* common/kem.c: Move constants to the top. Add some documentation.
* g10/pkglue.c (pk_encrypt): Add arguments session_key and factor code
out to ...
(do_encrypt_rsa_elg): here,
(do_encrypt_ecdh): and here,
(do_encrypt_kem): and here.
* g10/encrypt.c (write_pubkey_enc): Call with session key algorithm.
--
This makes it easier to review the code.
* common/openpgp-oid.c (openpgp_is_curve_supported): Allow the
abbreviated curve name.
* g10/pkglue.c (pk_encrypt): Add debug output.
* g10/seskey.c (encode_session_key): Handle Kyber session key like
ECDH. This is just a stub.
* g10/keygen.c (ecckey_from_sexp): Use the modern OID for cv25519.
(parse_key_parameter_part): Allow more Kyber variants.
--
Test by creating an ed25519 key and using
gpg --quick-add-key --batch --passphrase "" <fingerprint> <algo>
to create several subkeys. Tested with ALGOs:
kyber768
kyber1024
ky768_cv25519
ky768_bp256
kyber768_nistp256
ky1024_cv448
All curves capable of encryption should work.
GnuPG-bug-id: 6815
* agent/cvt-openpgp.c (extract_private_key): Support Kyber algorithms.
* common/openpgp-oid.c (map_gcry_pk_to_openpgp): Map KEM to Kyber.
* common/sexputil.c (get_pk_algo_from_key): Increase buffer for use
with "kyber1024".
* g10/call-agent.c (agent_get_keyinfo): Fix warning.
* g10/keygen.c (do_create_from_keygrip): Support Kyber.
(ask_algo): Ditto.
--
To test create a standard key and the use --edit-key and "addkey" with
selection 13 and use the comma delimited keygrips.
GnuPG-bug-id: 7014
* common/openpgpdefs.h (PUBKEY_ALGO_KY768_25519): Remove.
(PUBKEY_ALGO_KY1024_448): Remove.
(PUBKEY_ALGO_KYBER): New. Use them everywhere instead of the removed.
* g10/build-packet.c (gpg_mpi_write_nohdr): Rename to
(gpg_mpi_write_opaque_nohdr): this. Change callers.
(gpg_mpi_write_opaque_32): New.
(do_key): Support Kyber keys using the revised format.
* g10/gpg.h (MAX_EXTERN_KEYPARM_BITS): New.
* g10/parse-packet.c (read_octet_string): Add arg nbytes so support
reading with a length prefix. Adjust callers.
(parse_key): Parse Kyber public keys.
* g10/misc.c (pubkey_get_npkey): Support Kyber.
(pubkey_get_nskey): Ditto.
* g10/keyid.c (pubkey_string): Support dual algorithms.
(do_hash_public_key): Support Kyber.
(nbits_from_pk): Ditto.
(keygrip_from_pk): Return the Kyber part for the ECC+Kyber dual algo.
* g10/keygen.c (struct common_gen_cb_parm_s): Add genkey_result2.
Note that this callback is not yet used.
(ecckey_from_sexp): Add optional arg sexp2 and use it for Kyber.
Change callers.
(ecckey_from_sexp): Do not leak LIST in case of an error.
(common_gen): Add arg keyparms2, change callers, and support Kyber.
(gen_kyber): New.
(get_keysize_range): Support Kyber.
(fixup_keysize): Simplify and support Kyber.
(do_create): Handle Kyber.
(parse_key_parameter_part): Remove algo strings "ky768" and "ky1024"
and add a generic "kyber" with default parameters.
--
This uses a revised format which is more aligned with the usual
OpenPGP structure. A lot of things are still missing. For example
support for handling two keygrips and checking both of them in a -K
listing. There is also only ky768_bp384 as fixed algorithm for now.
No passphrase for the Kyber part of the dual algorithm is on purpose.
A test was done using
gpg --quick-gen-key pqc1 nistp256
and then running
gpg -v --quick-add-key <fingerprint> kyber
which creates a v5 subkey on a v4 primary key. A second test using
gpg --quick-gen-key pqc2 Ed448
followed by a --quick-add-key created a v5 key with a v5 subkey.
GnuPG-bug-id: 6815
* common/openpgp-oid.c (oidtable): Add column "abbr" and set them for
Brainpool.
(openpgp_oid_to_curve): Rename arg "canon" to "mode" and implement
mode 2.
--
For dual algorithms (PQC) we need shorter versions of brainpool to
avoid names which otherwise might be capped when printed.
* common/kmac.c (compute_kmac256): Return an error for older gcrypt
versions.
--
Except for the new KEM module there is no hard requirement for
libgcrypt 1.11 *yet*.
