* sm/minip12.c (parse_shrouded_key_bag): Increase size of salt buffer.
--
Reported on the mailing list. The change does not seem to have a big
regression risk, thus applied. See below for the mail
# ------------------------ >8 ------------------------
https://lists.gnupg.org/pipermail/gnupg-users/2024-September/067312.html
* sm/gpgsm.h (COMPAT_NO_CHAIN_CACHE): New.
(struct cert_cache_item_s, cert_cache_item_t): New.
(struct server_control_s): Add parent_cert_cache.
* sm/gpgsm.c (compatibility_flags): Add "no-chain-cache".
(parent_cache_stats): New.
(gpgsm_exit): Print the stats with --debug=memstat.
(gpgsm_deinit_default_ctrl): Release the cache.
* sm/certchain.c (gpgsm_walk_cert_chain): Cache the certificates.
(do_validate_chain): Ditto.
--
This gives another boost of 30% (from 6.5 to 4.0 seconds in the test
environment with ~1000 certs). do_validate_chain actually brings us
the speedup becuase the gpgsm_walk_cert_chain is not used during a key
listing. For the latter we actually cache all certificates because
that was easier.
GnuPG-bug-id: 7308
Adjusted for 2.2:
- Add gpgsm_deinit_default_ctrl
- Remove ctrl arg from keydb_new
* kbx/keybox-search.c (keybox_get_cert): Store the blob clags in the
cert object.
* sm/certchain.c (do_validate_chain): Skip clearing of the ephemeral
flag if we know that it is not set.
--
GnuPG-bug-id: 7308
* g10/trustdb.c (copy_key_item): New.
(validate_keys): Use a stripped down UTK list w/o expired keys.
--
This patch makes sure that an expired trusted key is not used for
trust computation. The test case is to delete a trusted key from the
keyring, import a copy of that key which has already expired, check
that a signed key is not anymore fully trusted and finally import a
prolonged version of the trusted key and check that the signed key is
now again fully trusted.
GnuPG-bug-id: 7200
* g10/import.c (import_one_real): Rename non_self to non_self_or_utk.
If not set after chk_self_sigs check whether the imported key is an
ultimately trusted key.
--
The revalidation mark was only set if the imported key had a new key
signature. This is in general correct but not if the imported key is
a trusted key.
GnuPG-bug-id: 7200
* g10/trustdb.c (store_validation_status): Remove arg 'stored'.
(validate_keys): Remove keyhashtable 'stored' which was never used.
--
This has been here since 2003. The variable was never evaluated -
only stored.
Also added some comments.
* g10/decrypt-data.c (struct decode_filter_context_s): Add flag
checktag_failed.
(aead_checktag): Set flag.
(decrypt_data): Initially clear that flag and check the flag after the
decryption.
* g10/mainproc.c (proc_encrypted): Revert the log_get_errorcount based
check.
--
This fixes a bug where for an OCB encrypted and signed message with
the signing key missing during decryption the DECRYPTION_FAILED status
line was printed along with "WARNING: encrypted message has been
manipulated". This was because we use log_error to show that the
signature could not be verified due to the missing pubkey; the
original fix looked at the error counter and thus triggered the
decryption failed status.
Fixes-commit: 82b39fe254703776209cebb88f428bf2d1eb596b
GnuPG-bug-id: 7042
* g10/import.c (read_block): Don't show a warning for unbnown version
also for non-primary-key packets.
* g10/parse-packet.c (parse_key): Use log_info for unsupported v5
packets.
--
This fixes the problem that 2.2 can't import keys with a v5 subkey.
This fix allows a gnupg 2.6. version to export a key with an
additional PQC subkey and 2.2 can still import the other subkeys.
The second patch avoids that gpg returns with an error code.
Updates-commit: de70a2f377c1647417fb8a2b6476c3744a901296
* g10/options.h (flags): Add proc_all_sigs.
