* g10/sign.c (sign_file): Use iobuf_read instead of iobuf_get for
reading data from detached file.
--
This patch reduces iobuf_read per byte processing overhead and speeds
up detached signing.
Detached signing speed on AMD Ryzen 5800X (4.3GiB file, SHA256):
gpg process
user time
before: 3.951s
after: 1.898s (2.0x faster)
GnuPG-bug-id: T5826
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
Backported-from-master: f8943ce098f6f193da791faf43fff823568697d9
Numbers above are given for the version in master but should be
similar for here (2.2).
* g10/keydb.h (GETPASSWORD_FLAG_SYMDECRYPT): New.
(passphrase_to_dek_ext): Remove this obsolete prototype.
* g10/passphrase.c (passphrase_get): Add arg flags. Use new flag
value.
(passphrase_to_dek): Add arg flags and pass it on.
* g10/mainproc.c (proc_symkey_enc): Use new flag.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/divert-scd.c (divert_pksign): Add arg 'grip'. Replace OPENPGP
key reference to keygrips.
(divert_pkdecrypt): Ditto.
* agent/protect.c (parse_shadow_info): Trim spaces.
* agent/pkdecrypt.c (agent_pkdecrypt): Pass the keygrip.
* agent/pksign.c (agent_pksign_do): Ditto.
* g10/mainproc.c (print_pkenc_list): Print extra info for an invalid
id error.
* g10/sign.c (do_sign): Ditto.
--
Using the keygrip instead of the identifier works on OpenPGP cards and
thus we use that to make sure that we are working on the right card.
For other cards we better don't do that to avoid regressions. Those
other cards are also usually provided and do not allow to
self-generate the keys.
Note that old versions of the code (gpg 1.4) used the fingerprint as
additional check but that was eventually removed and now that we use
the keygrip all over the place, it is best to use this to identify a
key.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/keydb.h (pref_hint): Change from union to struct and add field
'exact'. Adjust callers.
* g10/pkclist.c (algo_available): Take care of the exact hint.
* g10/sign.c (sign_file): Fix indentation. Rework the hash from
recipient prefs.
--
This fixes a encrypt+sign case like: One recipient key has SHA512 as
highest ranked hash preference but the the signing key is a 256 bit
curve. Because we don't want to use a truncated hash with ECDSA, we
need to have an exact match - this is in particular important for
smartcard which check that the hash matches the curves.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/misc.c (is_weak_digest): New.
(print_digest_algo_note): Use it here.
* g10/sig-check.c (check_signature_end_simple): Use it.
* g10/sign.c (hash_for): Do not use recipient_digest_algo if it is in
the least of weak digest algorithm.
--
If a message is signed and encrypted to several recipients, the to be
used digest algorithm is deduced from the preferences of the
recipient. This is so that all recipients are able to check the the
signature. However, if the sender has a declared an algorithm as
week, that algorithm shall not be used - in this case we fallback to
the standard way of selecting an algorithm.
Note that a smarter way of selecting the algo is to check this while
figuring out the algorithm - this needs more testing and thus we do it
the simple way.
Reported-by: Phil Pennock
Signed-off-by: Werner Koch <wk@gnupg.org>
Backported-from-master: 15746d60d492f5792e4a179ab0a08801b4049695
* common/compliance.h (PK_ALGO_FLAG_RSAPSS): New.
* common/compliance.c (gnupg_pk_is_compliant): Add arg alog_flags and
test rsaPSS. Adjust all callers.
* common/util.c (pubkey_algo_to_string): New.
(gnupg_pk_is_allowed): Ditto.
* sm/misc.c (gpgsm_ksba_cms_get_sig_val): New wrapper function.
(gpgsm_get_hash_algo_from_sigval): New.
* sm/certcheck.c (gpgsm_check_cms_signature): Change type of sigval
arg. Add arg pkalgoflags. Use the PK_ALGO_FLAG_RSAPSS.
* sm/verify.c (gpgsm_verify): Use the new wrapper and new fucntion to
also get the algo flags. Pass algo flags along. Change some of the
info output to be more like current master.
--
Signed-off-by: Werner Koch <wk@gnupg.org>
This backport from master
commit 969abcf40cdfc65f3ee859c5e62889e1a8ccde91
also includes some changes taken from
commit a759fa963a42e0652134130029217270b6d5d00b
(sm: Improve readability of the data verification output.)
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/openpgpdefs.h (SIGSUBPKT_KEY_BLOCK): New.
* g10/gpg.c (oIncludeKeyBlock): New.
