* agent/command-ssh.c (ssh_send_available_keys): Do not bump
key_counter for ignored keys. Also use opt.debug instead of
opt.verbose and fix a memory leak.
--
The error shown by "ssh-add -l" before this fix was:
error fetching identities: incomplete messag
Fixes-commit: 193fcc2f7a8cca5240ce50499c54f99235a87e1c
GnuPG-bug-id: 5996
* agent/gpg-agent.c (oNoUserTrustlist,oSysTrustlistName): New.
(opts): Add new option names.
(parse_rereadable_options): Parse options.
(finalize_rereadable_options): Reset allow-mark-trusted for the new
option.
* agent/agent.h (opt): Add fields no_user_trustlist and
sys_trustlist_name.
* agent/trustlist.c (make_sys_trustlist_name): New.
(read_one_trustfile): Use here.
(read_trustfiles): Use here. Implement --no-user-trustlist.
--
With the global options we can now avoid that a user changes the
Root-CA trust by editing the trustlist.txt. However, to implement
this we need a new option so that we don't need to rely on some magic
like --no-allow-mark-trusted has been put into a force section.
The second option makes system administration easier as it allows to
keep the trustlist in a non-distributed file.
GnuPG-bug-id: 5990
* agent/command.c (get_keyinfo_on_cards): Make it static. Don't
return bogus value on error. Return NULL when scdaemon is disabled.
--
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* agent/findkey.c (agent_key_from_file): Change the semantics of
GRIP. Now, it's NULL for use by PKDECRYPT and PKSIGN/PKAUTH.
* agent/pkdecrypt.c (agent_pkdecrypt): Set GRIP=NULL.
* agent/pksign.c (agent_pksign_do): Likewise.
--
GnuPG-bug-id: 5099
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* agent/gpg-agent.c (main): Mark --supervised as deprecated.
* dirmngr/dirmngr.c (main): Ditto.
--
The supervised thing causes more trouble than it pretends to solve.
* agent/command-ssh.c (SSH_OPT_CONSTRAIN_MAXSIGN): New.
(SSH_OPT_CONSTRAIN_EXTENSION): New.
(ssh_handler_add_identity): Ignore them.
(ssh_handler_extension): Take success for session-bind.
--
OpenSSH 8.9 does not gracefully allow communication with older agent
implementations. Until this new OpenSSH feature has been settled we
return a faked response.
Code has not yet been tested.
GnuPG-bug-id: 5931
* agent/command-ssh.c (start_command_handler_ssh): Use es_sysopen.
--
With new (not-yet-released) libgpg-error, gpg-agent should be able to
handle connection from Cygwin version of OpenSSH.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* agent/command-ssh.c (get_client_info): Use type gnupg_fd_t for
socket, until call of socket API.
(start_command_handler_ssh): Don't convert here.
--
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* agent/command.c (cmd_keytocard): Timestamp at "Created:" field is
only used when time is not specified.
--
Fixes-commit: c795be79c14fac01b984bdc2e2041d2141f27612
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* agent/call-daemon.c (wait_child_thread): Print the correct name.
--
This makes sure that the log print the actual used name if for example
--pinentry-program was used.
* agent/trustlist.c (struct trustitem_s): Add flag "qual".
(read_one_trustfile): Rename arg "allow_include" to "systrust" and
change callers. Parse new flag "qual".
(istrusted_internal): Print all flags.
* sm/call-agent.c (istrusted_status_cb): Detect the "qual" flag.
* sm/gpgsm.h (struct rootca_flags_s): Add flag "qualified".
* sm/certchain.c (do_validate_chain): Take care of the qualified flag.
* agent/command-ssh.c (sexp_key_construct): Do not put an empty string
into an S-expression.
(stream_read_string): Do not not try to a read a zero length block.
--
Actually we could handles this different by not putting a comment tag
into the s-expression, however this requires more code and at other
places we already return "(none)" instead of an empty comment.
The second fix is more or less a cosmetic thing to get better error
messages in case the underlying read system call returns an error.
GnuPG-bug-id: 5794
--
The current code uses the binary ctrl->keygrip, but all the passphrase
storage engines expect this to be a string, so convert the binary
keygrip to a hex one before passing it in as the keyid. This fixes a
crash seen in some libsecret implementations where a non-ascii keyid
isn't well handled.
