* common/iobuf.c (iobuf_read): Handle a case with IOBUF_INPUT_TEMP.
(iobuf_write): Handle a case with IOBUF_OUTPUT_TEMP.
--
GnuPG-bug-id: 5941
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* common/t-ssh-utils.c (FLAGS_NOFIPS): New.
(sample_keys): Add flags member.
(main): Detect if libgcrypt is in FIPS mode, try SHA256 fingerprints
first and expect the MD5 ones will fail.
--
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
* common/exechelp-w32.c (handle_to_fd): Use intptr_t.
(gnupg_wait_processes): Fix to use pid_to_handle.
--
Both of original MinGW and MinGW-W64 use intptr_t for the first
argument of _open_osfhandle. So, intptr_t is better here.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* common/dotlock.c (dotlock_take_unix): Do same when same PID
process detects stale lockfile. Add comment.
--
GnuPG-bug-id: 5884
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* common/mapstrings.c (struct intmapping_s): New.
(map_static_strings): New.
* common/stringhelp.c (do_strconcat): Rename to ...
(vstrconcat): this and make global.
* common/t-mapstrings.c (test_map_static_strings): New test.
* common/dotlock.c (read_lockfile): Return the file descriptor when
R_FD is available.
(dotlock_take_unix): Check the case the lockfile was already removed.
--
GnuPG-bug-id: 5884
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* common/t-dotlock.c (lock_and_unlock): Use usleep and faster.
Loop at least once. Use getrandom for random time.
(main): Add new option --one-shot to run lock/unlock once.
--
GnuPG-bug-id: 5884
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* common/iobuf.c (filter_flush): Remove "src_len == 0" check.
* g10/compress-bz2.c (do_compress): Exit early if flush not
forced and input length is zero.
* g10/compress.c (do_compress): Likewise.
--
Remove "(src_len == 0)" check in filter_flush which was
introduced to fix compress failure caused by zero length
flush from iobuf_close. However this check broke enarmoring
file with length of zero. Patch instead fixes zero length
flush problem in compress filters.
GnuPG-bug-id: T5828
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
* common/iobuf.c (filter_flush): Use 'iobuf->e_d' if configured.
(iobuf_write): Configure 'iobuf->e_d' for 'filter_flush' if
'iobuf->d.buf' is empty and external buffer is larger than threshold.
--
Zero-copy operation in iobuf_write() and filter_flush() allow bypassing
'iobuf->d.buf' for greater performance. This mainly helps OCB
performance where additional memory copies through iobuf stack
can take significant portion of program time.
GnuPG-bug-id: T5828
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
* common/iobuf.h (iobuf_struct): Add 'e_d' substructure and members.
* common/iobuf.c (IOBUF_ZEROCOPY_THRESHOLD): New.
(iobuf_alloc): Clear 'iobuf->e_d'.
(underflow_target): Use 'iobuf->e_d' when configured to bypass copying
through 'iobuf->d.buf'.
(iobuf_read): Configure 'iobuf->e_d' for 'underflow' if 'iobuf->d.buf'
is empty and external buffer is larger than threshold.
--
Zero-copy operation in iobuf_read() and underflow() allow bypassing
'iobuf->d.buf' for greater performance. This mainly helps OCB
performance where additional memory copies through iobuf stack
can take significant portion of program time.
GnuPG-bug-id: T5828
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
* common/iobuf.c (iobuf_copy): Use iobuf buffer size for temporary
buffers.
* g10/plaintext.c (handle_plaintext, do_hash): Likewise.
* g10/sign.c (sign_file): Likewise.
--
As iobuf will have zerocopy operation for read/write, it is better to
use same size buffers as iobuf for temporary copy buffers.
GnuPG-bug-id: T5828
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
* common/i18n.c (i18n_init): Use gnupg_localedir() instead of LOCALEDIR.
(i18n_localegettext): Ditto.
* tools/gpgconf-comp.c (my_dgettext): Ditto.
