* scd/app-openpgp.c (struct app_local_s): New field pincache.
(cache_pin): Set it.
(pin_from_cache): Consult it.
* scd/app-piv.c (struct app_local_s): New field pincache.
(cache_pin): Set it.
(pin_from_cache): Consult it.
--
Signed-off-by: Werner Koch <wk@gnupg.org>
* scd/command.c (pincache_put): Add arg pinlen and change all callers
to provide it.
* scd/app-piv.c (cache_pin): New.
(pin_from_cache): New.
(ask_and_prepare_chv): Add args no_cache and r_unpaddedpinlen. Take
PIN from the cache. Return the unpadded length.
(verify_chv): Add arg ctrl. Cache the PIN.
(do_change_chv): Clear PIN cache.
--
The PIV pins are padded but we want to store the unpadded PIN. Thus
the changes to the function.
Code has has been tested by commenting the no_cache parameter because
we the current test certificate was created for PIV.9C which requires
a verification for each use. More testing is required.
GnuPG-bug-id: 4791
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/command.c (struct server_local_s): Add last_card_keyinfo.
(eventcounter): Add maybe_key_change.
(cmd_genkey, cmd_scd, cmd_import_key, cmd_delete_key): Bump new
counter.
(cmd_keyinfo): Cache the keyinfo from the card.
(start_command_handler): Release the cache.
--
This cache speeds up processing of commands like "gpg -K" because
scdaemon does not need to be asked for each key as long as nothing
changed with the card.
We should have a better notification service from scdaemon to make
sure that we get only the relevant events. What we do right now is a
bit course but sufficient.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/sysutils.c (inhibit_set_foregound_window): New var.
(gnupg_inhibit_set_foregound_window): New func.
(gnupg_allow_set_foregound_window): Use var.
* g10/gpg.c (main): Inhibit in loopback mode.
* sm/gpgsm.c (main): Ditto.
--
Signed-off-by: Werner Koch <wk@gnupg.org>
* scd/app.c (maybe_switch_app): Factor reselect code out to ...
(run_reselect): new.
(app_write_learn_status): Tweak diagnostics.
(app_do_with_keygrip): Run reselect if a card has more than one
switchable application.
* agent/call-scd.c (agent_card_serialno): Ditto.
* tools/card-call-scd.c (start_agent): Use option --all with SERIALNO.
(scd_serialno): Ditto.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/command-ssh.c (card_key_available): Supply KEYINFO argument.
Call agent_card_readkey by KEYGRIP of KEYINFO.
Don't use $AUTHKEYID, but IDSTR of KEYINFO.
(ssh_handler_request_identities): Follow the change of
card_key_available.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* scd/app-openpgp.c (wipe_and_free, wipe_and_free_string): Use them
everywere where we do a wipememory followed by a free.
(pin2hash_if_kdf): Change interface. The input PIN is not anymore
changed. Further there are no more assumptions about the length of
the provided buffer.
(cache_pin): Restructure.
(chvno_to_keyref): New.
(pin_from_cache): New.
(verify_a_chv): Add arg CTRL. Adjust for changed pin2hash_if_kdf.
Chache and retrieve the PIN here.
(verify_chv2): Do not cache the PIN here.
(build_enter_admin_pin_prompt): Add arg 'r_remaining'.
(verify_chv3): Adjust for changed pin2hash_if_kdf. Implement the PIN
cache.
(do_change_pin): Clear the PIN cache. Do not change the PIN here.
Lots of adjustments to cope with the chnaged pin2hash_if_kdf.
(do_sign): Do not cache the PIN here.
--
Note that some of the changes are required because we can't rely that
the PIN is always cached in secure memory. Thus it is better to do an
explicit wipe.
Testing the PIN cache can currently only be done my modifying the
code to do a verification for each operation. Only some basic testing
has been done. Clearing the PIN cache is also not fully
implemented. With the forthcoming changes to app-piv we should be
enter able to test the PIN cache.
Missing stuff:
- The agent should be able to selectively clear the cache on a per
slot base.
- We should replace AESWRAP by OCB.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/call-scd.c (handle_pincache_put): Do not decrypt.
(handle_pincache_get): New.
(inq_needpin): Call it.
* scd/command.c (set_key_for_pincache): New.
(pincache_put): Restructure and set key.
(pincache_get): Ditto.
--
This change does away with encrypting and decrypting the cached PIN in
the agent. There is no need for this we can simply cache the
cryptogram and let scdaemon decrypt after retrieving it from the
agent. This way we do not need to share a key between gpg-agent and
scdaemon and further the cached content is invalid after the scdaemon
process has been replaced.
Signed-off-by: Werner Koch <wk@gnupg.org>
* scd/command.c (pincache_put): New. Uses a dummy key for now.
(pincache_get): New.
