aboutsummaryrefslogtreecommitdiffstats
path: root/g10 (unfollow)
Commit message (Collapse)AuthorFilesLines
2017-09-28gpg: Workaround for junk after --trusted-key.Werner Koch1-0/+20
* g10/trust.c (register_trusted_key): Cut off everthing starting as a hash sign. -- This problem is fallout from commit f99830b72812395da5451152bdd2f2d90a7cb7fb which fixes GnuPG-bug-id: 1206 The same could happen with other options taking keyids but we won't change that because a trailing '#' does not indicate a comment. So this is really only a workaround and eventually we will deprecate --trusted-key anyway or require a fingerprint as a value. Signed-off-by: Werner Koch <[email protected]>
2017-09-27gpg: Prepare for a longer fingerprintWerner Koch8-22/+65
* g10/card-util.c (change_cafpr): Use MAX_FINGERPRINT_LEN. * g10/cipher.c (write_header): Use snprintf. * g10/gpg.h (MAX_FINGERPRINT_LEN): Change to 32. (MAX_FORMATTED_FINGERPRINT_LEN): Change to 59 * g10/keyid.c (format_hexfingerprint): Add v5 fingerprint format. * g10/tofu.c (get_policy): Use MAX_FINGERPRINT_LEN for the buffer but keep the raw length for now. -- Note that this patch only increases the size of the buffer and adds a new formatting for v5 fingerprints. Moe work is required to fix internal data structures like those in trustdb.gpg and the tofu tables. Signed-off-by: Werner Koch <[email protected]>
2017-09-27common: Add constant KEYGRIP_LEN.Werner Koch2-4/+4
* common/util.h (KEYGRIP_LEN): New. * g10/call-agent.c (agent_probe_any_secret_key): Use that constant. * g10/keyid.c (keygrip_from_pk): Ditto. Signed-off-by: Werner Koch <[email protected]>
2017-09-27gpg: Let --debug clock time sign and verify.Werner Koch2-0/+10
* configure.ac (ENABLE_LOG_CLOCK): New ac_define and option. * common/logging.c (log_clock): Use ENABLE_LOG_CLOCK to enable timestamp printing. * g10/call-agent.c (agent_pksign): Time signing. * g10/sig-check.c (check_signature_end_simple): Time verification. -- Timing for verification is limited to data signatures because this is the most common thing to evaluate. We should consider to change log_clock to printf style so that we could print the signature class and other info. Signed-off-by: Werner Koch <[email protected]>
2017-09-26g10: Select a secret key by checking availability under gpg-agent.NIIBE Yutaka1-4/+11
* g10/getkey.c (finish_lookup): Add WANT_SECRET argument to confirm by agent_probe_secret_key. (get_pubkey_fromfile, lookup): Supply WANT_SECRET argument. -- GnuPG-bug-id: 1967 Signed-off-by: NIIBE Yutaka <[email protected]>
2017-09-13gpgv: Initialize compliance checker.Werner Koch1-0/+1
* g10/gpgv.c (main): Call gnupg_initialize_compliance. -- The compliance checker needs to be initialize so that it won't let spit out a "not suitable" message. We use the module name of gpg. Because there is no option to change the compliance mode in gpgv we will always be in the default (CO_GNUPG) mode. It also does not make much sense to have it here because gpgv expects a "curated" keyring. GnuPG-bug-id: 3404 Signed-off-by: Werner Koch <[email protected]>
2017-09-12gpg: Fix "Fix key generation with only an email part".Werner Koch1-1/+1
* g10/keygen.c (proc_parameter_file): Don't check the result of stpcpy. -- Fixes-commit: 7089dcc54099a4909ce7d386c07ab87e1398e2eb Signed-off-by: Werner Koch <[email protected]>
2017-09-11gpg: Fix key generation with only an email part.Werner Koch1-1/+8
* g10/keygen.c (proc_parameter_file): Special case the email only case. -- Using a parameter file like %ask-passphrase key-type: RSA key-length: 2048 key-usage: sign subkey-type: RSA subkey-length: 2048 subkey-usage: encrypt name-email: [email protected] with "gpg --gen-key --patch" the result was this key pub rsa2048 2017-09-11 [SC] 63A8C1BA12CC289A0E8072C971C7F8D4A18CE0BE uid [ultimate] <[email protected]> sub rsa2048 2017-09-11 [E] At least the the extra leading space the left angle bracket is wrong. Further some mail providers reject keys which consist of more than just a plain mail address. Using just a mail address is anyway the new new suggested content for a user id. With this patch the key will be pub rsa2048 2017-09-11 [SC] B302343C20EA6DECDB6A155135352F2520397080 uid [ultimate] [email protected] sub rsa2048 2017-09-11 [E] Signed-off-by: Werner Koch <[email protected]>
