* common/agent-opt.c: New.
* common/shareddefs.h: New.
* common/Makefile.am: Add new files.
* agent/agent.h: Include shareddefs.h.
(pinentry_mode_t): Factor out to shareddefs.h.
* agent/command.c (option_handler): Use parse_pinentry_mode.
* agent/command-ssh.c (SPEC_FLAG_IS_ECDSA): New.
(struct ssh_key_type_spec): Add fields CURVE_NAME and HASH_ALGO.
(ssh_key_types): Add types ecdsa-sha2-nistp{256,384,521}.
(ssh_signature_encoder_t): Add arg spec and adjust all callers.
(ssh_signature_encoder_ecdsa): New.
(sexp_key_construct, sexp_key_extract, ssh_receive_key)
(ssh_convert_key_to_blob): Support ecdsa.
(ssh_identifier_from_curve_name): New.
(ssh_send_key_public): Retrieve and pass the curve_name.
(key_secret_to_public): Ditto.
(data_sign): Add arg SPEC and change callers to pass it.
(ssh_handler_sign_request): Get the hash algo from SPEC.
* common/ssh-utils.c (get_fingerprint): Support ecdsa.
* agent/protect.c (protect_info): Add flag ECC_HACK.
(agent_protect): Allow the use of the "curve" parameter.
* agent/t-protect.c (test_agent_protect): Add a test case for ecdsa.
* agent/command-ssh.c (ssh_key_grip): Print a better error code.
--
The 3 standard curves are now supported in gpg-agent's ssh-agent
protocol implementation. I tested this with all 3 curves and keys
generated by OpenSSH 5.9p1.
Using existing non-ssh generated keys will likely fail for now. To fix
this, the code should first undergo some more cleanup; then the fixes
are pretty straightforward. And yes, the data structures are way too
complicated.
* agent/command-ssh.c: Remove dirent.h.
(control_file_s): Add struct item.
(rewind_control_file): New.
(search_control_file): Factor code out to ...
(read_control_file_item): New.
(ssh_handler_request_identities): Change to iterate over entries in
sshcontrol.
--
Formerly we scanned the private key directory for matches of entries
in sshcontrol. This patch changes it to scan the sshcontrol file and
thus considers only keys configured there. The rationale for this is
that it is common to have only a few ssh keys but many private keys.
Even if that assumption does not hold true, the scanning of the
sshcontrol file is faster than reading the directory and only then
scanning the ssh control for each directory entry.
* agent/command-ssh.c (SSH_CONTROL_FILE_NAME): New macro to replace
the direct use of the string.
(struct control_file_s, control_file_t): New.
(open_control_file, close_control_file): New. Use them instead of
using fopen/fclose directly.
* agent/gpg-agent.c (main): Pass new envar gnupg_SSH_AUTH_SOCK_by to
an invoked process.
--
This environment variable is useful for debugging if
--use-standard-socket is used (which is the default since 2.1).
Commonly you should have this in your init script (e.g. ~/.bashrc):
unset GPG_AGENT_INFO
unset SSH_AGENT_PID
SSH_AUTH_SOCK="${HOME}/.gnupg/S.gpg-agent.ssh"
export SSH_AUTH_SOCK
The problem is that gpg-agent won't be able to override the
SSH_AUTH_SOCK envvar if gpg-agent has been invoked as
gpg-agent --enable-ssh-support --daemon /bin/bash
To fix this you should instead use this code in the init script:
unset GPG_AGENT_INFO
unset SSH_AGENT_PID
if [ ${gnupg_SSH_AUTH_SOCK_by:-0} -ne $$ ]; then
export SSH_AUTH_SOCK="${HOME}/.gnupg/S.gpg-agent.ssh"
fi
This will work in all cases and thus allows to start gpg-agent for
testing purposes with a different homedir and use this gpg-agent as an
ssh-agent. Example:
GNUPGHOME=$(pwd) gpg-agent --enable-ssh-support --daemon /bin/bash
gnupg_SSH_AUTH_SOCK_by is set to the PID of the exec-ed process and
thus will work safely if called recursively.
* configure.ac (HAVE_BROKEN_TTYNAME): New ac_define set for Android
systems.
* common/util.h (gnupg_ttyname): New macro. Change all callers of
ttyname to use this macro instead.
(ttyname) [W32]: Rename to _gnupg_ttyname and use also if
HAVE_BROKEN_TTYNAME is defined.
* common/simple-pwquery.c (agent_send_all_options): Keep on using
ttyname unless HAVE_BROKEN_TTYNAME is set. This is because this file
may be used standalone.
* scd/command.c (MAXLEN_SETDATA): New.
(cmd_setdata): Add option --append.
* agent/call-scd.c (agent_card_pkdecrypt): Use new option for long
data.
* scd/app-openpgp.c (struct app_local_s): Add field manufacturer.
(app_select_openpgp): Store manufacturer.
(do_decipher): Print a note for broken cards.
--
Please note that I was not able to run a full test because I only have
broken cards (S/N < 346) available.
* findkey.c (agent_public_key_from_file): Fix use of
gcry_sexp_build_array.