* scd/app.c (select_application): Flush the PIN cache.
(scd_update_reader_status_file): Ditto.
(maybe_switch_app): Call the new prep_reselect function.
(app_write_learn_status): Ditto.
* scd/app-openpgp.c (cache_pin): New helper to cache a PIN.
(verify_chv2): Call it.
(verify_chv3): Call it.
(clear_chv_status): Call it.
(do_change_pin): Call it.
* scd/app-common.h (struct app_ctx_s): Add function 'prep_select'.
* scd/app-openpgp.c (do_prep_reselect): New stub function.
(app_select_openpgp): Set new stub function.
* scd/app-piv.c (do_prep_reselect): New stub function.
(app_select_piv): Set new stub function.
* scd/app-common.h (struct app_ctx_s): Add parameter ctrl to setattr,
sign, auth, decipher, and check_pin. Change all implementations and
callers to pass such a parameter.
--
This is work in progress.
Signed-off-by: Werner Koch <wk@gnupg.org>
* scd/command.c (open_card_with_request): Use NULL instead of
APPTYPE_NULL.
--
Using the enum is wrong because a string is expected. However it did
harm anything because the value of the enum is 0 and thus equivalent
to NULL.
Signed-off-by: Werner Koch <wk@gnupg.org>
* kbx/backend-sqlite.c: New.
* kbx/Makefile.am (keyboxd_SOURCES): Add it.
(keyboxd_CFLAGS, keyboxd_LDADD): Add SQLite flags.
* kbx/backend.h (enum database_types): Add DB_TYPE_SQLITE.
(be_sqlite_local_t): New typedef.
(struct db_request_part_s): Add field besqlite.
* kbx/backend-support.c (strdbtype): Add string for DB_TYPE_SQLITE.
(be_generic_release_backend): Support SQLite.
(be_release_request): Ditto.
(be_find_request_part): Ditto.
(is_x509_blob): Rename to ...
(be_is_x509_blob): this and make global.
* kbx/frontend.c (kbxd_set_database): Detect ".db" suffix and use that
for SQLite.
(kbxd_search): Support SQLite
(kbxd_store): Ditto.
(kbxd_delete): Ditto.
* kbx/frontend.h (kbxd_store_modes): Move to ...
* kbx/keyboxd.h (enum kbxd_store_modes): here.
* kbx/keyboxd.c (main): USe pubring.db for now. This is a temporary
hack.
* kbx/backend-kbx.c (be_kbx_delete): Remove unused var cert.
--
Take care: This is not finished and in particular filling the database
takes quite long.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/keylist.c (list_one): Probe for a secret key in --with-secret
mode.
--
In contrast to list_all(), list_one() did not tests for a secret key
and took MARK_TRUSTED verbatim as an indication for "secret key
available".
GnuPG-bug: 4061
Signed-off-by: Werner Koch <wk@gnupg.org>
* kbx/frontend.c (the_database): New var.
(db_desc_t): Remove.
(kbxd_add_resource): Renamed to ...
(kbxd_set_database): this. Simplify.
(kbxd_search): Change to use only one database.
(kbxd_store): Ditto.
(kbxd_delete): Ditto.
--
The original implementation was way to complicated and would have only
brought back the problems deciding which database to use for each key.
The new scheme used one configured database and only that. That
database needs to be set right at the start.
Signed-off-by: Werner Koch <wk@gnupg.org>
* sm/gpgsm.h (struct server_control_s): Add field 'current_time'.
* sm/certchain.c (find_up_search_by_keyid): Detect a corner case.
Also simplify by using ref-ed cert objects in place of an anyfound
var.
--
See the code for a description of the problem. Tested using the certs
from the bug report and various command lines
gpgsm --faked-system-time=XXXX --disable-crl-checks \
-ea -v --debug x509 -r 0x95599828
with XXXX being 20190230T000000 -> target cert too young
with XXXX being 20190330T000000 -> okay
with XXXX being 20190830T000000 -> okay, using the long term cert
with XXXX being 20220330T000000 -> target cert expired
The --disabled-crl-checks option is required because in our a simple
test setting dirmngr does not know about the faked time.
GnuPG-bug-id: 4696
Signed-off-by: Werner Koch <wk@gnupg.org>
* dirmngr/server.c (proc_wkd_get): Print new NOTE status
"wkd_cached_result".
* g10/call-dirmngr.c (ks_status_cb): Detect this and print a not ein
verbose mode.
--
This little patch is helpful to see why a WKD change still does not
work after it has been updated on the server.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/getkey.c (get_pubkey_byname): Extend is_mbox checking.
(get_best_pubkey_byname): Ditto.
--
With this patch it is now possible to use
gpg -e -r '<foo@example.org>'
and auto key locate will find the key. Without that a plain mail
address; i.e.
gpg -e -r 'foo@example.org'
was required.