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
* agent/gpg-agent.c (main): Move detection up.
--
The problem is that PARGS is re-used and when detecting a possible
incorrect use, the flag that "--" has already been seen has gone.
* agent/gpg-agent.c (oStealSocket): New.
(opts): Add option.
(steal_socket): New file global var.
(main): Set option.
(create_server_socket): Implement option.
* dirmngr/dirmngr.c (oStealSocket): New.
(opts): Add option.
(steal_socket): New file global var.
(main): Set option. Add comment to eventually implement it.
--
Note that --steal-socket has currently no effect on dirmngr because
dirmngr does this anway.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/sexp-secret.c (fixup_when_ecc_private_key): Initialize buffer to
avoid its use on unexpected inputs.
--
GnuPG-bug-id: 5393
Co-authored-by: NIIBE Yutaka <gniibe@fsij.org>
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
* agent/protect.c (USE_CLOCK_GETTIME): New macro.
(calibrate_get_time): Only use clock_gettime if USE_CLOCK_GETTIME.
--
GnuPG-bug-id: 5623
Fixes-commit: 380bce13d94ff03c96e39ac1d834f382c5c730a1
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* agent/command.c (cmd_get_passphrase): Do not deref PI. PI is always
NULL.
--
Fixes-commit: b89b1f35c29ceaebe39b31444936aa66c9297f2c
GnuPG-bug-id: 5577
* agent/agent.h (PINENTRY_STATUS_PASSWORD_GENERATED): New.
(MAX_GENPIN_TRIES): Remove.
* agent/call-pinentry.c (struct entry_parm_s):
(struct inq_cb_parm_s): Add genpinhash and genpinhas_valid.
(is_generated_pin): New.
(inq_cb): Suppress constraints checking for a generated passphrase.
No more need for several tries to generate the passphrase.
(do_getpin): Store a generated passphrase/pin in the status field.
(agent_askpin): Suppress constraints checking for a generated
passphrase.
(agent_get_passphrase): Ditto.
* agent/command.c (cmd_get_passphrase): Ditto.
--
A generated passphrase has enough entropy so that all kind of extra
checks would only reduce the actual available entropy. We thus detect
if a passphrase has been generated (and not changed) and skip all
passphrase constraints checking.
* agent/call-pinentry.c (DEFAULT_GENPIN_BYTES): Replace by ...
(DEFAULT_GENPIN_BITS): this and increase to 150.
(generate_pin): Make sure that we use at least 128 bits.
* agent/call-pinentry.c: Remove unused assert.h.
(inq_cb): Fix use use of assuan_end_confidential in case of nested
use.
(do_getpin): Ditto.
(setup_formatted_passphrase): Escape the help string.
(setup_enforced_constraints): Ignore empty help strings.
--
(Ported from 2.2)
* agent/gpg-agent.c (oCheckSymPassphrasePattern): New.
(opts): Add --check-sym-passphrase-pattern.
(parse_rereadable_options): Set option.
(main): Return option info.
* tools/gpgconf-comp.c: Add new option.
* agent/agent.h (opt): Add var check_sym_passphrase_pattern.
(struct pin_entry_info_s): Add var constraints_flags.
(CHECK_CONSTRAINTS_NOT_EMPTY): New to replace a hardwired 1.
(CHECK_CONSTRAINTS_NEW_SYMKEY): New.
* agent/genkey.c (check_passphrase_pattern): Rename to ...
(do_check_passphrase_pattern): this to make code reading
easier. Handle the --check-sym-passphrase-pattern option.
(check_passphrase_constraints): Replace arg no_empty by a generic
flags arg. Also handle --check-sym-passphrase-pattern here.
* agent/command.c (cmd_get_passphrase): In --newsymkey mode pass
CHECK_CONSTRAINTS_NEW_SYMKEY flag.
* agent/call-pinentry.c (struct entry_parm_s): Add constraints_flags.
(struct inq_cb_parm_s): New.
(inq_cb): Use new struct for parameter passing. Pass flags to teh
constraints checking.
(do_getpin): Pass constraints flag down.