2017-09-08gpg: default to AES-256.Daniel Kahn Gillmor1-1/+3
* g10/main.h (DEFAULT_CIPHER_ALGO): Prefer AES256 by default. -- It's 2017, and pretty much everyone has AES-256 available. Symmetric crypto is also rarely the bottleneck (asymmetric crypto is much more expensive). AES-256 provides some level of protection against large-scale decryption efforts, and longer key lengths provide a hedge against unforseen cryptanalysis. Signed-off-by: Daniel Kahn Gillmor <[email protected]>
2017-09-08gpg: default to 3072-bit RSA keys.Daniel Kahn Gillmor2-7/+6
* agent/command.c (hlp_genkey): update help text to suggest the use of 3072 bits. * doc/wks.texi: Make example match default generation. * g10/keygen.c (DEFAULT_STD_KEY_PARAM): update to rsa3072/cert,sign+rsa3072/encr, and fix neighboring comment, (gen_rsa, get_keysize_range): update default from 2048 to 3072). * g10/keyid.c (pubkey_string): update comment so that first example is the default 3072-bit RSA. -- 3072-bit RSA is widely considered to be 128-bit-equivalent security. This is a sensible default in 2017. Signed-off-by: Daniel Kahn Gillmor <[email protected]> Gbp-Pq: Topic update-defaults Gbp-Pq: Name 0015-gpg-default-to-3072-bit-RSA-keys.patch
2017-08-24gpg: Fix memory leak while running --check-trustdb.Werner Koch1-0/+3
* g10/trustdb.c (update_min_ownertrust): Free PK. -- This bug was revealed by the new trust-pgp-2.scm test. Signed-off-by: Werner Koch <[email protected]>
2017-08-24gpg: Fix memory leak in sig-check.Werner Koch1-2/+5
* g10/sig-check.c (check_signature_over_key_or_uid): Remove useless condition. Actually free when SIGNER was allocated by us. -- SIGNER_ALLOCATED never received a value of -1 but that was tested. IF SIGNER_ALLOCATED was 2 the memory was never freed: if (signer_allocated == 1) if (signer_allocated == 2) free() Fixes-commit: 44cdb9d73f1a0b7d2c8483a119b9c4d6caabc1ec This function needs to be audited more thoroughly. Signed-off-by: Werner Koch <[email protected]>
2017-08-24indent: Change comment style on two functionsWerner Koch1-63/+71
-- This is to make those function better readable. if (foo) /* Comment */ { } is bad style because it requires extra time to notice the begin of the block and vice versa when noticing the block it is not clear whether this is an conditioned or unconditioned block. Having asterisks on the left is better for view impaired people and for b/w printouts. Signed-off-by: Werner Koch <[email protected]>
2017-08-23gpgconf: Swap "auto-key-retrieve" and "no-auto-key-retrieve".Werner Koch1-6/+7
* g10/gpg.c (gpgconf_list): Announce "auto-key-retrieve". (main): Simplify setting of KEYSERVER_AUTO_KEY_RETRIEVE. * tools/gpgconf-comp.c: Make "no-auto-key-retrieve" invisible. Make "auto-key-retrieve" an expert option. -- This basically reverts 9bb13a0e819334681caca38c9074bd7bfc04e45e because --no-auto-key-retrieve is again the default. Note that we allow both options for the sake of profiles. Signed-off-by: Werner Koch <[email protected]>
2017-08-21gpgconf: Make WoT settings configurable by gpgconf.Damien Goutte-Gattat1-0/+3
* tools/gpgconf-comp.c (gc_options_gpg): Add max-cert-depth, completes-needed, and marginals-needed options. * g10/gpg.c (gpgconf_list): Likewise. -- Some tests to come for the PGP trust model will need to manipulate these parameters. Signed-off-by: Damien Goutte-Gattat <[email protected]>
2017-08-11gpg: default to --no-auto-key-retrieve.Daniel Kahn Gillmor1-2/+1
