* common/homedir.c (w32_rootdir): Use Unicode fucntion not only for
WinCE.
--
This uses the same code We used for WindowsCE. It has not been tested
with a Unicode requiring installation directory.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/gpg.c (list_config): Replace print_sanitized_string2 by
es_write_sanitized.
* common/stringhelp.c (print_sanitized_buffer2): Remove.
(print_sanitized_buffer, print_sanitized_utf8_buffer): Remove.
(print_sanitized_utf8_buffer, print_sanitized_utf8_string): Remove.
(print_sanitized_string): Remove.
* sm/certdump.c (print_dn_part, print_dn_parts): Remove arg FP.
(pretty_print_sexp, gpgsm_print_name2, gpgsm_print_name): Remove.
--
Mixing stdio and estream is never a good idea. This fix also allows
us to remove a lot of garbage.
Reported-by: Jason A. Donenfeld <Jason@zx2c4.com>
GnuPG-bug-id: 1822
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/iobuf.c: (iobuf_open): initialize len
--
In iobuf_open, IOBUFCTRL_DESC and IOBUFCTRL_INIT commands are invoked
(via file_filter()) on fcx, passing in a pointer to an uninitialized
len.
With these two commands, file_filter doesn't actually do anything with
the value of len, so there's no actual risk of use of uninitialized
memory in the code as it stands.
However, some static analysis tools might flag this situation with a
warning, and initializing the value doesn't hurt anything, so i think
this trivial cleanup is warranted.
Debian-Bug-Id: 773469
* common/dotlock.c: (dotlock_create_unix) avoid double-close()
in unusual situations.
--
close(2) says:
close() should not be retried after an EINTR since this may
cause a reused descriptor from another thread to be closed.
Before this patch was applied, if close(fd) failed with EINTR, it
would be closed again in the write_failed: block.
It could also have been closed a second time in the case that
(use_hardlinks_p (h->tname)) evaluated to something other than 0 or 1.
This patch avoids both of those scenarios.
Note that close() could still be called twice on the same file
descriptor if the first close(fd) fails but errno is not EINTR. I'm
not sure the right thing to do in that scenario. An alternate
resolution could be to unequivocally set fd to -1 after the first
failed close(fd), avoiding the errno == EINTR test.
Debian-Bug-Id: 773423
* agent/command-ssh.c (setup_ssh_env): Move code to ...
* agent/gpg-agent.c (agent_copy_startup_env): .. new function. Change
calllers.
* agent/command.c (start_command_handler): Call that fucntion for
restricted connections.
--
A remote connection is and should not be able to setup the local
session environment. However, unless --keep-display is used we would
be left without an environment and thus pinentry can't be used. The
fix is the same as used for ssh-agent connection: We use the default
environment as used at the startup of the agent.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/util.h (GPG_ERR_OBJ_TERM_STATE): New.
* scd/iso7816.c (map_sw): Add this error code.
* scd/app-openpgp.c (do_getattr): Return the life cycle indicator.
* scd/app.c (select_application): Allow a return value of
GPG_ERR_OBJ_TERM_STATE.
* scd/scdaemon.c (set_debug): Print the DBG_READER value.
* g10/call-agent.c (start_agent): Print a status line for the
termination state.
(agent_scd_learn): Make arg "info" optional.
(agent_scd_apdu): New.
* g10/card-util.c (send_apdu): New.
(factory_reset): New.
(card_edit): Add command factory-reset.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/http.c (send_request): Print TLS alert info
(connect_server): Detect bogus DNS entry.
--
1. Prints the TLS alert description.
2. Detect case where the DNS returns an IP address but the server is
not reachable at this address. This may happen for a server which
is reachable only at IPv6 but but the local machine has no full
IPv6 configuration.
* agent/agent.h (opt): Add field extra_socket.
(server_control_s): Add field restricted.
* agent/command.c: Check restricted flag on many commands.
* agent/gpg-agent.c (oExtraSocket): New.
(opts): Add option --extra-socket.
(socket_name_extra): New.