GnuPG-bug-id: 4726
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/gpg.h (MAX_FORMATTED_FINGERPRINT_LEN): Increase by one.
* g10/keyid.c (format_hexfingerprint): Change v5 key formatting.
--
The new format are groups of 5 letters to resemble the CW speed
measurement rules. Many hams are used to such blocks from their CW
lessons. Note that there is no double blank in the middle as with the
v4 keys.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/call-keyboxd.c (struct store_parm_s): New.
(store_inq_cb): New.
(keydb_update_keyblock): Implement the keyboxd part.
(keydb_insert_keyblock): Ditto.
(keydb_delete_keyblock): Ditto.
(keydb_search_reset): Clear ubid flag. Also use the correct union
member for building the search string.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/util.h (UBID_LEN): New. Use it at all places.
* kbx/keybox-blob.c (create_blob_finish): Do not write the UBID item.
* kbx/keybox-dump.c (print_ubib): Remove.
(_keybox_dump_blob): Do not print the now removed ubid flag.
* kbx/keybox-search-desc.h (struct keydb_search_desc): Use constants
for the size of the ubid and grip.
* kbx/keybox-search.c (blob_cmp_ubid): New.
(has_ubid): Make it a simple wrapper around blob_cmp_ubid.
(keybox_get_data): Add arg 'r_ubid'.
* kbx/frontend.h (enum kbxd_store_modes): New.
* kbx/kbxserver.c (cmd_store): Add new option --insert.
* kbx/backend-cache.c (be_cache_initialize): New.
(be_cache_add_resource): Call it here.
* kbx/backend-kbx.c (be_kbx_seek): Remove args 'fpr' and 'fprlen'.
(be_kbx_search): Get the UBID from keybox_get_data.
* kbx/backend-support.c (be_fingerprint_from_blob): Replace by ...
(be_ubid_from_blob): new. Change all callers.
* kbx/frontend.c (kbxd_add_resource): Temporary disable the cache but
use the new cache init function.
(kbxd_store): Replace arg 'only_update' by 'mode'. Seek using the
ubid. Take care of the mode.
--
It turned out that using the hash of the entire blob was not helpful.
Thus we redefine the Unique-Blob-ID (UBID) as the primary fingerprint
of the blob. In case this is a v5 OpenPGP key a left truncated
version of the SHA-256 hash is used; in all other cases the full SHA-1
hash. Using a SHA-256 hash does not make sense because v4 keys are
and will for some time be the majority of keys and thus padding them
with zeroes won't make any difference. Even if fingerprint collisions
can eventually be created we will assume that the keys are bogus and
that it does not make sense to store its twin also in our key storage.
We can also easily extend the update code to detect a collision and
reject the update.
Signed-off-by: Werner Koch <wk@gnupg.org>
* dirmngr/dirmngr.c (time_for_housekeeping_p): Build with --fwrapv.
Replace protecting macro.
--
I once replaced all JNLIB macros by GPGRT macros. However, I missed
to get this right for the push pragma test and used
GPGRT_GCC_HAVE_PUSH_PRAGMA
instead of the correct
GPGRT_HAVE_PRAGMA_GCC_PUSH
Now after fixing this it turned out that no-strict-overflow is not
anymore supported as oprimize pragma and thus I replaced that too
according to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58454 .
Tested with 32 bit Windows but the real test will be in 2038.
Signed-off-by: Werner Koch <wk@gnupg.org>
* dirmngr/http.h (struct parsed_uri_s): Add flag is_ldap.
* dirmngr/http.c (do_parse_uri): Set flag. Do not error out for a
missing slashes in an http scheme if NO_SCHEME_CHECK is active.
* dirmngr/t-http.c (main): Print new flag.
* dirmngr/ks-engine-ldap.c (ks_ldap_help): Use flag instead of
checking the scheme.
* dirmngr/ldap-parse-uri.c (ldap_uri_p): Re-implement using
http_parse_uri.
* dirmngr/t-ldap-parse-uri.c (main): Add option --verbose.
--
This patch merely remove the separate parser for checking for an LDAP
scheme. It is better to let our generic URI parser handle this. Also
fixes this bug
|| url[4] == 'i' || url[4] == 'i')
to make the rarely used ldapi scheme case-insensitive.
More changes to the LDAP code are planned.
Signed-off-by: Werner Koch <wk@gnupg.org>
* configure.ac: Do not build dirmngr if no TLS is available.
* dirmngr/http.c: Remove all uses of the USE_TLS macro.
Signed-off-by: Werner Koch <wk@gnupg.org>
* kbx/keybox.h: Fix naming of arguments.
* scd/ccid-driver.c (print_error) [TEST]: Add missing break. Note
that this is anyway an impossible case.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/call-pinentry.c (atfork_cb): Factor code out to ...