* g10/gpg.c (main): remove KEYSERVER_AUTO_KEY_RETRIEVE from the default keyserver options. * doc/gpg.texi: document this change. -- This is a partial reversion of 7e1fe791d188b078398bf83c9af992cb1bd2a4b3. Werner and i discussed it earlier today, and came to the conclusion that: * the risk of metadata leakage represented by a default --auto-key-retrieve, both in e-mail (as a "web bug") and in other contexts where GnuPG is used to verified signatures, is quite high. * the advantages of --auto-key-retrieve (in terms of signature verification) can sometimes be achieved in other ways, such as when a signed message includes a copy of its own key. * when those other ways are not useful, a graphical, user-facing application can still offer the user the opportunity to choose to fetch the key; or it can apply its own policy about when to set --auto-key-retrieve, without needing to affect the defaults. Note that --auto-key-retrieve is specifically about signature verification. Decisions about how and whether to look up a key during message encryption are governed by --auto-key-locate. This change does not touch the --auto-key-locate default of "local,wkd". The user deliberately asking gpg to encrypt to an e-mail address is a different scenario than having an incoming e-mail trigger a potentially unique network request. Signed-off-by: Daniel Kahn Gillmor <[email protected]>
2017-08-09g10: Write status error on error of --quick-revoke-uid.Marcus Brinkmann1-1/+4
* g10/keyedit.c (keyedit_quick_revuid): Write status error on error. Signed-off-by: Marcus Brinkmann <[email protected]> GnuPG-bug-id: 2963
2017-08-08gpg: Add option '--disable-dirmngr'.Justus Winter3-0/+9
* doc/gpg.texi: Document new option. * g10/call-dirmngr.c (create_context): Fail if option is given. * g10/gpg.c (cmd_and_opt_values): New value. (opts): New option. (gpgconf_list): Add new option. (main): Handle new option. * g10/options.h (struct opt): New field 'disable_dirmngr'. * tools/gpgconf-comp.c (gc_options_gpg): New option. GnuPG-bug-id: 3334 Signed-off-by: Justus Winter <[email protected]>
2017-08-04gpg: Fix memory leak in parse_auto_key_locate.Werner Koch1-3/+4
* g10/getkey.c (parse_auto_key_locate): Fix freeing of OPTIONS. -- It was probably too late for me to hack. Signed-off-by: Werner Koch <[email protected]>
2017-08-04gpg: Make --no-auto-key-retrieve gpgconf-igurable.Werner Koch1-1/+1
* g10/gpg.c (gpgconf_list): Print no-auto-key-retrieve instead of auto-key-retrieve. * tools/gpgconf-comp.c (gc_options_gpg): Replace auto-key-retrieve by no-auto-key-retrieve and chnage level from invisible to advanced. Signed-off-by: Werner Koch <[email protected]>
2017-08-04gpg: Default to --auto-key-locate "local,wkd" and --auto-key-retrieve.Werner Koch3-4/+19
* g10/gpg.c (main): Add KEYSERVER_AUTO_KEY_RETRIEVE to the default keyserver options. Set the default for --auto-key-locate to "local,wkd". Reset that default iff --auto-key-locate has been given in the option file or in the commandline. * g10/getkey.c (parse_auto_key_locate): Work on a copy of the arg. -- GnuPG-bug-id: 3324 Signed-off-by: Werner Koch <[email protected]>
2017-08-04gpg: Avoid double fingerprint printing with import-show.Werner Koch1-1/+2
* g10/import.c (import_one) <IMPORT_SHOW>: Take care of fingerprint options. Signed-off-by: Werner Koch <[email protected]>
2017-08-04gpg: New import option show-only.Werner Koch2-1/+6
* g10/options.h (IMPORT_DRY_RUN): New. * g10/import.c (parse_import_options): Add "show-only". (import_one): use that as alternative to opt.dry_run. -- This is just a convenience thing for --import-options import-show --dry-run Signed-off-by: Werner Koch <[email protected]>
2017-08-02g10: Always save standard revocation certificate in file.Marcus Brinkmann1-0/+4
* g10/revoke.c (gen_standard_revocation): Set opt.outfile to NULL temporarily to create certificate in right place. Signed-off-by: Marcus Brinkmann <[email protected]> GnuPG-bug-id: 3015
2017-08-01Revert "g10: Always save standard revocation certificate in file."Marcus Brinkmann7-18/+16
This reverts commit ebc65ff459e6c228fb7406e375819a9fe5637abe.