(opts): New option --include-key-block.
(main): Implement.
* g10/options.h (opt): New flag include_key_block.
* g10/parse-packet.c (dump_sig_subpkt): Support SIGSUBPKT_KEY_BLOCK.
(parse_one_sig_subpkt): Ditto.
(can_handle_critical): Ditto.
* g10/sign.c (mk_sig_subpkt_key_block): New.
(write_signature_packets): Call it for data signatures.
--
This patch adds support for a to be proposed OpenPGP ferature:
Introduce the Key Block subpacket to align OpenPGP with CMS.
This new subpacket may be used similar to the CertificateSet of
CMS (RFC-5652) and thus allows to start encrypted communication
after having received a signed message. In practice a stripped down
version of the key should be including having only the key material
and the self-signatures which are really useful and shall be used by
the recipient to reply encrypted.
#### Key Block
(1 octet with value 0, N octets of key data)
This subpacket MAY be used to convey key data along with a signature
of class 0x00, 0x01, or 0x02. It MUST contain the key used to create
the signature; either as the primary key or as a subkey. The key
SHOULD contain a primary or subkey capable of encryption and the
entire key must be a valid OpenPGP key including at least one User ID
packet and the corresponding self-signatures.
Implementations MUST ignore this subpacket if the first octet does not
have a value of zero or if the key data does not represent a valid
transferable public key.
GnuPG-bug-id: 4856
Signed-off-by: Werner Koch <wk@gnupg.org>
Backported from master.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/sign.c (update_keysig_packet): Convert digest algo when needed.
--
Several gpg commands try to keep most properties of a key signature
when updating (i.e. creating a new version of a key signature). This
included the use of the current hash-algorithm. This patch changes
this so that SHA-1 or RMD160 are replaced by SHA-256 if
possible (i.e. for RSA signatures). Affected commands are for example
--quick-set-expire and --quick-set-primary-uid.
GnuPG-bug-id: 4508
Signed-off-by: Werner Koch <wk@gnupg.org>
* 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 <mb@g10code.com>
GnuPG-bug-id: 3015
* common/compliance.c (gnupg_rng_is_compliant): New.
* g10/call-agent.c (start_agent) [W32]: Check rng compliance.
* sm/call-agent.c (start_agent) [W32]: Ditto.
* g10/encrypt.c (encrypt_simple, encrypt_crypt): Check that the RNG is
compliant.
* sm/encrypt.c (gpgsm_encrypt): Ditto.
* g10/sign.c (do_sign): Ditto.
* sm/sign.c (gpgsm_sign): Ditto.
--
Under Windows we need to check that the Jitter RNG is active in de-vs
mode. Under Linux this is not necessary because /dev/random can be
scrutinized and is believed to provide enough entropy.
Signed-off-by: Werner Koch <wk@gnupg.org>
--
For proper operations as a server we need to avoid global variables.
Thus we need to pass the session state CTRL to most functions. Quite
a lot of changes but fortunately straightforward to do.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/packet.h (struct parse_packet_ctx_s): Add fields LAST_PKT and
FREE_LAST_PKT.
(init_parse_packet): Clear them.
(deinit_parse_packet): New macro. Change all users if
init_parse_packet to also call this macro.
* g10/free-packet.c (free_packet): Add arg PARSECTX and handle shallow
packet copies in the context. Change all callers.
* g10/parse-packet.c (parse): Store certain packets in the parse
context.
--
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/sign.c (write_signature_packets): Free SIG. Also replace
xcalloc by xtrycalloc.
--
If do_sign fails SIG was not released. Note that in the good case SIG
is transferred to PKT and freed by free_packet.
Reported-by: Stephan Müller
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/options.h (struct opt): Add field 'sender_list'.
* g10/gpg.c: Include mbox-util.h.
(oSender): New.
(opts): Add option "--sender".
(main): Parse option.
--
This option will eventually be used for more advanced purposes.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/passphrase.c (passphrase_get): Remove arg KEYID. Change arg
MODE to NOCACHE.
(passphrase_to_dek): Remove args KEYID and PUBKEY_ALGO. Split arg
MODE into CREATE and NOCACHE. Change all callers and adjust stubs.
(passphrase_clear_cache): Remove args KEYID and ALGO. They are not
used. Change caller.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/gpg.c (oMimemode): New.
(opts): Add --mimemode.
(main): Use --mimemode only in rfc4880bis compliance mode.
* g10/options.h (struct opt): Add field "mimemode".
* g10/build-packet.c (do_plaintext): Allow for mode 'm'.
