, 'B') "@." - End of operation. The final control packet forces signature verification "@>" - End of stream Other Notes =========== * For packet version 3 we calculate the keyids this way: RSA := low 64 bits of n ELGAMAL := build a v3 pubkey packet (with CTB 0x99) and calculate a rmd160 hash value from it. This is used as the fingerprint and the low 64 bits are the keyid. * Revocation certificates consist only of the signature packet; "import" knows how to handle this. The rationale behind it is to keep them small. Keyserver Message Format ========================= The keyserver may be contacted by a Unix Domain socket or via TCP. The format of a request is: ==== command-tag "Content-length:" digits CRLF ======= Where command-tag is NOOP GET PUT DELETE The format of a response is: ====== "GNUPG/1.0" status-code status-text "Content-length:" digits CRLF ============ followed by bytes of data Status codes are: o 1xx: Informational - Request received, continuing process o 2xx: Success - The action was successfully received, understood, and accepted o 4xx: Client Error - The request contains bad syntax or cannot be fulfilled o 5xx: Server Error - The server failed to fulfill an apparently valid request Documentation on HKP (the http keyserver protocol): A minimalistic HTTP server on port 11371 recognizes a GET for /pks/lookup. The standard http URL encoded query parameters are this (always key=value): - op=index (like pgp -kv), op=vindex (like pgp -kvv) and op=get (like pgp -kxa) - search=. This is a list of words that must occur in the key. The words are delimited with space, points, @ and so on. The delimiters are not searched for and the order of the words doesn't matter (but see next option). - exact=on. This switch tells the hkp server to only report exact matching keys back. In this case the order and the "delimiters" are important. - fingerprint=on. Also reports the fingerprints when used with 'index' or 'vindex' The keyserver also recognizes http-POSTs to /pks/add. Use this to upload keys. A better way to do this would be a request like: /pks/lookup/?op= This can be implemented using Hurd's translator mechanism. However, I think the whole key server stuff has to be re-thought; I have some ideas and probably create a white paper. GnuPG Frequently Asked Questions Version: 1.5.8 Last-Modified: Oct 8, 2002 Maintained-by: David D. Scribner, This is the GnuPG FAQ. The latest HTML version is available here. The index is generated automatically, so there may be errors. Not all questions may be in the section they belong to. Suggestions about how to improve the structure of this FAQ are welcome. Please send additions and corrections to the maintainer. It would be most convenient if you could provide the answer to be included here as well. Your help is very much appreciated! Please, don't send message like "This should be a FAQ - what's the answer?". If it hasn't been asked before, it isn't a FAQ. In that case you could search in the mailing list archive. 1. GENERAL 1.1) What is GnuPG? 1.2) Is GnuPG compatible with PGP? 1.3) Is GnuPG free to use for personal or commercial use? 2. SOURCES of INFORMATION 2.1) Where can I find more information on GnuPG? 2.2) Where do I get GnuPG? 3. INSTALLATION 3.1) Which OSes does GnuPG run on? 3.2) Which random data gatherer should I use? 3.3) How do I include support for RSA and IDEA? 4. USAGE 4.1) What is the recommended key size? 4.2) Why does it sometimes take so long to create keys? 4.3) And it really takes long when I work on a remote system. Why? 4.4) What is the difference between options and commands? 4.5) I can't delete a user ID on my secret keyring because it has already been deleted on my public keyring. What can I do? 4.6) I can't delete my secret key because the public key disappeared. What can I do? 4.7) What are trust, validity and ownertrust? 4.8) How do I sign a patch file? 4.9) Where is the "encrypt-to-self" option? 4.10) How can I get rid of the Version and Comment headers in armored messages? 4.11) What does the "You are using the xxxx character set." mean? 4.12) How can a get list of key IDs used to encrypt a message? 4.13) I can't decrypt my symmetrical-only (-c) encrypted messages with a new version of GnuPG. 4.14) How can I use GnuPG in an automated environment? 4.15) Which email-client can I use with GnuPG? 4.16) Can't we have a gpg library? 4.17) I have successfully generated a revocation certificate, but I don't understand how to send it to the key servers. 4.18) How do I put my keyring in a different directory? 5. COMPATIBILITY ISSUES 5.1) How can I encrypt a message with GnuPG so that PGP is able to decrypt it? 5.2) How do I migrate from PGP 2.x to GnuPG? 5.3) (removed) 5.4) Why is PGP 5.x not able to encrypt messages with some keys? 5.5) Why is PGP 5.x not able to verify my messages? 5.6) How do I transfer owner trust values from PGP to GnuPG? 5.7) PGP does not like my secret key. 5.8) I just installed the most recent version of GnuPG and don't have a ~/.gnupg/options file. Is this missing from the installation? 6. PROBLEMS and ERROR MESSAGES 6.1) Why do I get "gpg: Warning: using insecure memory!" 6.2) Large File Support doesn't work ... 6.3) In the edit menu the trust values are not displayed correctly after signing uids. Why? 6.4) What does "skipping pubkey 1: already loaded" mean? 6.5) GnuPG 1.0.4 doesn't create ~/.gnupg ... 6.6) An ElGamal signature does not verify anymore since version 1.0.2 ... 