2017-08-01g10: Always save standard revocation certificate in file.Marcus Brinkmann7-16/+18
* g10/main.h (open_outfile): New parameter NO_OUTFILE. * g10/openfile.c (open_outfile): New parameter NO_OUTFILE. If given, never use opt.outfile. * g10/revoke.c (create_revocation): If FILENAME is true, also set NO_OUTFILE to true (for standard revocation certificates). * g10/dearmor.c, g10/encrypt.c, g10/export.c, g10/revoke.c, g10/sign.c: Adjust all other callers. Signed-off-by: Marcus Brinkmann <[email protected]> GnuPG-bug-id: 3015
2017-08-01gpg,sm: Error out on compliance mismatch while decrypting.Werner Koch1-18/+24
* g10/pubkey-enc.c (get_session_key): Bail out if the algo is not allowed in the current compliance mode. * sm/decrypt.c (gpgsm_decrypt): Ditto. -- The idea here is that the owner of the key created a non-compliant key and later receives a mail encrypted to that key. The sender should have checked this key too but we can't guarantee that. By hard failing here the owner of the key will notice that he had created a non-compliant key and thus has a chance to generate a new compliant key. In case the compliant criteria changes and the owner wants to decrypt an old message he can still switch gpg to another compliant mode. Fixes-commit: a0d0cbee7654ad7582400efaa92d493cd8e669e9 GnuPG-bug-id: 3308 Signed-off-by: Werner Koch <[email protected]>
2017-07-31Explain the "server is older than xxx warning".Werner Koch2-0/+15
* g10/call-agent.c (warn_version_mismatch): Print a note on how to restart the servers. * g10/call-dirmngr.c (warn_version_mismatch): Ditto. * sm/call-agent.c (warn_version_mismatch): Ditto. * sm/call-dirmngr.c (warn_version_mismatch): Ditto. -- We should move this fucntion to common. However, the status output functions are different and would need to be streamlined too. GnuPG-bug-id: 3117 Debian-bug-id: 860745 Signed-off-by: Werner Koch <[email protected]>
2017-07-28gpg,sm: String changes for compliance diagnostics.Werner Koch7-31/+27
Signed-off-by: Werner Koch <[email protected]>
2017-07-28gpg: Minor rework for better readibility of get_best_pubkey_byname.Werner Koch2-17/+17
* g10/getkey.c (get_best_pubkey_byname): Change return type to gpg_error_t. Use var name err instead of rc. Move a gpg_error_from_syserror closer to the call. -- Signed-off-by: Werner Koch <[email protected]>
2017-07-28gpg: Fix segv in get_best_pubkey_byname.Werner Koch1-1/+1
* g10/getkey.c (get_best_pubkey_byname): Init NEW. -- We call free_user_id on NEW.uid and thus it needs to be initialized. This fixes the ref-count or invisible segv bug from GnuPG-bug-id: 3266 Signed-off-by: Werner Koch <[email protected]>
2017-07-27gpg: Tweak compliance checking for verificationWerner Koch2-4/+15
* common/compliance.c (gnupg_pk_is_allowed): Rework to always allow verification. * g10/mainproc.c (check_sig_and_print): Print a con-compliant warning. * g10/sig-check.c (check_signature2): Use log_error instead of log_info. -- We should be able to verify all signatures. So we only print a warning. That is the same beheavour as for untrusted keys etc. GnuPG-bug-id: 3311 Signed-off-by: Werner Koch <[email protected]>
2017-07-27gpg,sm: Allow encryption (with warning) to any key in de-vs mode.Werner Koch1-10/+6
* g10/encrypt.c (encrypt_crypt): Do not abort for a non-compliant key. * sm/encrypt.c (gpgsm_encrypt): Ditto. -- GnuPG-bug-id: 3306 Signed-off-by: Werner Koch <[email protected]>
2017-07-27gpg,sm: Fix compliance checking for decryption.Werner Koch1-24/+18
* common/compliance.c (gnupg_pk_is_compliant): Remove the Elgamal signing check. We don't support Elgamal signing at all. (gnupg_pk_is_allowed) <de-vs>: Revert encryption/decryption for RSA. Check the curvenames for ECDH. * g10/pubkey-enc.c (get_session_key): Print only a warning if the key is not compliant. * sm/decrypt.c (gpgsm_decrypt): Ditto. Use the same string as in gpg so that we have only one translation. -- We always allow decryption and print only a note if the key was not complaint at the encryption site. GnuPG-bug-id: 3308 Signed-off-by: Werner Koch <[email protected]>
2017-07-27indent: Wrap an overlong line.Werner Koch1-1/+3
-- Folks, please set your editors to 80 columns to notice such flaws.