--
A test case leading to a segv in Libgcrypt is
gpg-connect-agent \
"READKEY 9277C5875C8AFFCB727661C18BE4E0A0DEED9260" /bye
The keygrip was created by "monkeysphere s", which has a comment.
gcry_sexp_build_array expects pointers to the arguments which is quite
surprising. Probably ARG_NEXT was accidentally implemented wrongly.
Anyway, we can't do anything about it and thus need to fix the check
the users of this function.
Some-comments-by: Werner Koch <wk@gnupg.org>
The asymmetric quotes used by GNU in the past (`...') don't render
nicely on modern systems. We now use two \x27 characters ('...').
The proper solution would be to use the correct Unicode symmetric
quotes here. However this has the disadvantage that the system
requires Unicode support. We don't want that today. If Unicode is
available a generated po file can be used to output proper quotes. A
simple sed script like the one used for en@quote is sufficient to
change them.
The changes have been done by applying
sed -i "s/\`\([^'\`]*\)'/'\1'/g"
to most files and fixing obvious problems by hand. The msgid strings in
the po files were fixed with a similar command.
* agent/trustlist.c (clear_trusttable): New.
(agent_reload_trustlist): Use new function.
(read_trustfiles): Require to be called with lock held.
(agent_istrusted): Factor all code out to ...
(istrusted_internal): new. Add ALREADY_LOCKED arg. Make sure the
table islocked. Do not print TRUSTLISTFLAG stati if called internally.
(agent_marktrusted): Replace calls to agent_reload_trustlist by
explicit code.
--
In contrast to pth, npth does not use recursive mutexes by default.
However, the code in trustlist.c assumed recursive locks and thus we
had to rework it.
GnuPG requires that options are given before other arguments. This
can sometimes be confusing. We now print a warning if we found an
argument looking alike a long option without being preceded by the
stop option. This is bug#1343.
* common/argparse.h (ARGPARSE_FLAG_STOP_SEEN): New.
* common/argparse.c (arg_parse): Set new flag.
* g10/gpg.c (main): Print the warning.
* agent/gpg-agent.c (main): Ditto.
* dirmngr/dirmngr.c (main): Ditto.
* g13/g13.c (main): Ditto.
* scd/scdaemon.c (main): Ditto.
* sm/gpgsm.c (main): Ditto.
* tools/gpg-connect-agent.c (main): Ditto.
* tools/gpgconf.c (main): Ditto.
This is not yet fully implemented. It will eventually allow to
support pinpad equipped readers which do not support variable length
pin lengths.
* agent/protect.c (parse_shadow_info): Add optional arg R_PINLEN and
parse pinlen info. Change all callers to pass NULL for it.
* agent/command.c: Include asshelp.h.
(cmd_pkdecrypt, cmd_genkey, cmd_preset_passphrase)
(pinentry_loopback): Use print_assuan_status for INQUIRE_MAXLEN.
For the shared code parts it is cumbersome to pass an error sourse
variable to each function. Its value is always a constant for a given
binary and thus a global variable makes things a lot easier than the
former macro stuff.
* common/init.c (default_errsource): New global var.
(init_common_subsystems): Rename to _init_common_subsystems. Set
DEFAULT_ERRSOURCE.
* common/init.h: Assert value of GPG_ERR_SOURCE_DEFAULT.
(init_common_subsystems): New macro.
* common/util.h (default_errsource): Add declaration.
* kbx/keybox-defs.h: Add some GPG_ERR_SOURCE_DEFAULT trickery.
Since there isn't a way to prompt the user to insert the smartcard when
pinentry-mode=loopback, return GPG_ERR_CARD_NOT_PRESENT instead of
GPG_ERR_NO_PIN_ENTRY.
* agent/divert-scd.c (ask_for_card): Return GPG_ERR_CARD_NOT_PRESENT
when pinentry-mode=loopback.
Fix pinentry-mode=loopback when cancelling an inquire from scdaemon.
This is similar to commit 4f21f8d but for both protocol command
cancellation and pinentry cancellation.
* agent/call-scd.c (agent_card_pkdecrypt): Check for
GPG_ERR_ASS_CANCELED.
(agent_card_pksign): Ditto.
* agent/gpg-agent.c (get_agent_ssh_socket_name): Use
INVALID_HANDLE_VALUE instead of 0.
(handle_signal) [!HAVE_W32_SYSTEM]: Don't define.
(handle_connections): Port Windows code to NPTH.
* dirmngr/dirmngr.c (handle_connections): Port Windows code to NPTH.
* g13/g13.c (handle_connections): Port Windows code to NPTH.
* scd/scdaemon.c (handle_connections): Port Windows code to NPTH.
This status message is used to inform the client of the maximum length
of an inquired passphrase and is used in pinentry-mode=loopback.
* agent/command.c (pinentry_loopback): Send the INQUIRE_MAXLEN status
message before doing the inquire.
* agent/gpg-agent.c (main): Supply omitted "break" statement for
lc-messages option. Otherwise, control would fall through to the
following oXauthority case and use the same value there.
Similar to commit 29af488 but also fixes PKDECRYPT and PKSIGN.