(agent_askpin): Take constrainst flag from the supplied pinentry
struct.
--
Requirements for a passphrase to protect a private key and for a
passphrase used for symmetric encryption are different. Thus a
the use of a different pattern file will be useful. Note that a
pattern file can be used to replace the other passphrase constraints
options and thus we don't need to duplicate them for symmetric
encryption.
GnuPG-bug-id: 5517
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/agent.h (opt): Change type of pinentry_formatted_passphrase
to int (as for other flags).
* agent/call-pinentry.c (setup_formatted_passphrase): Remove no longer
needed translated strings. Write option without value to Assuan
connection.
* agent/gpg-agent.c (opts): Use ARGPARSE_s_n for
oPinentryFormattedPassphrase.
(parse_rereadable_options): Set option to 1.
--
GnuPG-bug-id: 5553, 5517
* agent/call-pinentry.c (inq_cb): Handle checkpin inquiry.
(setup_enforced_constraints): New.
(agent_get_passphrase): Call setup_enforced_constraints if new
passphrase is requested.
--
This implements the gpg-agent side for checking whether a new passphrase
entered by the user in pinentry satisfies the passphrase constraints.
Performing a checkpin inquiry is only allowed if the passphrase
constraints are enforced. setup_enforced_constraints sends necessary
options and translated strings to pinentry.
GnuPG-bug-id: 5517, 5532
* agent/cvt-openpgp.c (do_unprotect): Only modify SKEY when it is
correctly decrypted.
--
GnuPG-bug-id: 5122
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* agent/command.c (cmd_genkey): Use goto instead of return.
* agent/cvt-openpgp.c (convert_from_openpgp_main): Ditto.
* agent/genkey.c (agent_ask_new_passphrase): Fix typo to free correct
pointer
(agent_genkey): Release memory
* agent/gpg-agent.c (check_own_socket): Free sockname
* agent/protect-tool.c (read_key): Free buf.
(agent_askpin): Free passphrase
--
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Changed original patch to not add a free before a GPG_ERR_BUG.
Signed-off-by: Werner Koch <wk@gnupg.org>
GnuPG-bug-id: 5393
* agent/command.c (cmd_keyinfo): Factor some code out to ...
(get_keyinfo_on_cards): ... new.
(cmd_havekey): Add --list mode.
* g10/gpg.h (struct server_control_s): Add new caching vars.
* g10/gpg.c (gpg_deinit_default_ctrl): Release cache.
* g10/call-agent.c (agent_probe_any_secret_key): Init and try to use
the keygrip cache.
(agent_genkey): Clear the cache.
(agent_import_key): Ditto.
* g10/keylist.c (list_all, list_one): Pass ctrl to
agent_probe_any_secret_key.
* g10/getkey.c (lookup): Ditto.
--
With this change we first ask the agent for a list of all secret
keygrips and use that list instead of asking the agent for each public
key. Speeds up my "gpg -K" with a lot of secret and public keys by
more than 25%.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/command-ssh.c (ssh_handler_request_identities): Remove double
check of ERR.
* g10/getkey.c (get_pubkey_byname): Remove double use of break.
* g10/pkglue.c (pk_encrypt): Handle possible NULL-ptr access due to
failed malloc.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/command.c (cmd_readkey): Add option --no-data and special
handling for $SIGNKEYID and $AUTHKEYID.
* g10/call-agent.c (agent_scd_getattr): Create shadow keys for KEY-FPR
output.
* g10/skclist.c (enum_secret_keys): Automagically get a missing public
key for the current card.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/command-ssh.c (ssh_handler_request_identities): Skip unknown
curves.
--
For example when using my standard ed25519 token and testing cards
with only Brainpool support, the ssh-agent failed due to the unknown
curves seen on the card. This patches fixes this by ignoring keys
with unknown curves.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/pksign.c (agent_pksign_do): Support ECC and DSA verification
and print some debug info in the error case.
--
Note that the addition of do_encode_dsa has no immediate effect
because we use the code only for RSA. However, for debugging it can
be useful to change the code to check also other created signatures.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/command.c (cmd_keytotpm): New.
(agent/command.c): Register new command KEYTOTPM.
* g10/call-agent.c (agent_keytotpm): New.