* common/sysutils.c (no_translate_sys2libc_fd) [W32]: New global.
(disable_translate_sys2libc_fd): New.
(translate_sys2libc_fd): Make static and cobuild only for Windows.
(translate_sys2libc_fd_int): Use no_translate_sys2libc_fd flag.
* g10/gpg.c, sm/gpgsm.c (oDisableFdTranslation): New const.
(opts): Add option "disable-fd-translation".
(main): Set option.
--
GnuPG-bug-id: 7060
* common/homedir.c (gpgconf_ctl): Add field "gnupg".
(parse_gpgconf_ctl): Support keyword "gnupg".
(my_gnupg_dirname): New.
(my_fixed_default_homedir): New.
(gnupg_registry_dir): New.
(standard_homedir): Use my_gnupg_dirname and my_fixed_default_homedir.
(default_homedir): Use gnupg_registry_dir and
my_fixed_default_homedir.
(_gnupg_socketdir_internal): Use my_gnupg_dirname. Increase size of
prefixbuffer.
(gnupg_sysconfdir): Use my_gnupg_dirname.
* tools/gpgconf.c (list_dirs): Use gnupg_registry_dir.
(show_other_registry_entries): Ditto.
--
This will be useful to install versions of GnuPG VS-Desktop and GnuPG
Desktop in addition to a standard GnuPG version. Only basic tests on
Unix done; Windows testing is still outstanding.
GnuPG-bug-id: 7040
* common/homedir.c (gpgconf_ctl): new struct.
(string_is_true): New.
(parse_gpgconf_ctl): New. Based on the former code in unix_rootdir.
(check_portable_app): Use parse_gpgconf_ctl and the new struct.
(unix_rootdir): Ditto.
--
This is a unification of the gpgconf.ctl mechanism. For backward
compatibility we need to keep the empty (or actually only comments)
method as used formerly under Windows. Iff one really wants a
portable application the new portable keyword should be used, though.
Noet that the Windows portable stuff has not been tested for quite
some time.
* g13/g13.c (oNoMount): New.
(opts): Add --no-mount.
(main): Implement this.
* g13/g13-common.h (opt): Add field no_mount.
* common/status.h (STATUS_PLAINDEV): New.
* g13/sh-cmd.c (has_option): Uncomment.
(cmd_mount): Add option --no-mount and pass down.
* g13/sh-dmcrypt.c (sh_dmcrypt_mount_container): Add arg nomount and
emit PLAINDEV status line.
(sh_dmcrypt_umount_container): Rund findmnt before umount.
--
This option can be used to decrypt a device but not to mount it. For
example to run fsck first. A command or option to run fsck before a
mount will eventually be added.
The use of findmnt is needed so that we can easily remove a device
which has not been mounted.
* common/homedir.c (copy_dir_with_fixup) [W32]: Fold double
backslashes.
--
This is in general no problem but when we hash or compare the directory
to test whether tit is the standard home directory, we may use a
different socket file and thus a second instance of a daemon.
GnuPG-bug-id: 6833
* scd/app-common.h (CARDTYPE_SCE7): New.
* scd/app.c (strcardtype): Support it.
(atr_to_cardtype): New.
(app_new_register): Try to get the cardtype from atr_to_cardtype.
* scd/app-piv.c (app_select_piv): Tweak for SCE7. Add general method
to construct a S/N from the Card UUID.
--
The test cards I have are rsa2048 with X.509 certificates. I don't
have the entire chain but loading the certificates work. For testing
I created an OpenPGP key from the keys and tested signing and
decryption.
GnuPG-bug-id: 6919
* common/homedir.c (enum wantdir_values): New enums.
(unix_rootdir): Change arg to use the enums. Adjust all callers. Add
support for the socketdir keyword.
(_gnupg_socketdir_internal): Take care of the socketdir keyword in
gpgconf.ctl.
* doc/tools.texi (Files used by gpgconf): Briefly explain the
gpgconf.ctl syntax.
* common/dotlock.h (DOTLOCK_LOCK_BY_PARENT, DOTLOCK_LOCKED): New.
* common/dotlock.c [HAVE_POSIX_SYSTEM]: Include <dirent.h>.
(dotlock_get_process_id, dotlock_detect_tname): New.
(dotlock_create_unix): Handle the case when no_write option is
specified. Not creating the lock file, but detect the the file of
tname.
(dotlock_create) [HAVE_POSIX_SYSTEM]: Add support of
DOTLOCK_LOCK_BY_PARENT and DOTLOCK_LOCKED for dotlock CLI util.
(dotlock_take_unix): Support the case of DOTLOCK_LOCK_BY_PARENT.
--
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* common/dotlock.h (enum dotlock_reasons): New.