6.7) Old versions of GnuPG can't verify ElGamal signatures 6.8) When I use --clearsign, the plain text has sometimes extra dashes in it - why? 6.9) What is the thing with "can't handle multiple signatures"? 6.10) If I submit a key to a keyserver, nothing happens ... 6.11) I get "gpg: waiting for lock ..." 6.12) Older gpg binaries (e.g., 1.0) have problems with keys from newer gpg binaries ... 6.13) With 1.0.4, I get "this cipher algorithm is deprecated ..." 6.14) Some dates are displayed as ????-??-??. Why? 6.15) I still have a problem. How do I report a bug? 6.16) Why doesn't GnuPG support X.509 certificates? 6.17) Why do national characters in my user ID look funny? 6.18) I get 'sed' errors when running ./configure on Mac OS X ... 6.19) Why does GnuPG 1.0.6 bail out on keyrings used with 1.0.7? 6.20) I've upgraded to GnuPG version 1.0.7 and now it takes longer to load my keyrings. What can I do? 7. ADVANCED TOPICS 7.1) How does this whole thing work? 7.2) Why are some signatures with an ELG-E key valid? 7.3) How does the whole trust thing work? 7.4) What kind of output is this: "key C26EE891.298, uid 09FB: ...."? 7.5) How do I interpret some of the informational outputs? 7.6) Are the header lines of a cleartext signature part of the signed material? 7.7) What is the list of preferred algorithms? 7.8) How do I change the list of preferred algorithms? 8. ACKNOWLEDGEMENTS 1. GENERAL 1.1) What is GnuPG? GnuPG stands for GNU Privacy Guard and is GNU's tool for secure communication and data storage. It can be used to encrypt data and to create digital signatures. It includes an advanced key management facility and is compliant with the proposed OpenPGP Internet standard as described in RFC 2440. As such, it is aimed to be compatible with PGP from NAI, Inc. 1.2) Is GnuPG compatible with PGP? In general, yes. GnuPG and newer PGP releases should be implementing the OpenPGP standard. But there are some interoperability problems. See question 5.1 for details. 1.3) Is GnuPG free to use for personal or commercial use? Yes. GnuPG is part of the GNU family of tools and applications built and provided in accordance with the Free Software Foundation (FSF) General Public License (GPL). Therefore the software is free to copy, use, modify and distribute in accordance with that license. Please read the file titled COPYING that accompanies the application for more information. 2. SOURCES of INFORMATION 2.1) Where can I find more information on GnuPG? On-line resources: The documentation page is located at . Also, have a look at the HOWTOs and the GNU Privacy Handbook (GPH, available in English, Spanish and Russian). The latter provides a detailed user's guide to GnuPG. You'll also find a document about how to convert from PGP 2.x to GnuPG. At you'll find an online archive of the GnuPG mailing lists. Most interesting should be gnupg-users for all user-related issues and gnupg-devel if you want to get in touch with the developers. In addition, searchable archives can be found on MARC, e.g.: gnupg-users: , gnupg-devel: . *PLEASE:* Before posting to a list, read this FAQ and the available documentation. In addition, search the list archive - maybe your question has already been discussed. This way you help people focus on topics that have not yet been resolved. The GnuPG source distribution contains a subdirectory: ./doc where some additional documentation is located (mainly interesting for hackers, not the casual user). 2.2) Where do I get GnuPG? You can download the GNU Privacy Guard from its primary FTP server or from one of the mirrors: The current stable version is 1.2.x. Please upgrade to this version as it includes additional features, functions and security fixes that may not have existed in prior versions. 3. INSTALLATION 3.1) Which OSes does GnuPG run on? It should run on most Unices as well as Windows versions (including Windows NT/2000) and Macintosh OS/X. A list of OSes reported to be OK is presented at: 3.2) Which random data gatherer should I use? "Good" random numbers are crucial for the security of your encryption. Different operating systems provide a variety of more or less quality random data. Linux and *BSD provide kernel generated random data through /dev/random - this should be the preferred choice on these systems. Also Solaris users with the SUNWski package installed have a /dev/random. In these cases, use the configure option: --enable-static-rnd=linux In addition, there's also the kernel random device by Andi Maier , but it's still beta. Use at your own risk! On other systems, the Entropy Gathering Daemon (EGD) is a good choice. It is a perl-daemon that monitors system activity and hashes it into random data. See the download page to obtain EGD. Use: --enable-static-rnd=egd here. If the above options do not work, you can use the random number generator "unix". This is *very* slow and should be avoided. The random quality isn't very good so don't use it on sensitive data. 3.3) How do I include support for RSA and IDEA? RSA is included as of GnuPG version 1.0.3. The official GnuPG distribution does not contain IDEA due to a patent restriction. The patent does not expire before 2007 so don't expect official support before then. However, there is an unofficial module to include it even in earlier versions of GnuPG. It's available from . Look for: idea.c Compilation directives are in the headers of these files. You will then need to add the following line to your ~/.gnupg/options file: load-extension idea 4. USAGE 4.1) What is the recommended key