* g10/mainproc.c (check_sig_and_print): Do not stop signature checking
if this new option is used.
* g10/gpg.c (oProcAllSigs): New.
(opts): Add "proc-all-sigs".
(main): Set it.
--
GnuPG-bug-id: 7261
Backported-from-master: 1eb382fb1f431575872b47dc160807858b7df3e5
(cherry picked from commit cb739bb2a5796b8d48cbb92c615538bf53bcd482)
--
Due to the recently introduced use of STARTUPINFOEXW in gpgrt we now
need at least Windows Vista. Version 8 of Mingw defaults to XP SP2
which requires us to explicit override that default.
The SO number of libassuan needs an update too.
(cherry picked from commit d56b63a66181b4f029e91dcea095e9eb149895b7)
* autogen.sh: Update to version 2024-07-04 from libgpg-error.
* configure.ac (BUILD_REVISION): Rename the ac_define by
BUILD_COMMITID.
* tools/gpgconf.c (show_version_gnupg): Use it here.
--
(cherry picked from commit 736579331b14035db5b857812d18b493678f7769)
* g10/keydb.h (GET_PUBKEYBLOCK_FLAG_ADSK): New constant.
* g10/packet.h (PUBKEY_USAGE_XENC_MASK): New constant.
* g10/pubkey-enc.c (get_it): Print a note if an ADSK key was used.
Use the new get_pubkeyblock flag.
* g10/getkey.c (struct getkey_ctx_s): Add field allow_adsk.
(get_pubkeyblock): Factor all code out to ...
(get_pubkeyblock_ext): new.
(finish_lookup): Add new arg allow_adsk and make use of it.
--
This patch solves two purposes:
- We write a note that the ADSK key was used for decryption
- We avoid running into a
"oops: public key not found for preference check\n"
due to ADSK keys. The error is mostly harmless but lets gpg return
with an exit code of 2.
Backported-from-master: 882ab7fef9bf4440900c32d7463469307224f11a
* agent/findkey.c (agent_write_private_key): Init FP.
--
With
Fixes-commit: 3b69d8bf7146b8d10737d0cfea9c97affc60ad73
we took a shortcut but forgot to init FP which is usually set after
the new check. At least in the AppImage this always triggered a segv
and gpg or Kleopatra had to restart the agent. Not a problem in 2.4
becuase this code is a hack we don't need in 2.4 due to that new
ephemeral key storing feature.
GnuPG-bug-id: 6944
* common/util.h (GPG_ERR_BAD_RESET_CODE): New replacement.
(GPG_ERR_NO_RESET_CODE): New.
(GPG_ERR_BAD_PUK): New.
--
But as least 1.38 is required. Tested with 1.42
* common/tlv.h (TLV_PARSER_FLAG_T5793): New.
(tlv_parser_new): New macro. Rename function with an underscore.
(tlv_next_with_flag): New.
* common/tlv-parser.c (struct tlv_parser_s): Remove const from buffer.
Add fields crammed, lasttlv, and origoff. Remove bufferlist ands ist
definition.
(dump_to_file): New but disabled debug helper.
(parse_tag): Print more info on error.
(_tlv_parser_new): Add args lasttlv and LNO. Take a copy of the data.
(_tlv_parser_release): Free the copy of the buffer and return the
recorded TLV object from tlv_parser_new.
(_tlv_peek, tlv_parser_peek, _tlv_parser_peek_null): Remove.
(_tlv_push): Record crammed length.
(_tlv_pop): Restore crammed length.
(_tlv_parser_next): Add arg flags. More debug output. Handle cramming
here. Take care of cramming here.
(tlv_expect_object): Simplify to adjust for changes in _tlv_parser_next.
(tlv_expect_octet_string): Remove arg encapsulates. Adjust for
changes in _tlv_parser_next. Change all allers.
(tlv_expect_null): New.
(cram_octet_string): Rewrite.