(atfork_core): new.
--
We convey certain envvars directly via the environment to Pinentry and
thus they don't show up in the Assuan logging. Because we better
don't call a logging function in an atfork handle, this patch splits
the code up and uses the same code to display what was done in at fork
after the connection has been established.
Signed-off-by: Werner Koch <wk@gnupg.org>
* doc/DETAILS: Specify new status code "NOTE".
* dirmngr/ks-engine-http.c (ks_http_fetch): Print a NOTE status for a
bad TLS certificate.
* g10/call-dirmngr.c (ks_status_cb): Detect this status.
--
For example a
gpg -v --locate-external-keys dd9jn@posteo.net
now yields
gpg: Note: server uses an invalid certificate
gpg: (further info: bad cert for 'posteo.net': \
Hostname does not match the certificate)
gpg: error retrieving 'dd9jn@posteo.net' via WKD: Wrong name
gpg: error reading key: Wrong name
(without -v the "further info" line is not shown). Note that even
after years Posteo is not able to provide a valid certificate for
their .net addresses. Anyway, this help to show the feature.
Signed-off-by: Werner Koch <wk@gnupg.org>
* dirmngr/http.c: Include dirmngr-status.h
(http_prepare_redirect): Emit WARNING status lines for redirection
problems.
* dirmngr/http.h: Include fwddecl.h.
(struct http_redir_info_s): Add field ctrl.
* dirmngr/ks-engine-hkp.c (send_request): Set it.
* dirmngr/ks-engine-http.c (ks_http_fetch): Set it.
* g10/call-dirmngr.c (ks_status_cb): Detect the two new warnings.
--
This should make it easier to diagnose problems with bad WKD servers.
Signed-off-by: Werner Koch <wk@gnupg.org>
* dirmngr/dns-stuff.c (cached_inet_support): New variable.
(dns_stuff_housekeeping): New.
(check_inet_support): New.
* dirmngr/http.c (connect_server): Use only detected interfaces.
* dirmngr/dirmngr.c (housekeeping_thread): Flush the new cache.
--
This currently works only for Windows but that is where users really
ran into problems. The old workaround was to configure disable-ipv4
or disable-ipv6.
Signed-off-by: Werner Koch <wk@gnupg.org>
* dirmngr/dirmngr.c (parse_ldapservers_file): Tune down error
in case no such file exists.
--
This is hardly an error because by default such a file
does not exist. This caused trouble in support because
people think "Oh the system prints an error".
GnuPG-Bug-Id: T3539
* configure.ac (USE_LDAPWRAPPER): Remove ac_define and conditional.
* dirmngr/Makefile.am: Remove USE_LDAPWRAPPER and considere true.
* dirmngr/ldap-wrapper-ce.c: Remove.
* dirmngr/ldap-wrapper.c: Remove USE_LDAPWRAPPER stuff. Minor chnages
to debug output.
* dirmngr/dirmngr_ldap.c: Remove USE_LDAPWRAPPER stuff. Remove
my_ldap macros.
(fetch_ldap) [W32]: Use ldap_sslinit.
--
Note that the ldap_sslinit stuff for Windows has not been tested.
This patch removes the whole wrapper-but-linked business because we do
not need it anymore. WindowsCE for which we used it is pretty dead.
Signed-off-by: Werner Koch <wk@gnupg.org>
* dirmngr/ldapserver.h: Rename LDAPSERVER_OPT.
--
Using _OPT is not good because it does not describe for what this
configuration group is used. _SESSION immediately shows the purpose.
Signed-off-by: Werner Koch <wk@gnupg.org>
* sm/gpgsm.h (struct keyserver_spec): Add field use_ldaps.
* sm/gpgsm.c (parse_keyserver_line): Parse flags.
* sm/call-dirmngr.c (prepare_dirmngr): Send ldaps flag to the dirmngr.
* dirmngr/dirmngr.h (struct ldap_server_s): Add field use_ldaps.
* dirmngr/ldapserver.c (ldapserver_parse_one): Parse flags.
* dirmngr/ldap.c (start_cert_fetch_ldap): Call wrapper with --tls.
* dirmngr/dirmngr_ldap.c: New option --tls.
(fetch_ldap): Make use of that option.
--
There was no way to specify an LDAPS server in
dirmngr_ldapserver.socnf or with gpgsm's --keyserver option. This
patch fixes this. Eventually we should allow to replace host and port
by a partial URI in the same way ldap_initialize does it. For backward
compatibility we do not yet do that.
Although the dirmngr code accepts an URL (eg. taken from a
certificate), I can't see how the scheme was ever used. Thus the
patch also detects an ldaps scheme and uses this. That part has not
been tested, though.
Signed-off-by: Werner Koch <wk@gnupg.org>