size? 1024 bit for DSA signatures; even for plain ElGamal signatures. This is sufficient as the size of the hash is probably the weakest link if the key size is larger than 1024 bits. Encryption keys may have greater sizes, but you should then check the fingerprint of this key: gpg --fingerprint As for the key algorithms, you should stick with the default (i.e., DSA signature and ElGamal encryption). An ElGamal signing key has the following disadvantages: the signature is larger, it is hard to create such a key useful for signatures which can withstand some real world attacks, you don't get any extra security compared to DSA, and there might be compatibility problems with certain PGP versions. It has only been introduced because at the time it was not clear whether there was a patent on DSA. 4.2) Why does it sometimes take so long to create keys? The problem here is that we need a lot of random bytes and for that we (on Linux the /dev/random device) must collect some random data. It is really not easy to fill the Linux internal entropy buffer; I talked to Ted Ts'o and he commented that the best way to fill the buffer is to play with your keyboard. Good security has its price. What I do is to hit several times on the shift, control, alternate, and caps lock keys, because these keys do not produce output to the screen. This way you get your keys really fast (it's the same thing PGP2 does). Another problem might be another program which eats up your random bytes (a program (look at your daemons) that reads from /dev/random). 4.3) And it really takes long when I work on a remote system. Why? Don't do this at all! You should never create keys or even use GnuPG on a remote system because you normally have no physical control over your secret key ring (which is in most cases vulnerable to advanced dictionary attacks) - I strongly encourage everyone to only create keys on a local computer (a disconnected laptop is probably the best choice) and if you need it on your connected box (I know, we all do this) be sure to have a strong password for both your account and for your secret key, and that you can trust your system administrator. When I check GnuPG on a remote system via ssh (I have no Alpha here) ;-) I have the same problem. It takes a *very* long time to create the keys, so I use a special option, --quick-random, to generate insecure keys which are only good for some tests. 4.4) What is the difference between options and commands? If you do a 'gpg --help', you will get two separate lists. The first is a list of commands. The second is a list of options. Whenever you run GPG, you *must* pick exactly one command (with one exception, see below). You *may* pick one or more options. The command should, just by convention, come at the end of the argument list, after all the options. If the command takes a file (all the basic ones do), the filename comes at the very end. So the basic way to run gpg is: gpg [--option something] [--option2] [--option3 something] --command file Some options take arguments. For example, the --output option (which can be abbreviated as -o) is an option that takes a filename. The option's argument must follow immediately after the option itself, otherwise gpg doesn't know which option the argument is supposed to paired with. As an option, --output and its filename must come before the command. The --recipient (-r) option takes a name or keyID to encrypt the message to, which must come right after the -r argument. The --encrypt (or -e) command comes after all the options and is followed by the file you wish to encrypt. Therefore in this example the command-line issued would be: gpg -r alice -o secret.txt -e test.txt If you write the options out in full, it is easier to read: gpg --recipient alice --output secret.txt --encrypt test.txt If you're encrypting to a file with the extension ".txt", then you'd probably expect to see ASCII-armored text in the file (not binary), so you need to add the --armor (-a) option, which doesn't take any arguments: gpg --armor --recipient alice --output secret.txt --encrypt test.txt If you imagine square brackets around the optional parts, it becomes a bit clearer: gpg [--armor] [--recipient alice] [--output secret.txt] --encrypt test.txt The optional parts can be rearranged any way you want: gpg --output secret.txt --recipient alice --armor --encrypt test.txt If your filename begins with a hyphen (e.g. "-a.txt"), GnuPG assumes this is an option and may complain. To avoid this you have to either use "./-a.txt", or stop the option and command processing with two hyphens: "-- -a.txt". *The exception to using only one command:* signing and encrypting at the same time. For this you can combine both commands, such as in: gpg [--options] --sign --encrypt foo.txt 4.5) I can't delete a user ID on my secret keyring because it has already been deleted on my public keyring. What can I do? Because you can only select from the public key ring, there is no direct way to do this. However it is not very complicated to do anyway. Create a new user ID with exactly the same name and you will see that there are now two identical user IDs on the secret ring. Now select this user ID and delete it. Both user IDs will be removed from the secret ring. 4.6) I can't delete my secret key because the public key disappeared. What can I do? To select a key a search is always done on the public keyring, therefore it is not possible to select a secret key without having the public key. Normally it shoud never happen that the public key got lost but the secret key is still available. The reality is different, so GnuPG implements a special way to deal with it: Simply use the long keyID to specify the key to delete, which can be obtained by using the --with-colons options (it is the fifth field in the lines beginning with "sec"). 