(DOTLOCK_PREPARE_CREATE): New flag.
* common/dotlock.c (struct dotlock_handle): Add info_cb and
info_cb_value.
(dotlock_create): Support the new flag.
(dotlock_finish_create): New.
(read_lockfile): Silence in case of ENOENT.
(dotlock_set_info_cb): New. Use callback after all error and info
messages.
(dotlock_take_unix, dotlock_take_w32): Allow termination by callback.
* common/name-value.c (struct name_value_container): Add flag
"modified".
(nvc_modified): New.
(nvc_new): Set flag.
(_nvc_add): Set flag.
(nvc_delete): Set flag.
(nvc_set): Set flag unless value did not change.
(nve_set): Add arg PK. Change the caller.
* agent/findkey.c (agent_write_private_key): Update only if modified.
--
This helps software which uses a file system watcher to track changes
to private keys. In particular smartcard triggered changes are a
problem for such software because this may at worst trigger another
smartcard read.
GnuPG-bug-id: 6829
* common/miscellaneous.c (is_openpgp_compressed_packet)
(is_file_compressed): Moved to ...
* common/iobuf.c: ... in this file.
(is_file_compressed): Change the argument to INP, the iobuf.
* common/util.h (is_file_compressed): Remove.
* common/iobuf.h (is_file_compressed): Add.
* g10/cipher-aead.c (write_header): Don't call write_status_printf
here.
(cipher_filter_aead): Call write_status_printf when called with
IOBUFCTRL_INIT.
* g10/cipher-cfb.c (write_header): Don't call write_status_printf
here.
(cipher_filter_cfb): Call write_status_printf when called with
IOBUFCTRL_INIT.
* g10/encrypt.c (encrypt_simple): Use new is_file_compressed function,
after call of iobuf_push_filter.
(encrypt_crypt): Likewise.
* g10/sign.c (sign_file): Likewise.
--
Cherry-pick from master commit of:
2f872fa68c
GnuPG-bug-id: 6481
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* g10/misc.c (pubkey_get_nenc): Add ky768 and ky1024 values.
* g10/parse-packet.c (read_octet_string): New.
(read_size_body): Rename to ...
(read_sized_octet_string): this and change args to update-able PKTLEN.
(parse_pubkeyenc): Split general parsing loop for easier reading.
Implement parser for the Kyber algorithms.
--
Take care: this has not been tested at all, it merely passes the
regression test for the other algos.
Kyber is also known as ML-KEM in FIPS-203.
The list mode is slighly changed: In case of a parsing error no data
is printed - before that already parsed data was printed.
GnuPG-bug-id: 6815
* common/mischelp.c (timegm): Move to ...
* common/gettime.c (timegm): here. On Windows use timegm_u32.
(timegm_u32): New.
(isotime2epoch): Factor code out to ...
(isotime_make_tm): new helper.
(isotime2epoch_u64): New.
(_win32_timegm): Remove duplicated code.
(parse_timestamp): Use of timegm.
(scan_isodatestr): Fallback to isotime2epoch_u64.
--
This mainly helps on 32 bit Windows. For Unix we assume everyone is
using 64 bit or shall wait until the libc hackers finally provide a
time64_t.
GnuPG-bug-id: 6736
* sm/minip12.c: Factor parsing code out to ...
* common/tlv-parser.c: new. Extend function names and provide a few
extra functions.
* common/Makefile.am (common_sources): Add new file.
* sm/minip12.c: Adjust to use the new parser API.
* common/dotlock.c (next_wait_interval): New.
(dotlock_take_unix): Use new function.
(dotlock_take_w32): Ditto.
--
In particular when using a dotlock file for protecting the spawning
and several processes try to spawn the agent or another component, we
often run into long delays. The solution is to is to exponential
backoff and also to reduce the initial delay from 50ms to 4ms. We
further limit the maximum wait period to about 2 seconds and then
repeat at intervals of 512, 1024 and 2048ms. In the wait-forever case
we add a small random value to have different intervals per process.
GnuPG-bug-id: 3380
For testing this code snippet in the spawning function might be
useful:
const char *s;
if ((s=getenv("hold_gpg_file")))
while (!gnupg_access (s, F_OK))
gnupg_sleep (1);
* common/b64dec.c (b64decode): New.
* common/t-b64.c: Change license to LGPL.
(oops): New macro.
(hex2buffer): New.
(test_b64decode): New.
(main): Default to run the new test.
* common/Makefile.am (module_maint_tests): Move t-b64 to ...
(module_tests): here.
--
Sometimes we have a short base64 encoded string we need todecode.
This function makes it simpler.
License change of the test module justified because I am the single
author of the code.
Werner Koch