2017-07-27gpg: Avoid output to the tty during import.Werner Koch7-74/+96
* g10/key-check.c (key_check_all_keysigs): Add arg mode and change all output calls to use it. * g10/keyedit.c (keyedit_print_one_sig): Add arg fp and chnage all output calls to use it. (keyedit_menu): Adjust for changes. * g10/gpgcompose.c (keyedit_print_one_sig): Add dummy arg fp. * g10/import.c (import_one): Call key_check_all_keysigs with output to the log stream. -- Fixes-commit: 404fa8211b6188a0abe83ef43a4b44d528c0b035 GnuPG-bug-id: 3288 Signed-off-by: Werner Koch <[email protected]>
2017-07-26g10: Make sure exactly one fingerprint is output with --quick-gen-key.Marcus Brinkmann1-1/+2
* g10/keygen.c (do_generate_keypair): Only set fpr in list_keyblock_direct invocation if neither --fingerprint nor --with-fingerprints are given. Signed-off-by: Marcus Brinkmann <[email protected]> GnuPG-bug-id: 2741
2017-07-25gpg: Update key origin info during import merge.Werner Koch1-4/+97
* g10/import.c (update_key_origin): New. (merge_blocks): Add arg curtime. (import_one): Pass curtime to merge_blocks. Call update_key_origin. -- We probably need to refine the rules on how this is done. But it is a start. Signed-off-by: Werner Koch <[email protected]>
2017-07-25gpg: Store key origin for new userids during import merge.Werner Koch1-107/+157
* g10/import.c (apply_meta_data): Rename to ... (insert_key_origin): this. Factor code out to ... (insert_key_origin_pk, insert_key_origin_uid): new funcs. (import_one): Move insert_key_origin behind clean_key. (merge_blocks): Add args options, origin, and url. (append_uid): Rename to ... (append_new_uid): this. Add args options, curtime, origin, and url. Call insert_key_origin_uid for new UIDs. -- This is a straightforward change to handle new user ids. How to test: With an empty keyring run gpg --with-key-origin --locate-key \ --auto-key-locate clear,nodefault,wkd [email protected] and then append a new keyid using gpg --with-key-origin --locate-key \ --auto-key-locate clear,nodefault,wkd [email protected] Works with my current key 80615870F5BAD690333686D0F2AD85AC1E42B367. Signed-off-by: Werner Koch <[email protected]>
2017-07-24gpg: Extend --key-origin to take an optional URL arg.Werner Koch5-2/+44
* g10/getkey.c (parse_key_origin): Parse appended URL. * g10/options.h (struct opt): Add field 'key_origin_url'. * g10/gpg.c (main) <aImport>: Pass that option to import_keys. * g10/import.c (apply_meta_data): Extend for file and url. * g10/keyserver.c (keyserver_fetch): Pass the url to import_keys_es_stream. -- Example: gpg --key-origin url,myscheme://bla --import FILE Signed-off-by: Werner Koch <[email protected]>
2017-07-24gpg: Store key origin info for new keys from a keyserverWerner Koch2-29/+66
* g10/keyserver.c (keyserver_get_chunk): Use KEYORG_KS if request was done by fingerprint. * g10/import.c (apply_meta_data): Implement that. Signed-off-by: Werner Koch <[email protected]>
2017-07-24gpg: Store key origin info for new DANE and WKD retrieved keys.Werner Koch7-38/+101