4.7) What are trust, validity and ownertrust? With GnuPG, the term "ownertrust" is used instead of "trust" to help clarify that this is the value you have assigned to a key to express how much you trust the owner of this key to correctly sign (and thereby introduce) other keys. The "validity", or calculated trust, is a value which indicates how much GnuPG considers a key as being valid (that it really belongs to the one who claims to be the owner of the key). For more information on trust values see the chapter "The Web of Trust" in The GNU Privacy Handbook. 4.8) How do I sign a patch file? Use "gpg --clearsign --not-dash-escaped ...". The problem with --clearsign is that all lines starting with a dash are quoted with "- "; obviously diff produces many lines starting with a dash and these are then quoted and that is not good for a patch ;-). To use a patch file without removing the cleartext signature, the special option --not-dash-escaped may be used to suppress generation of these escape sequences. You should not mail such a patch because spaces and line endings are also subject to the signature and a mailer may not preserve these. If you want to mail a file you can simply sign it using your MUA (Mail User Agent). 4.9) Where is the "encrypt-to-self" option? Use "--encrypt-to your_keyID". You can use more than one of these options. To temporarily override the use of this additional key, you can use the option "--no-encrypt-to". 4.10) How can I get rid of the Version and Comment headers in armored messages? Use "--no-version --comment ''". Note that the left over blank line is required by the protocol. 4.11) What does the "You are using the xxxx character set." mean? This note is printed when UTF-8 mapping has to be done. Make sure that the displayed character set is the one you have activated on your system. Since "iso-8859-1" is the character set most used, this is the default. You can change the charset with the option "--charset". It is important that your active character set matches the one displayed - if not, restrict yourself to plain 7 bit ASCII and no mapping has to be done. 4.12) How can a get list of key IDs used to encrypt a message? gpg --batch --decrypt --list-only --status-fd 1 2>/dev/null | \ awk '/^\[GNUPG:\] ENC_TO / { print $3 }' 4.13) I can't decrypt my symmetrical-only (-c) encrypted messages with a new version of GnuPG. There was a bug in GnuPG versions prior to 1.0.1 which affected messages only if 3DES or Twofish was used for symmetric-only encryption (this has never been the default). The bug has been fixed, but to enable decryption of old messages you should run gpg with the option "--emulate-3des-s2k-bug", decrypt the message and encrypt it again without this option. The option will be removed in version 1.1 when released, so please re-encrypt any affected messages now. 4.14) How can I use GnuPG in an automated environment? You should use the option --batch and don't use passphrases as there is usually no way to store it more securely than on the secret keyring itself. The suggested way to create keys for an automated environment is: On a secure machine: If you want to do automatic signing, create a signing subkey for your key (use the interactive key editing menu by issueing the command 'gpg --edit-key keyID', enter "addkey" and select the DSA key type). Make sure that you use a passphrase (needed by the current implementation). gpg --export-secret-subkeys --no-comment foo >secring.auto Copy secring.auto and the public keyring to a test directory. Change to this directory. gpg --homedir . --edit foo and use "passwd" to remove the passphrase from the subkeys. You may also want to remove all unused subkeys. Copy secring.auto to a floppy and carry it to the target box. On the target machine: Install secring.auto as the secret keyring. Now you can start your new service. It's also a good idea to install an intrusion detection system so that you hopefully get a notice of an successful intrusion, so that you in turn can revoke all the subkeys installed on that machine and install new subkeys. 4.15) Which email-client can I use with GnuPG? Using GnuPG to encrypt email is one of the most popular uses. Several mail clients or mail user agents (MUAs) support GnuPG to varying degrees. Simplifying a bit, there are two ways mail can be encrypted with GnuPG: the "old style" ASCII armor (i.e. cleartext encryption), and RFC 2015 style (previously PGP/MIME, now OpenPGP). The latter has full MIME support. Some MUAs support only one of them, so whichever you actually use depends on your needs as well as the capabilities of your addressee. As well, support may be native to the MUA, or provided via "plug-ins" or external tools. The following list is not exhaustive: MUA OpenPGP ASCII How? (N,P,T) --------------------------------------------------------------- Calypso N Y P (Unixmail) Elm N Y T (mailpgp,morepgp) Elm ME+ N Y N Emacs/Gnus Y Y T (Mailcrypt,gpg.el) Emacs/Mew Y Y N Emacs/VM N Y T (Mailcrypt) Evolution Y Y N Exmh Y Y N GNUMail.app Y Y P (PGPBundle) GPGMail Y Y N KMail (<=1.4.x) N Y N KMail (1.5.x) Y(P) Y(N) P/N Mozilla Y Y P (Enigmail) Mulberry