(cleanup): Cleanup that socket name.
(main): Implement oExtraSocket.
(create_socket_name): Add arg homedir and change all callers.
(create_server_socket): Rename arg is_ssh to primary and change
callers.
(start_connection_thread): Take ctrl as arg.
(start_connection_thread_std): New.
(start_connection_thread_extra): New.
(handle_connections): Add arg listen_fd_extra and replace the
connection starting code by parameterized loop.
* common/asshelp.c (start_new_gpg_agent): Detect the use of the
restricted mode and don't fail on sending the pinentry environment.
* common/util.h (GPG_ERR_FORBIDDEN): New.
* common/openpgp-oid.c (openpgp_oid_to_str): Fix unsigned underflow.
* common/t-openpgp-oid.c (BADOID): New.
(test_openpgp_oid_to_str): Add test cases.
--
The code has an obvious error by not considering invalid encoding for
arc-2. A first byte of 0x80 can be used to make a value of less then
80 and we then subtract 80 from that value as required by the OID
encoding rules. Due to the unsigned integer this results in a pretty
long value which won't fit anymore into the allocated buffer.
The fix is obvious. Also added a few simple test cases. Note that we
keep on using sprintf instead of snprintf because managing the
remaining length of the buffer would probably be more error prone than
assuring that the buffer is large enough. Getting rid of sprintf
altogether by using direct conversion along with membuf_t like code
might be possible.
Reported-by: Hanno Böck
Signed-off-by: Werner Koch <wk@gnupg.org>
Ported from libksba commit f715b9e156dfa99ae829fc694e5a0abd23ef97d7
* common/argparse.c (initialize): Use correct value.
--
This avoids a dead path in the argparse code.
It's not clear that this is needed, however, since
ARGPARSE_AMBIGUOUS_COMMAND is never actually used in the code.
Another approach would be to trim out ARGPARSE_AMBIGUOUS_COMMAND
entirely.
* gl/: Remove entire tree.
* configure.ac: Remove gnulib tests and the gl/ Makefile.
(setenv): Add to AC_CHECK_FUNCS.
* autogen.rc (extra_aclocal_flags): Set to empty.
* Makefile.am (ACLOCAL_AMFLAGS): Remove -I gl/m4
(SUBDIRS): Remove gl/.
* agent/Makefile.am (common_libs): Remove ../gl/gnulib.a
* common/Makefile.am (t_common_ldadd): Ditto.
* dirmngr/Makefile.am (dirmngr_LDADD): Ditto.
(dirmngr_ldap_LDADD, dirmngr_client_LDADD): Ditto.
* g10/Makefile.am (needed_libs): Ditto.
* g13/Makefile.am (g13_LDADD): Ditto.
* kbx/Makefile.am (kbxutil_LDADD): Ditto.
($(PROGRAMS)): Ditto.
* scd/Makefile.am (scdaemon_LDADD): Ditto.
* sm/Makefile.am (common_libs): Ditto.
* tools/Makefile.am (common_libs, commonpth_libs): Ditto.
* agent/gpg-agent.c: Remove "mkdtemp.h"
* g10/exec.c: Ditto.
* scd/scdaemon.c: Ditto.
* tools/symcryptrun.c: Ditto.
* common/sysutils.c: Remove "setenv.h"
* common/t-timestuff.c: Use putenv if setenv is not available.
--
gnulib has always been a cause of trouble in GnuPG because we used
only a very few functions and the complex include machinery of gnulib
is quite complex and the cause for many build problems for example on
OS X. This is not gnulib's fault but due to our limited use of gnulib
and that we only rarely update the gnulib code to avoid regressions.
In part two we will address the functions
mkdtemp
setenv
unsetenv
strpbrk
which may bot be implemented on all platforms. They are not required
on a libc based system.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/http.c (write_server) [W32]: Rework to use send() instead of
write even when build with npth.
(cookie_read) [W32]: Rework to use recv() instead of read even when
build with npth.
* configure.ac <adns>: Use adns_free as probe function for libadns.