* g10/keyedit.c (cmdKEYTOTPM): New command "keytotpm".
(keyedit_menu): Implement.
--
The plumbing is done in two parts: the agent is modified to understand
a KEYTOTPM assuan command taking the key grip as an argument. This
simply obtains the key s expression and calls the existing writeky
diversion to the tpm2daemon. The daemon reponds with the TPM
conversion of the key and that key is then stored in the keyfile as a
shadowed-private-key with "tpm2-v1" type.
To effect the conversion, all the user does from gpg --edit-key is
select which private key they wish to move (or move the primary if no
key is selected) and type keytotpm. The conversion to TPM form is
instantaneous and once converted, the actual key cannot be recovered,
meaning that if you want your gpg key to move to a new laptop you must
keep an unconverted backup copy in a safe location.
When you do a list command, all TPM keys show up as
card-no: TPM-Protected
The key is stored encrypted to the TPM2 storage seed and since each
TPM has a unique seed, only the single TPM contained in your laptop
can now read the key. This means you cannot simply copy the shadowed
key file over to a new laptop, you must copy over the backup copy and
then convert it to TPM form on the new laptop.
To decomission your laptop, execute a tssclear command which
regenerates the storage seed and effectively shreds all keys. Note
when you have done this *every* TPM2 shadowed private key becomes
unreadable by any TPM and all are effectively destroyed.
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Very minor cosmetic changes.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/call-tpm2d.c: New.
* divert-tpm2.c: New.
* agent/Makefile.am: Add new files.
* agent/agent.h (DAEMON_TPM2D): New. Add stub fucntions.
* agent/call-daemon.c (GNUPG_MODULE_NAME_TPM2DAEMON): New.
* agent/command.c (do_one_keyinfo): Handle tpmv2.
* agent/gpg-agent.c (oTpm2daemonProgram): New.
(opts): New option --tpm2daemon-program.
(parse_rereadable_options): Handle option.
* agent/pkdecrypt.c (agent_pkdecrypt): Divert to tpm2d.
(agent_pksign_do): Ditto.
---
A new shadow key type: "tpm2-v1" is introduced signalling that the
shadowed key is handled by the tpm2daemon. A function to identify
this type is introduced and diversions to the tpm2daemon functions are
conditioned on this function for pkign and pkdecrypt where the same
diversions to scd are currently done. The (info) field of the
shadowed key stores the actual TPM key. The TPM key is encrypted so
only the physical TPM it was created on can read it (so no special
protection is required for the info filed), but if the (info) field
becomes corrupt or damaged, the key will be lost (unlike the token
case, where the key is actually moved inside the token).
Note, this commit adds handling for existing TPM format shadow keys,
but there is still no way to create them.
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
Additional changes:
* Add ChangeLog entries.
* Some minor indentation fixes.
* agent/Makefile.am (gpg_agent_SOURCES): Change to make distcheck
work.
* agent/agent.h [!HAVE_LIBTSS]: Do not return -EINVAL but an
gpg_error_t. Mark args as unused.
* agent/protect.c (agent_is_tpm2_key): Free BUF.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/call-daemon.c (struct wait_child_thread_parm_s) [W32]: Do not
use HANDLE for pid_t.
(wait_child_thread): Ditto.
--
Mingw has its own definition of pid_t as does libassuan. We should use
this instead of using HANDLE. Things are a bit complicated in
Windows, because Windows also has a pid_t but that is mostly useless;
in particular because you can't wait on a real pid but need a handle.
* common/session-env.c (session_env_list_stdenvnames): Extend to allow
return all names as one string.
* agent/command-ssh.c (SSH_REQUEST_EXTENSION): New.
(SSH_RESPONSE_EXTENSION_FAILURE): New.
(request_specs): Add handler for the extension command.
(ssh_handler_extension): New.
--
The extension mechanism is specified in
https://tools.ietf.org/html/draft-miller-ssh-agent-04
Signed-off-by: Werner Koch <wk@gnupg.org>
* configure.ac: Require at least Libgcrypt 1.9.0. Remove all
GCRYPT_VERSION_NUMBER dependent code.
--
Only Libgcrypt 1.9 implements EAX which is a mandatory algorithm in
RFC4880bis.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/pkdecrypt.c (agent_pkdecrypt): Support decryption with card
but without a stub key.