Y Y P Mutt Y Y N Sylpheed Y Y N Sylpheed-claws Y Y N TkRat Y Y N XEmacs/Gnus Y Y T (Mailcrypt) XEmacs/Mew Y Y N XEmacs/VM N Y T (Mailcrypt) XFmail Y Y N N - Native, P - Plug-in, T - External Tool The following table lists proprietary MUAs. The GNU Project suggests against the use of these programs, but they are listed for interoperability reasons for your convenience. MUA OpenPGP ASCII How? (N,P,T) --------------------------------------------------------------- Apple Mail Y Y P (GPGMail) Becky2 Y Y P (BkGnuPG) Eudora Y Y P (EuroraGPG) Eudora Pro Y Y P (EudoraGPG) Lotus Notes N Y P Netscape 4.x N Y P Netscape 7.x Y Y P (Enigmail) Novell Groupwise N Y P Outlook N Y P (G-Data) Outlook Express N Y P (GPGOE) Pegasus N Y P (QDPGP,PM-PGP) Pine N Y T (pgpenvelope,(gpg|pgp)4pine) Postme N Y P (GPGPPL) The Bat! N Y P (Ritlabs) Good overviews of OpenPGP-support can be found at: , and . 4.16) Can't we have a gpg library? This has been frequently requested. However, the current viewpoint of the GnuPG maintainers is that this would lead to several security issues and will therefore not be implemented in the foreseeable future. However, for some areas of application gpgme could do the trick. You'll find it at . 4.17) I have successfully generated a revocation certificate, but I don't understand how to send it to the key servers. Most keyservers don't accept a 'bare' revocation certificate. You have to import the certificate into gpg first: gpg --import my-revocation.asc then send the revoked key to the keyservers: gpg --keyserver certserver.pgp.com --send-keys mykeyid (or use a keyserver web interface for this). 4.18) How do I put my keyring in a different directory? GnuPG keeps several files in a special homedir directory. These include the options file, pubring.gpg, secring.gpg, trustdb.gpg, and others. GnuPG will always create and use these files. On unices, the homedir is usually ~/.gnupg; on Windows "C:\gnupg\". If you want to put your keyrings somewhere else, use: --homedir /my/path/ to make GnuPG create all its files in that directory. Your keyring will be "/my/path/pubring.gpg". This way you can store your secrets on a floppy disk. Don't use "--keyring" as its purpose is to specify additional keyring files. 5. COMPATIBILITY ISSUES 5.1) How can I encrypt a message with GnuPG so that PGP is able to decrypt it? It depends on the PGP version. PGP 2.x You can't do that because PGP 2.x normally uses IDEA which is not supported by GnuPG as it is patented (see 3.3), but if you have a modified version of PGP you can try this: gpg --rfc1991 --cipher-algo 3des ... Please don't pipe the data to encrypt to gpg but provide it using a filename; otherwise, PGP 2 will not be able to handle it. As for conventional encryption, you can't do this for PGP 2. PGP 5.x and higher You need to provide two additional options: --compress-algo 1 --cipher-algo cast5 You may also use "3des" instead of "cast5", and "blowfish" does not work with all versions of PGP 5. You may also want to put: compress-algo 1 into your ~/.gnupg/options file - this does not affect normal GnuPG operation. This applies to conventional encryption as well. 5.2) How do I migrate from PGP 2.x to GnuPG? PGP 2 uses the RSA and IDEA encryption algorithms. Whereas the RSA patent has expired and RSA is included as of GnuPG 1.0.3, the IDEA algorithm is still patented until 2007. Under certain conditions you may use IDEA even today. In that case, you may refer to Question 3.3 about how to add IDEA support to GnuPG and read to perform the migration. 5.3) (removed) (empty) 5.4) Why is PGP 5.x not able to encrypt messages with some keys? PGP, Inc. refuses to accept ElGamal keys of type 20 even for encryption. They only support type 16 (which is identical at least for decryption). To be more inter-operable, GnuPG (starting with version 0.3.3) now also uses type 16 for the ElGamal subkey which is created if the default key algorithm is chosen. You may add a type 16 ElGamal key to your public key, which is easy as your key signatures are still valid. 5.5) Why is PGP 5.x not able to verify my messages? PGP 5.x does not accept v4 signatures for data material but OpenPGP requests generation of v4 signatures for all kind of data, that's why GnuPG defaults to them. Use the option "--force-v3-sigs" to generate v3 signatures for data. 5.6) How do I transfer owner trust values from PGP to GnuPG? There is a script in the tools directory to help you. After you have imported the PGP keyring you can give this command: $ lspgpot pgpkeyring | gpg --import-ownertrust where pgpkeyring is the original keyring and not the GnuPG keyring you might have created in the first step. 5.7) PGP does not like my secret key. Older PGPs probably bail out on some private comment packets used by GnuPG. These packets are fully in compliance with OpenPGP; however PGP is not really OpenPGP aware. A workaround is to export the secret keys with this command: $ gpg --export-secret-keys --no-comment -a your-key-id Another possibility is this: by default, GnuPG encrypts your secret key using the Blowfish symmetric algorithm. Older PGPs will only understand 3DES, CAST5, or IDEA symmetric algorithms. Using the following method you can re-encrypt your secret gpg key with a different algo: $ gpg --s2k-cipher-algo=CAST5 --s2k-digest-algo=SHA1 \ --compress-algo=1 --edit-key Then use passwd to change the password (just change it to the same thing, but it will encrypt the key with CAST5 this time). Now you can export it and PGP should be able to handle it. For PGP 6.x the following options work to export a key: $ gpg --s2k-cipher-algo 3des --compress-algo 1 --rfc1991 \ --export-secret-keys 5.8) I just installed the most recent version of GnuPG and don't have a ~/.gnupg/options file. Is this missing from the installation? No. The ~/.gnupg/options file has been renamed to ~/.gnupg/conf for new installs as of version 1.1.92. If an existing ~/.gnupg/options file is found during an upgrade it will still be used, but this change was required to have a more consistent naming scheme with forthcoming tools. 