* g10/encrypt.c (encrypt_simple, encrypt_crypt): Use 'm' if requested.
* g10/plaintext.c (handle_plaintext): Handle 'm' mode.
* g10/sign.c (write_plaintext_packet): Handle 'm' mode.
(sign_file, sign_symencrypt_file): Use 'm' if requested.
--
Thsi patch prepares for a proposed change in RFC4880bis to support a
MIME flag. A literal data packet with the mime flag set is handled
like a 't' or 'u' but CR are not removed. The PLAINTEXT status line
will also indicate a MIME content.
If --mimemode is used without --rfc4880bis 't' will be used.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/openpgpdefs.h (SIGSUBPKT_ISSUER_FPR): New.
* g10/build-packet.c (build_sig_subpkt_from_sig): Add arg PKSK and
insert the issuer fpr if needed.
* g10/sign.c (write_signature_packets): Pass signing key.
(make_keysig_packet): Ditto.
(update_keysig_packet): Ditto.
* g10/parse-packet.c (dump_sig_subpkt): Print issuer fpr.
(parse_one_sig_subpkt): Detect issuer fpr.
(can_handle_critical): Add issuer fpr.
* g10/mainproc.c (check_sig_and_print): Try to get key via fingerprint.
* g10/gpgv.c (keyserver_import_fprint): New stub.
* g10/test-stubs.c (keyserver_import_fprint): New stub.
--
This support is enabled with the --rfc4880bis option and intended to
test to recently proposed issuer fpr.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/gpg.c (oDisableSignerUID): New.
(opts): New option '--disable-signer-uid'.
(main): Set option.
* g10/options.h (opt): Add field flags.disable_signer_uid.
* g10/sign.c: Include mbox-util.h.
(mk_notation_policy_etc): Embed the signer's uid.
* g10/mainproc.c (check_sig_and_print): Do not use WKD for auto key
retrieval if --disable-signer-uid is used.
--
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/call-agent.c, g10/call-agent.h (agent_get_keyinfo): add
r_cleartext parameter to report whether a key is stored without
passphrase protection.
* g10/gpgv.c, g10/test-stubs.c: augment dummy agent_get_keyinfo to
match new API.
* g10/export.c, g10/keyedit.c, g10/keygen.c, g10/keylist.c,
g10/sign.c: pass NULL to agent_get_keyinfo since we do not yet
need to know whether agent is passphrase-protected.
--
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
* g10/pkclist.c (build_pk_list): Remove parameter use, which is always
called set to PUBKEY_USAGE_ENC. Update callers.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>
* g10/encrypt.c (use_mdc): Make it a global func.
* g10/sign.c (sign_symencrypt_file): Use that function to decide
whether to use an MDC.
* tests/openpgp/conventional-mdc.test: Add a simple test case.
--
We used --force-mdc in sign+symenc mode (-cs) only with --force-mdc.
That broke our assumption from commit 625e292 (GnuPG 2.1.9) that all
uses of modern ciphers are using MDC.
Reported-by: Ben Kibbey <bjk@luxsci.net>
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/sign.c (do_sign): Let verify signature by gpg-agent.
* agent/pksign.c (agent_pksign_do): Call gcry_pk_verify for RSA.
--
RSA signature verification should be done to prevent attacks against
RSA CRT implementations and not to return invalid signature to
adversary. Newer libgcrypt does so. For older libgcrypt and
smartcards, gpg-agent does signature verification.
* g10/gpg.c (opts): Remove --no-sig-create-check.
* g10/options.h (struct opt): Remove field no_sig_create_check.
* g10/sign.c (do_sign): Always check unless it is RSA and we are using
Libgcrypt 1.7.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/sign.c (sign_file): Use log_printf instead of stderr.
* g10/tdbdump.c (export_ownertrust): Use estream fucntions.
(import_ownertrust): Ditto.
* g10/tdbio.c (tdbio_dump_record): Ditto. Change arg to estream_t.
--
Reported-by: Guilhem Moulin <guilhem@fripost.org>
Needed for unattended key edits with --status-fd, because since 2.1
status prompts are preceded by es_fflush (in cpr.c:do_get_from_fd)
not fflush(3), so the standard output may not be flushed before each
prompt. (Which breaks scripts using select(2) to multiplex between
the standard and status outputs.)
His patch only affected print_and_check_one_sig_colon() but there are
many more places where stdio and estream are mixed. This patch now
replaces most of them in g10/. At some places stdio is still used,
but that is local to a function and should not have side effects.
Signed-off-by: Werner Koch <wk@gnupg.org>