* g10/import.c (apply_meta_data): Remove arg 'merge'. Add arg 'url'. Implement WKD and DANE key origin. (import_keys_internal): Add arg 'url' and change all callers. (import_keys_es_stream): Ditto. (import): Ditto. (import_one): Ditto. * g10/keylist.c (list_keyblock_print): Fix update URL printing. * g10/call-dirmngr.c (gpg_dirmngr_wkd_get): Add arg 'r_url' to return the SOURCE. Pass ks_status_cb to assuan_transact. * g10/keyserver.c (keyserver_import_wkd): Get that URL and pass it to the import function. -- Note that this only for new keys. Merging this info will be added soon. Signed-off-by: Werner Koch <[email protected]>
2017-07-24gpg: Filter keys received via DANEWerner Koch1-5/+27
* g10/keyserver.c (keyserver_import_cert): Use an import filter in DANE mode. -- We only want to see the user ids requested via DANE and not any additional ids. This filter enables this in the same way we do this in WKD. Signed-off-by: Werner Koch <[email protected]>
2017-07-24g10: Make sure to emit NEED_PASSPHRASE on --import of secret key.Marcus Brinkmann3-3/+9
* call-agent.h (agent_import_key): Add keyid parameters. * call-agent.c (agent_import_key): Set keyid parameters. * import.c (transfer_secret_keys): Pass keyid parameters. Signed-off-by: Marcus Brinkmann <[email protected]> GnuPG-bug-id: 2667
2017-07-24g10: Make sure to emit NEED_PASSPHRASE on --export-secret-key.Marcus Brinkmann3-3/+9
* call-agent.h (agent_export_key): Add keyid parameters. * call-agent.c (agent_export_key): Set keyid parameters. * export.c (receive_seckey_from_agent): Pass keyid parameters. Signed-off-by: Marcus Brinkmann <[email protected]> GnuPG-bug-id: 2667
2017-07-21g10: Avoid caching passphrase for failed symmetric encryption.Marcus Brinkmann1-1/+2
* g10/mainproc.c (proc_encrypted): If error code is GPG_ERR_CIPHER_ALGO, assume the symmetric passphrase was wrong and invalidate the cache. Signed-off-by: Marcus Brinkmann <[email protected]> GnuPG-bug-id: 2270
2017-07-21gpg: Extend --quick-set-expire to allow subkey expiration setting.Werner Koch3-26/+108
* g10/keyedit.c (keyedit_quick_set_expire): Add new arg subkeyfprs. (menu_expire): Rename arg force_mainkey to unattended and allow unattended changing of subkey expiration. * g10/gpg.c (main): Extend --quick-set-expire. Signed-off-by: Werner Koch <[email protected]>
2017-07-21gpg: Fix possible double free of the card serialno.Werner Koch1-0/+6
* g10/free-packet.c (copy_public_key): Copy fields serialno and updateurl. -- The PK->serialno is used to get the version of the card to decide whether it does support other algorithms than SHA-1. This value is cached but no deep copy was done when calling copy_public_key. Bug detected by importing some public keys and then importing a secret key which led to a double free. Signed-off-by: Werner Koch <[email protected]>
2017-07-21gpg: Use macros to check the signature class.Werner Koch1-24/+24
* g10/import.c: Use the extistin macros for better readability. Signed-off-by: Werner Koch <[email protected]>
2017-07-21g10: Clean keyblock on initial commit.Marcus Brinkmann1-0/+5
* g10/import.c (import_one): If option import-clean is set, also clean on initial import, not only for merge. Signed-off-by: Marcus Brinkmann <[email protected]> GnuPG-bug-id: 2401
generated by cgit v1.2.3 (git 2.25.1) at 2025-06-23 22:56:22 +0000