6. PROBLEMS and ERROR MESSAGES 6.1) Why do I get "gpg: Warning: using insecure memory!" On many systems this program should be installed as setuid(root). This is necessary to lock memory pages. Locking memory pages prevents the operating system from writing them to disk and thereby keeping your secret keys really secret. If you get no warning message about insecure memory your operating system supports locking without being root. The program drops root privileges as soon as locked memory is allocated. To setuid(root) permissions on the gpg binary you can either use: chmod u+s /path/to/gpg or chmod 4755 /path/to/gpg Some refrain from using setuid(root) unless absolutely required for security reasons. Please check with your system administrator if you are not able to make these determinations yourself. On UnixWare 2.x and 7.x you should install GnuPG with the 'plock' privilege to get the same effect: filepriv -f plock /path/to/gpg If you can't or don't want to install GnuPG setuid(root), you can use the option "--no-secmem-warning" or put: no-secmem-warning in your ~/.gnupg/options file (this disables the warning). On some systems (e.g., Windows) GnuPG does not lock memory pages and older GnuPG versions (<=1.0.4) issue the warning: gpg: Please note that you don't have secure memory This warning can't be switched off by the above option because it was thought to be too serious an issue. However, it confused users too much, so the warning was eventually removed. 6.2) Large File Support doesn't work ... LFS works correctly in post-1.0.4 versions. If configure doesn't detect it, try a different (i.e., better) compiler. egcs 1.1.2 works fine, other gccs sometimes don't. BTW, several compilation problems of GnuPG 1.0.3 and 1.0.4 on HP-UX and Solaris were due to broken LFS support. 6.3) In the edit menu the trust values are not displayed correctly after signing uids. Why? This happens because some information is stored immediately in the trustdb, but the actual trust calculation can be done after the save command. This is a "not easy to fix" design bug which will be addressed in some future release. 6.4) What does "skipping pubkey 1: already loaded" mean? As of GnuPG 1.0.3, the RSA algorithm is included. If you still have a "load-extension rsa" in your options file, the above message occurs. Just remove the load command from the options file. 6.5) GnuPG 1.0.4 doesn't create ~/.gnupg ... That's a known bug, already fixed in newer versions. 6.6) An ElGamal signature does not verify anymore since version 1.0.2 ... Use the option --emulate-md-encode-bug. 6.7) Old versions of GnuPG can't verify ElGamal signatures Update to GnuPG 1.0.2 or newer. 6.8) When I use --clearsign, the plain text has sometimes extra dashes in it - why? This is called dash-escaped text and is required by OpenPGP. It always happens when a line starts with a dash ("-") and is needed to make the lines that structure signature and text (i.e., "-----BEGIN PGP SIGNATURE-----") to be the only lines that start with two dashes. If you use GnuPG to process those messages, the extra dashes are removed. Good mail clients remove those extra dashes when displaying such a message. 6.9) What is the thing with "can't handle multiple signatures"? Due to different message formats GnuPG is not always able to split a file with multiple signatures unambiguously into its parts. This error message informs you that there is something wrong with the input. The only way to have multiple signatures in a file is by using the OpenPGP format with one-pass-signature packets (which is GnuPG's default) or the cleartext signed format. 6.10) If I submit a key to a keyserver, nothing happens ... You are most likely using GnuPG 1.0.2 or older on Windows. That's feature isn't yet implemented, but it's a bug not to say it. Newer versions issue a warning. Upgrade to 1.0.4 or newer. 6.11) I get "gpg: waiting for lock ..." A previous instance of gpg has most likely exited abnormally and left a lock file. Go to ~/.gnupg and look for .*.lock files and remove them. 6.12) Older gpg binaries (e.g., 1.0) have problems with keys from newer gpg binaries ... As of 1.0.3, keys generated with gpg are created with preferences to TWOFISH (and AES since 1.0.4) and that also means that they have the capability to use the new MDC encryption method. This will go into OpenPGP soon, and is also suppoted by PGP 7. This new method avoids a (not so new) attack on all email encryption systems. This in turn means that pre-1.0.3 gpg binaries have problems with newer keys. Because of security and bug fixes, you should keep your GnuPG installation in a recent state anyway. As a workaround, you can force gpg to use a previous default cipher algo by putting: cipher-algo cast5 into your options file. 6.13) With 1.0.4, I get "this cipher algorithm is deprecated ..." If you just generated a new key and get this message while encrypting, you've witnessed a bug in 1.0.4. It uses the new AES cipher Rijndael that is incorrectly being referred as "deprecated". Ignore this warning, more recent versions of gpg are corrected. 