--
On Unix, gnupg_localedir() returns the locale directory relative to
the root directory of the gnupg installation if specified in the
gpgconf.ctl. Otherwise, it returns the built-in LOCALEDIR.
GnuPG-bug-id: 5999
* dirmngr/dirmngr.h (struct fingerprint_list_s): Add field binlen.
(opt): Add field ignored_certs.
* dirmngr/dirmngr.c: Add option --ignore-cert
(parse_rereadable_options): Handle that option.
(parse_ocsp_signer): Rename to ...
(parse_fingerprint_item): this and add two args.
* dirmngr/certcache.c (put_cert): Ignore all to be igored certs.
Change callers to handle the new error return.
--
This option is useful as a workaround in case we ill run into other
chain validation errors like what we fixed in
GnuPG-bug-id: 5639
* common/homedir.c (MYPROC_SELF_EXE): New.
(unix_rootdir): Use it here. Also support GNUPG_BUILD_ROOT as
fallback.
--
In addition this adds a fallback method for AIX etc which do not have
an easy way to get the info.
* common/homedir.c (unix_rootdir): Add arg want_sysconfdir.
(gnupg_sysconfdir): Return it.
--
Our regression test suite has the problem that we can't disable the
use of the global config files or test them using the regualr
binaries. This new keyword will allow us to overcome the problem.
* common/homedir.c (unix_rootdir): New.
(gnupg_bindir): Use it.
(gnupg_libexecdir): Use it.
(gnupg_libdir): Use it.
(gnupg_datadir): Use it.
(gnupg_localedir): Use it.
--
This feature is useful for building and using an AppImage version of
gnupg and probably also for some other use cases.
GnuPG-bug-id: 5999
Here is a sample gpgconf.ctl file
--8<---------------cut here---------------start------------->8---
# gpgconf.ctl
#
# This file is used to change the directories where the gpg components
# are installed. It does not change the configuration directories.
# The file is expected in the same directory as gpgconf. The physical
# installation directories are evaluated and no symlinks. Blank lines
# and lines starting with pound signed are ignored. No errors are
# printed for unknown keywords or commands. The only defined key for
# now is "rootdir" which must be followed by one optional space, an
# equal sign, and the value for the root directory. Environment
# variables are substituted in standard shell manner, the final value
# must start with a slash, trailing slashed are stripped.
rootdir = $APPDIR/gnupg
--8<---------------cut here---------------end--------------->8---
* common/stringhelp.c (substitute_envvars): New. Based on code in
gpg-connect-agent.
* common/t-stringhelp.c: Include sysutils.h.
(test_substitute_envvars): New.
--
GnuPG-bug-id: 5599
* common/membuf.c (put_membuf): Allow NULL for the second arg.
--
There has been such a use case in keybox-blob.c.
Signed-off-by: NIIBE Yutaka <gniibe@fsij.org>
* common/recsel.c (recsel_parse_expr): Add flag -t.
* common/stringhelp.c: Remove assert.h.
(strtokenize): Factor code out to do_strtokenize.
(strtokenize_nt): New.
(do_strtokenize): Add arg trim to support the strtokenize_nt.
* common/t-stringhelp.c (test_strtokenize_nt): New test cases.
* tools/wks-util.c (wks_list_key): Use strtokenize_nt and the recsel
flag -t.
--
This fixes a bug with user ids with leading spaces because:
wks-client lists all mail addresses from the key and matches them to the
requested mail address.
If there are several user-ids all with the same mail address
wks-client picks one of them and then extracts exactly that user id.
However, here it does not match by the mail address but by the full
user-id so that we can be sure that there will be only one user-id in
the final key.
The filter built expression unfortunately strips leading blanks but
requires a verbatim match. Thus it won't find the user id again and
errors out.
The new -t flag and a non-trimming strtokenize solves the problem.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/homedir.c (is_gnupg_default_homedir): Use standard_homedir
instead of the constant which makes a difference on Windows.
(_gnupg_socketdir_internal) [W32]: Move the directory to LOCAL_APPDATA.