--
GnuPG-bug-id: 5170
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* agent/call-pinentry.c (generate_pin): Use STRING random which is
sufficient for a passphrase.
(inq_cb): s/rc/err/. Do not print two errors in case generate_pin
fails. Lowercase strings as per GNU standards.
(setup_genpin): Fix translation test.
(setup_qualitybar): Ditto.
--
Our policy is to use VERY_STRONG for a long term public key but not
for symmetric keys.
The GNU standards to demand that strings start with a lowercase
because that seems to be needed if it follows a prefix of for example
"gpg:". I don't actually like this becuase we need different strings
in interactive and standard mode but the majory of strings in GnuPGare
written this way. If a translation is necessary it can be done with
an en_GB or so locale.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/call-pinentry.c (agent_get_passphrase): Setup genpin.
(do_getpin): Update with new name for inquire callback.
(inq_quality): Rename to inq_cb and add genpin support.
(inq_cb): Renamed form inq_quality.
(generate_pin): New helper to generate a pin.
(agent_askpin): Fix some typos.
(setup_genpin): Provide new strings for pinentry.
--
This implements the gpg-agent side for generating a pin
especially for symmetric encryption using libgcrypt randomness
and checking it against the gpg-agent constraints.
* common/sysutils.h (struct gnupg_dirent_s): New.
* common/sysutils.c: Include dirent.h.
(struct gnupg_dir_s): New.
(gnupg_opendir, gnupg_readdir, gnupg_closedir): New. Change all
callers of opendir, readdir, and closedir to use these functions.
--
GnuPG-bug-id: 5098
* common/sysutils.c (gnupg_stat): New.
* common/sysutils.h: Include sys/stat.h.
--
Yet another wrapper for Unicode support on Windows.
GnuPG-bug-id: 5098
Signed-off-by: Werner Koch <wk@gnupg.org>
--
We need to use es_fopen on Windows to cope with non-ascii file names.
This is quite a large but fortunately straightforward change. At a
very few places we keep using stdio (for example due to the use of
popen).
GnuPG-bug-id: 5098
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/sysutils.c (gnupg_access): New. Replace all calls to access
by this wrapper.
* common/homedir.c (w32_shgetfolderpath): Change to return UTF-8
directory name.
(standard_homedir): Adjust for change.
(w32_commondir, gnupg_cachedir): Ditto.
--
Also use SHGetFolderPathW instead of SHGetFolderPathA on Windows.
This is required to correctly handle non-ascii filenames on Windows.
GnuPG-bug-id: 5098
* scd/app-openpgp.c (send_keypair_info): Emit the algo string as part
of a KEYPAIRINFO.
* scd/command.c (do_readkey): Ditto.
* scd/app-piv.c (do_readkey): Ditto.
* scd/app-nks.c (do_learn_status_core): Ditto.
(struct fid_cache_s): Add field algostr.
(flush_fid_cache): Release it.
(keygripstr_from_pk_file): Fill it and add it to the cache. Use a
single exit label.
* scd/app-help.c (app_help_get_keygrip_string_pk): Add new arg
r_algostr. Change all callers.
--
This is helpful so that callers do not need to parse the key for this
basic information. Use "scd readkey --info-only" to return the info
status line instead of the key material; using just --info returns the
info in addition to the key material.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/stringhelp.h (split_fields): Use const * for the strings in
the ARRAY.
(split_fields_colon): Likewise.
* common/stringhelp.c (split_fields, split_fields_colon): Fix
the implementation.
* agent/call-scd.c, agent/command.c: Follow the change.
* common/t-stringhelp.c, dirmngr/loadswdb.c: Likewise.
* g10/call-agent.c, tools/card-call-scd.c: Likewise.
* tools/card-yubikey.c, tools/gpg-card.c: Likewise.
* tools/gpg-card.h, tools/gpg-wks-client.c: Likewise.
* tools/gpgconf-comp.c, tools/gpgconf.c: Likewise.
* tools/wks-util.c: Likewise.
--
The strings in the ARRAY don't need to be released by caller, as those
are references. It's easier to follow the code when it's explicitly
const *.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