6.14) Some dates are displayed as ????-??-??. Why? Due to constraints in most libc implementations, dates beyond 2038-01-19 can't be displayed correctly. 64-bit OSes are not affected by this problem. To avoid printing wrong dates, GnuPG instead prints some question marks. To see the correct value, you can use the options --with-colons and --fixed-list-mode. 6.15) I still have a problem. How do I report a bug? Are you sure that it's not been mentioned somewhere on the mailing lists? Did you have a look at the bug list (you'll find a link to the list of reported bugs on the documentation page). If you're not sure about it being a bug, you can send mail to the gnupg-devel list. Otherwise, use the GUUG bug tracking system . 6.16) Why doesn't GnuPG support X.509 certificates? GnuPG, first and foremost, is an implementation of the OpenPGP standard (RFC 2440), which is a competing infrastructure, different from X.509. They are both public-key cryptosystems, but how the public keys are actually handled is different. 6.17) Why do national characters in my user ID look funny? According to OpenPGP, GnuPG encodes user ID strings (and other things) using UTF-8. In this encoding of Unicode, most national characters get encoded as two- or three-byte sequences. For example, å (0xE5 in ISO-8859-1) becomes Ã¥ (0xC3, 0xA5). This might also be the reason why keyservers can't find your key. 6.18) I get 'sed' errors when running ./configure on Mac OS X ... This will be fixed after GnuPG has been upgraded to autoconf-2.50. Until then, find the line setting CDPATH in the configure script and place a: unset CDPATH statement below it. 6.19) Why does GnuPG 1.0.6 bail out on keyrings used with 1.0.7? There is a small bug in 1.0.6 which didn't parse trust packets correctly. You may want to apply this patch if you can't upgrade: http://www.gnupg.org/developer/gpg-woody-fix.txt 6.20) I've upgraded to GnuPG version 1.0.7 and now it takes longer to load my keyrings. What can I do? The way signature states are stored has changed so that v3 signatures can be supported. You can use the new --rebuild-keydb-caches migration command, which was built into this release and increases the speed of many operations for existing keyrings. 7. ADVANCED TOPICS 7.1) How does this whole thing work? To generate a secret/public keypair, run: gpg --gen-key and choose the default values. Data that is encrypted with a public key can only be decrypted by the matching secret key. The secret key is protected by a password, the public key is not. So to send your friend a message, you would encrypt your message with his public key, and he would only be able to decrypt it by having the secret key and putting in the password to use his secret key. GnuPG is also useful for signing things. Files that are encrypted with the secret key can be decrypted with the public key. To sign something, a hash is taken of the data, and then the hash is in some form encoded with the secret key. If someone has your public key, they can verify that it is from you and that it hasn't changed by checking the encoded form of the hash with the public key. A keyring is just a large file that stores keys. You have a public keyring where you store yours and your friend's public keys. You have a secret keyring that you keep your secret key on, and should be very careful with. Never ever give anyone else access to it and use a *good* passphrase to protect the data in it. You can 'conventionally' encrypt something by using the option 'gpg -c'. It is encrypted using a passphrase, and does not use public and secret keys. If the person you send the data to knows that passphrase, they can decrypt it. This is usually most useful for encrypting things to yourself, although you can encrypt things to your own public key in the same way. It should be used for communication with partners you know and where it is easy to exchange the passphrases (e.g. with your boy friend or your wife). The advantage is that you can change the passphrase from time to time and decrease the risk, that many old messages may be decrypted by people who accidently got your passphrase. You can add and copy keys to and from your keyring with the 'gpg --import' and 'gpg --export' option. 'gpg --export-secret-keys' will export secret keys. This is normally not useful, but you can generate the key on one machine then move it to another machine. Keys can be signed under the 'gpg --edit-key' option. When you sign a key, you are saying that you are certain that the key belongs to the person it says it comes from. You should be very sure that is really that person: You should verify the key fingerprint with: gpg --fingerprint user-id over the phone (if you really know the voice of the other person), at a key signing party (which are often held at computer conferences), or at a meeting of your local GNU/Linux User Group. Hmm, what else. You may use the option "-o filename" to force output to this filename (use "-" to force output to stdout). "-r" just lets you specify the recipient (which public key you encrypt with) on the command line instead of typing it interactively. Oh yeah, this is important. By default all data is encrypted in some weird binary format. If you want to have things appear in ASCII text that is readable, just add the '-a' option. But the