(need_octet_string_cramming): Remove.
* sm/minip12.c (dump_to_file): New. Enable in debug mode and if an
envvar ist set. Replace all explict but disabled dumping to call this
function.
(parse_bag_encrypted_data): Replace tlv_peek_null and a peeking for an
optional SET by non-peeking code.
(parse_cert_bag): Ditto.
(parse_shrouded_key_bag): Replace tlv_peek_null by non-peeking code.
(parse_bag_encrypted_data): Use the new TLV_PARSER_FLAG_T5793 to
enable the Mozilla workaround.
(parse_bag_encrypted_data): Replace the 'renewed_tlv' code by the new
tlv_parser_release semantics.
(parse_shrouded_key_bag): Ditto.
(parse_shrouded_key_bag): Create a new context instead of using the
former encapsulated mechanism for tlv_expect_octet_string.
(parse_bag_data): Ditto.
(p12_parse): Ditto.
* common/tlv-parser.c: New
* common/Makefile.am: Add new file.
--
GnuPG-bug-id: 7213
Backported-from-master: 690fd61a0cf2b4b51ee64811656692eb644d2918
This backport required to introduce the new tlv-parser.c file and
remove most source copied stuff from minip12.c. Thus the above
ChangeList is not fully correct.
* g10/getkey.c (lookup): Skip keys with unknown algos.
--
If the local store has private keys with an algorithm not supported by
thi version of gpg, gpg used to bail out. Thus decryption of proper
messages was not possible. This fix skips such secret keys.
* sm/certcheck.c (gpgsm_check_cms_signature): Do not free s_sig on
error. Its owned and freed by the caller.
--
This is part of
GnuPG-bug-id: 7129
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Fixes-commit: 969abcf40cdfc65f3ee859c5e62889e1a8ccde91
(cherry picked from commit dcb0b6fd4822107d68bcb046d4d0650d02c82522)
* agent/command.c (cmd_genkey): Moved init_membuf to the top.
--
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
This is part of
GnuPG-bug-id: 7129
(cherry picked from commit 379fc5569d604c4a7b5f12b2bbfc4106893c2a9e)
* tools/wks-util.c (install_key_from_spec_file): Initialize ERR in case
the loop is never run.
--
This is part of
GnuPG-bug-id: 7129
Co-authored-by: Jakub Jelen <jjelen@redhat.com>
(cherry picked from commit 021c27510b52f86a95ae70b5f4ed5d2c3886c3e8)
* scd/app-openpgp.c (pin2hash_if_kdf): DEK had been changed to pointer
to allocated memory, so, we need to use DEKLEN for the length.
--
GnuPG-bug-id: 7121
Fixes-commit: 20e85585ed20af67ce68e637ea5c3637615ba2e9
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* g10/options.h (COMPAT_VSD_ALLOW_OCB): Remove.
* g10/gpg.c (compatibility_flags): Remove "vsd-allow_ocb".
(main): Alwas set CO_EXTRA_INFO_VSD_ALLOW_OCB.
* g10/keygen.c (keygen_set_std_prefs): Always set OCB feature flag.
* g10/encrypt.c (use_aead): Always OCB also in de-vs mode.
* sm/gpgsm.h (COMPAT_ALLOW_ECC_ENCR): Remove.
* sm/gpgsm.c (compatibility_flags): Remove "allow-ecc-encr".
* sm/encrypt.c (encrypt_dek): Always allow ecc encryption.
* sm/certreqgen.c (proc_parameters): Likewise.
--
Both feature are meanwhile approved in de-vs mode thus there is no
more need for the flags.
* g10/keygen.c (print_key_flags): Print "RENC" if set.
(ask_key_flags_with_mask): Remove RENC from the possible set of
usages. Add a direct way to set it iff the key is encryption capable.
--
This could be done by using "set your own capabilities" for an RSA
key. In fact it was always set in this case.
GnuPG-bug-id: 7072