* agent/call-scd.c (agent_card_pkdecrypt): Check for GPG_ERR_CANCELED
when returning from the PKDECRYPT operation of scdaemon and cancel the
inquire.
(agent_card_pksign): Ditto.
(cancel_inquire): New.
* agent/gpg-agent.c, dirmngr/dirmngr.c, g13/g13.c, scd/scdaemon.c
(USE_GCRY_THREAD_CBS): New macro, defined if
GCRY_THREAD_OPTION_VERSION is 0.
(fixed_gcry_pth_init) [!USE_GCRY_THREAD_CBS]: Don't define.
(main) [!USE_GCRY_THREAD_CBS]: Do not install thread callbacks.
* agent/call-scd.c (pass_status_thru): Pass comment lines verbatim.
* tools/gpg-connect-agent.c (help_cmd_p): New.
(main): Treat an "SCD HELP" the same as "HELP".
* scripts/gitlog-to-changelog: New script. Taken from gnulib.
* scripts/git-log-fix: New file.
* scripts/git-log-footer: New file.
* doc/HACKING: Describe the ChangeLog policy
* ChangeLog: New file.
* Makefile.am (EXTRA_DIST): Add new files.
(gen-ChangeLog): New.
(dist-hook): Run gen-ChangeLog.
Rename all ChangeLog files to ChangeLog-2011.
* command-ssh.c (card_key_available): Change wording of no key
diagnostic.
(ssh_handler_request_identities): Do not call card_key_available
if the scdaemon is disabled.
When this mode is set an inquire will be sent to the client to retrieve
the passphrase. This adds a new inquire keyword "NEW_PASSPHRASE" that the
GENKEY and PASSWD commands use when generating a new key.
The protection used in the exported key used a different iteration
count than given in the S2K field. Thus all OpenPGP keys exported
from GnuPG 2.1-beta can't be imported again. Given that the actual
secret key material is kept in private-keys-v1.d/ the can be
re-exported with this fixed version.
2011-04-20 Marcus Brinkmann <mb@g10code.com>
* command.c (cmd_import_key): Release key from failed import
before converting openpgp private key in the openpgp-private-key
case.
Without Libgcrypt 1.5 is was not possible to use ECC keys. ECC is
major new feature and thus it does not make sense to allow building
with an older Libgcrypt without supporting ECC.
Also fixed a few missing prototypes.
This was a regression in 2.1 introduced due to having the agent do the
signing in contrast to the old "SCD PKSIGN" command which accesses the
scdaemon directly and passed the hash algorithm. The hash algorithm
is used by app-openpgp.c only for a sanity check.
We better do this once and for all instead of cluttering all future
commits with diffs of trailing white spaces. In the majority of cases
blank or single lines are affected and thus this change won't disturb
a git blame too much. For future commits the pre-commit scripts
checks that this won't happen again.
Wrote the ChangeLog 2011-01-13 entry for Andrey's orginal work modulo
the cleanups I did in the last week. Adjusted my own ChangeLog
entries to be consistent with that entry.
Nuked quite some trailing spaces; again sorry for that, I will better
take care of not saving them in the future. "git diff -b" is useful
to read the actual changes ;-).
The ECC-INTEGRATION-2-1 branch can be closed now.
Import and export of secret keys does now work. Encryption has been
fixed to be compatible with the sample messages.
This version tests for new Libgcrypt function and thus needs to be
build with a new Libgcrypt installed.
Quite some changes were needed but in the end we have less code than
before. Instead of trying to do everything with MPIs and pass them
back and forth between Libgcrypt and GnuPG, we know use the
S-expression based interface and make heavy use of our opaque MPI
feature.
Encryption, decryption, signing and verification work with
self-generared keys.
Import and export does not yet work; thus it was not possible to check
the test keys at https://sites.google.com/site/brainhub/pgpecckeys .
Changed order of some conditional to make to put the special case into
the true branch. Indentation changes. Minor other changes to make the
ECC code more similar to the rest of our code.
It builds but many sefltests still fail. Need to fix that before
using it with an ECDH enabled libgcrypt.
[/]
2011-01-21 Werner Koch <wk@g10code.com>
* configure.ac: Need Libgcrypt 1.4.6 due to AESWRAP.
(HAVE_GCRY_PK_ECDH): Add new test.
[agent/]
2011-01-21 Werner Koch <wk@g10code.com>
* cvt-openpgp.c (GCRY_PK_ECDH) [!HAVE_GCRY_PK_ECDH]: New.
[include/]
2011-01-21 Werner Koch <wk@g10code.com>
* cipher.h (GCRY_PK_USAGE_CERT): Remove compatibility macros
because we now require libgcrypt 1.4.6.
(GCRY_PK_ECDH): Add replacement.
The following works:
gpg2 --gen-key (ECC)
gpg2 --list-keys
gpg2 --list-packets ~/.gnupg/pubring.gpg
gpg2 --list-packets <private key from http://sites.google.com/site/brainhub/pgpecckeys>
ECDH doesn't work yet as the code must be re-written to adjust for gpg-agent refactoring.