(HAVE_ADNS_FREE): Remove bogus tests to set this and remove the macro.
(ADNSLIBS): Do not ac_subst - it is only used within configure.
--
adns_free is required on Windows anyway (for robustness reasons) and
it has been around for so long now that we do not need a separate
test. An upstream adns 1.5 has meanwhile been release but I doubt that
this has the required Windows code - and it is not libtool based
anyway.
--
Also fixed some of my own copyright notices due to the termination of
my assignment. The one displayed by --version is kept at FSF because
we had contributors in 2014 with FSF assignments and it gives the FSF
some visibility.
* agent/gpg-agent.c (oDebugQuickRandom): New.
(opts): New option --debug-quick-random.
(main): Use new option.
* common/asshelp.c (start_new_gpg_agent): Add hack to pass an
additional argument for the agent name.
* tests/openpgp/defs.inc: Pass --debug-quick-random to the gpg-agent
starting parameters.
* tests/openpgp/version.test: Ditto.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/argparse.h (ARGPARSE_INVALID_ARG): New.
* common/argparse.c: Include limits h and errno.h.
(initialize): Add error strings for new error constant.
(set_opt_arg): Add range checking.
Signed-off-by: Werner Koch <wk@gnupg.org>
* agent/agent.h (opt): Remove field use_standard_socket.
* agent/command.c (cmd_killagent): Always allow killing.
* agent/gpg-agent.c (main): Turn --{no,}use-standard-socket and
--write-env-file into dummy options. Always return true for
--use-standard-socket-p. Do not print the GPG_AGENT_INFO envvar
setting or set that envvar.
(create_socket_name): Simplify by removing non standard socket
support.
(check_for_running_agent): Ditto.
* common/asshelp.c (start_new_gpg_agent): Remove GPG_AGENT_INFO use.
* common/simple-pwquery.c (agent_open): Ditto.
* configure.ac (GPG_AGENT_INFO_NAME): Remove.
* g10/server.c (gpg_server): Do not print the AgentInfo comment.
* g13/server.c (g13_server): Ditto.
* sm/server.c (gpgsm_server): Ditto.
* tools/gpgconf.c (main): Simplify by removing non standard socket
support.
--
The indented fix to allow using a different socket than the one in the
gnupg home directory is to change Libassuan to check whether the
socket files exists as a regualr file with a special keyword to
redirect to another socket file name.
* configure.ac (NEED_NTBTLS_ABI, NEED_NTBTLS_VERSION): New.
(HTTP_USE_NTBTLS): New. Prefer over GNUTLS.
* m4/ntbtls.m4: New.
* m4/Makefile.am (EXTRA_DIST): Add new file.
* common/http.c: Add conditionals to eventually use NTBTLS.
--
This is only the configure stuff. If you have NTBTLS installed GNUTLS
will not be used but there won't be any https support either :-(.
This patch is used to have a real world test bench for the forthcoming
library.
* common/exechelp-posix.c: Remove weak pragmas.
* common/sexputil.c (make_canon_sexp_from_rsa_pk): Remove double
const.
--
We do not use Pth anymore and thus there is no more need for the weak
pragmas.
* common/openpgpdefs.h (PUBKEY_ALGO_EDDSA): Change to 22.
* g10/keygen.c (ask_curve): Reword the Curve25519 warning note.
--
In the hope that the IETF will eventually assign 22 for EdDSA using
the draft-koch-eddsa-for-openpgp-01 specs we start using this number.
* common/http.c (http_session_ref): Allow for NULL arg.
--
We always test for a an existing session and thus passing NULL as
session object should be allowed.
Reported-by: Jens Lechtenboerger
* common/openpgp-oid.c (struct oidtable): New.
(openpgp_curve_to_oid): Rewrite and allow OID as input.
(openpgp_oid_to_curve): Make use of the new table.
--
Due to the previous change we now usually store the OID with the
private key and not the name. Thus during import we do not anymore
need to map the name to an oid but can use the oid directly. We fix
that by extending openpgp_curve_to_oid to allow an oidstr as input.