(gnupg_cachedir): Remove unsued function.
* common/sysutils.c (gnupg_rmdir): New.
* tools/gpgconf.c (main): s/rmdir/gnupg_rmdir/.
--
That is actually a more correct directory than APPDATA. This fixes
a problem with installations where the APPDATA is non a network drive
and the resulting socket filename is truncated in our socket helper
function (because we use sockaddr also for our local socket
emulation on Windows).
LOCAL_APPDATA is expected to be on the local box and thus in the
majority of cases the resulting socket file name will be short enough.
GnuPG-bug-id: 5537
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/session-env.c (stdenvnames): Add XDG_SESSION_TYPE and
QT_QPA_PLATFORM.
--
On Unix systems (except Darwin), Qt uses those two environment
variables additionally to DISPLAY and WAYLAND_DISPLAY to figure out
whether to use X11 or Wayland. For example, QT_QPA_PLATFORM needs
to be set to "wayland" to make Qt use Wayland on Gnome.
GnuPG-bug-id: 3659
* kbx/keybox-search-desc.h (struct keydb_search_desc): New flag
name_used.
* common/userids.c (classify_user_id): Set flag.
* kbx/kbxserver.c (struct search_backing_store_s): New.
(cmd_search): use a backing store for the const pointers.
(kbxd_start_command_handler): Release the backing store.
--
Well, the search object partly uses buffers but also const
pointers (for strings and the serial number). This when assigning
such objects to an another one we should really take a deep copy and
not just copy the pointer. The more clean solution would have been to
provide a storage option the search object but that needs checking the
code at too many places so that I decided to use a separate backing
store array here.
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/t-sexputil.c (test_ecc_uncompress): Ignore unknwon curve
errors.
--
For unknown reasons some versions of Fedora or RHEL provide a patched
version of Libgcrypt with Brainpool support removed. We better ignore
this error because it is a regression in 2.2.28 although that older
versions could not used Brainpool keys.
GnuPG-bug-id: 5502
Signed-off-by: Werner Koch <wk@gnupg.org>
* common/ttyio.c (w32_write_console): Fallback to WriteConsoleA on
error.
--
To test this switch the Windows Console to "legacy mode"
set LANG=de
gpg --card-edit
and enter an invalid command. The response contains an Umlaut and old
Windows versions (and the legacy console) don't have a proper font
installed for this. Without this patch this runs into a log_fatal
error.
The mitigation we implement is to fallback to WriteConsoleA, that is
accepting wrong encoding and to print a note about the problem.
GnuPG-bug-id: 5491
* common/keyserver.h: Remove.
* sm/gpgsm.h (struct keyserver_spec): Remove.
(opt): Change keyserver to a strlist_t.
* sm/gpgsm.c (keyserver_list_free): Remove.
(parse_keyserver_line): Remove.
(main): Store keyserver in an strlist.
* sm/call-dirmngr.c (prepare_dirmngr): Adjust for the strlist. Avoid
an ambiguity in dirmngr by adding a prefix if needed.
* g10/options.h (struct keyserver_spec): Move definition from
keyserver.h to here. Remove most fields.
* g10/keyserver.c (free_keyserver_spec): Adjust.
(cmp_keyserver_spec): Adjust.
(parse_keyserver_uri): Simplify.
(keyidlist): Remove fakev3 arg which does not make any sense because
we don't even support v3 keys.
--
We now rely on the dirmngr to parse the keyserver specs. Thus a bad
specification will not be caught immediately. However, even before
that dirmngr had stricter tests.
Signed-off-by: Werner Koch <wk@gnupg.org>
Ported-from: 9f586700ec4ceac97fd47cd799878a8847342ffa
* common/compliance.c (gnupg_cipher_is_allowed): Allow GCM for gpgsm
in decrypt mode.
* tests/cms/samplemsgs/pwri-sample.gcm.p7m: Remove duplicated authtag
--
We allow GCM in de-vs mode for decryption although this has not been
evaluation. It is decryption and thus no serious harm may happen.