preferred method is to use a MIME aware mail reader (Mutt, Pine and many more). There is a small security glitch in the OpenPGP (and therefore GnuPG) system; to avoid this you should always sign and encrypt a message instead of only encrypting it. 7.2) Why are some signatures with an ELG-E key valid? These are ElGamal keys generated by GnuPG in v3 (RFC 1991) packets. The OpenPGP draft later changed the algorithm identifier for ElGamal keys which are usable for signatures and encryption from 16 to 20. GnuPG now uses 20 when it generates new ElGamal keys but still accepts 16 (which is according to OpenPGP "encryption only") if this key is in a v3 packet. GnuPG is the only program which had used these v3 ElGamal keys - so this assumption is quite safe. 7.3) How does the whole trust thing work? It works more or less like PGP. The difference is that the trust is computed at the time it is needed. This is one of the reasons for the trustdb which holds a list of valid key signatures. If you are not running in batch mode you will be asked to assign a trust parameter (ownertrust) to a key. You can see the validity (calculated trust value) using this command. gpg --list-keys --with-colons If the first field is "pub" or "uid", the second field shows you the trust: o = Unknown (this key is new to the system) e = The key has expired q = Undefined (no value assigned) n = Don't trust this key at all m = There is marginal trust in this key f = The key is full trusted u = The key is ultimately trusted; this is only used for keys for which the secret key is also available. r = The key has been revoked d = The key has been disabled The value in the "pub" record is the best one of all "uid" records. You can get a list of the assigned trust values (how much you trust the owner to correctly sign another person's key) with: gpg --list-ownertrust The first field is the fingerprint of the primary key, the second field is the assigned value: - = No ownertrust value yet assigned or calculated. n = Never trust this keyholder to correctly verify others signatures. m = Have marginal trust in the keyholders capability to sign other keys. f = Assume that the key holder really knows how to sign keys. u = No need to trust ourself because we have the secret key. Keep these values confidential because they express your opinions about others. PGP stores this information with the keyring thus it is not a good idea to publish a PGP keyring instead of exporting the keyring. GnuPG stores the trust in the trustdb.gpg file so it is okay to give a gpg keyring away (but we have a --export command too). 7.4) What kind of output is this: "key C26EE891.298, uid 09FB: ...."? This is the internal representation of a user ID in the trustdb. "C26EE891" is the keyid, "298" is the local ID (a record number in the trustdb) and "09FB" is the last two bytes of a ripe-md-160 hash of the user ID for this key. 7.5) How do I interpret some of the informational outputs? While checking the validity of a key, GnuPG sometimes prints some information which is prefixed with information about the checked item. "key 12345678.3456" This is about the key with key ID 12345678 and the internal number 3456, which is the record number of the so called directory record in the trustdb. "uid 12345678.3456/ACDE" This is about the user ID for the same key. To identify the user ID the last two bytes of a ripe-md-160 over the user ID ring is printed. "sig 12345678.3456/ACDE/9A8B7C6D" This is about the signature with key ID 9A8B7C6D for the above key and user ID, if it is a signature which is direct on a key, the user ID part is empty (..//..). 7.6) Are the header lines of a cleartext signature part of the signed material? No. For example you can add or remove "Comment:" lines. They have a purpose like the mail header lines. However a "Hash:" line is needed for OpenPGP signatures to tell the parser which hash algorithm to use. 7.7) What is the list of preferred algorithms? The list of preferred algorithms is a list of cipher, hash and compression algorithms stored in the self-signature of a key during key generation. When you encrypt a document, GnuPG uses this list (which is then part of a public key) to determine which algorithms to use. Basically it tells other people what algorithms the recipient is able to handle and provides an order of preference. 7.8) How do I change the list of preferred algorithms? In version 1.0.7 or later, you can use the edit menu and set the new list of preference using the command "setpref"; the format of this command resembles the output of the command "pref". The preference is not changed immediately but the set preference will be used when a new user ID is created. If you want to update the preferences for existing user IDs, select those user IDs (or select none to update all) and enter the command "updpref". Note that the timestamp of the self-signature is increased by one second when running this command. 8. ACKNOWLEDGEMENTS Many thanks to Nils Ellmenreich for maintaining this FAQ file for a long time, Werner Koch for the original FAQ file, and to all posters to gnupg-users and gnupg-devel. They all provided most of the answers. Also thanks to Casper Dik for providing us with a script to generate this FAQ (he uses it for the excellent Solaris2 FAQ). Copyright (C) 2000-2002 Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111, USA Verbatim copying and distribution of this entire article is permitted in any medium, provided this notice is preserved.