* common/estream-printf.c, common/estream-printf.h: Change license.
* common/estream.c, common/estream.h: Ditto.
--
g10 Code is the sole copyright holder of Libestream and thus as CEO I
have the rights to to change the license. This copy here in GnuPG is
currently the most current one thus the change is recorded in this
repository. This change is also deemed valid for all older versions.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/t-timestuff.c (test_timegm): Use timegm if available.
(main): Set TX to UTC if timegm is not available.
--
On OpenBSD 5.3 i386 that test failed due to the use of mktime.
Reported-by: Claus Assmann
* common/iobuf.c (direct_open): Add arg MODE700.
(iobuf_create): Ditto.
* g10/openfile.c (open_outfile): Add arg RESTRICTEDPERM. Change call
callers to pass 0 for it.
* g10/revoke.c (gen_desig_revoke, gen_revoke): Here pass true for new
arg.
* g10/export.c (do_export): Pass true for new arg if SECRET is true.
--
GnuPG-bug-id: 1653.
Note that this works only if --output has been used.
* g10/getkey.c (get_user_id): Do not call xmalloc with 0.
* common/xmalloc.c (xmalloc, xcalloc): Take extra precaution not to
pass 0 to the arguments.
--
The problem did not occur in 1.x because over there the xmalloc makes
sure to allocate at least one byte. With 2.x for most calls the
xmalloc of Libgcrypt is used and Libgcrypt returns an error insteead
of silent allocating a byte. Thus gpg 2.x bailed out with an
"Fatal: out of core while allocating 0 bytes".
The extra code in xmalloc.c is for more robustness for the other
xmalloc calls.
* common/http.c (http_open): Add arg httphost.
(http_open_document): Pass NULL for httphost.
(send_request): Add arg httphost. If given, use HTTPHOST instead of
SERVER. Use https with a proxy if requested.
(http_verify_server_credentials): Do not stop at the first error
message.
* dirmngr/ocsp.c (do_ocsp_request): Adjust call to http_open.
* keyserver/curl-shim.c (curl_easy_perform): Ditto.
* dirmngr/ks-engine-http.c (ks_http_fetch): Ditto.
* dirmngr/ks-engine-hkp.c (ks_hkp_help): Ditto.
* common/Makefile.am ($(PROGRAMS)): New rule
(t_http_LDADD): Use libcommontls.a without directory prefix.
* dirmngr/Makefile.am ($(PROGRAMS)): New rule.
* agent/cvt-openpgp.c (try_do_unprotect_arg_s): Add field "curve".
(get_keygrip): Add and use arg CURVE.
(convert_secret_key): Ditto.
(convert_transfer_key): Ditto.
(get_npkey_nskey): New.
(prepare_unprotect): Replace gcrypt functions by
get_npkey_nskey. Allow opaque MPIs.
(do_unprotect): Use CURVE instead of parameters.
(convert_from_openpgp_main): Ditto.
(convert_to_openpgp): Simplify.
* g10/import.c (one_mpi_from_pkey): Remove.
(transfer_secret_keys): Rewrite to use the curve instead of the
parameters.
* g10/parse-packet.c (parse_key): Mark protected MPIs with USER1 flag.
* common/openpgp-oid.c (openpgp_curve_to_oid): Allow the use of
"NIST P-256" et al.
* g10/keygen.c (ask_curve): Add arg ALGO.
(generate_keypair): Rewrite the ECC key logic.
* tests/openpgp/ecc.test: Provide the "ecc" passphrase.
* common/http.c (http_session_t): Add field "refcount".
(_my_socket_new, _my_socket_ref, _my_socket_unref): Add debug code.
(send_request, my_npth_read, my_npth_write): Use SOCK object for the
transport ptr.
(http_session_release): Factor all code out to ...
(session_unref): here. Deref SOCK.
(http_session_new): Init refcount and transport ptr.
(http_session_ref): New. Ref and unref all assignments.
--
Having the reference counted session objects makes it easier for the
application to pass around only an estream. Without that the
application would need to implement an es_onclose machinery for the
session object.
