* configure.ac: Added --enable-large-secmem option.
* g10/options.h: Add opt.flags.large_rsa.
* g10/gpg.c: Contingent on configure option: adjust secmem size,
add gpg --enable-large-rsa, bound to opt.flags.large_rsa.
* g10/keygen.c: Adjust max RSA size based on opt.flags.large_rsa
* doc/gpg.texi: Document --enable-large-rsa.
--
This is a cherry-pick of 534e2876acc05f9f8d9b54c18511fe768d77dfb5 from
STABLE-BRANCH-1-4 against STABLE-BRANCH-2-0
Some older implementations built and used RSA keys up to 16Kib, but
the larger secret keys now fail when used by more recent GnuPG, due to
secure memory limitations.
Building with ./configure --enable-large-secmem will make gpg
capable of working with those secret keys, as well as permitting the
use of a new gpg option --enable-large-rsa, which let gpg generate RSA
keys up to 8Kib when used with --batch --gen-key.
Debian-bug-id: 739424
Minor edits by wk.
GnuPG-bug-id: 1732
* g10/gpg.c: Cleanup argument parsing.
--
c76117f8b0165fe5cec5e7f234f55f5a4cd7f0ab mistakenly marked
compress-sigs and compress-keys as no-ops on the 2.0.x branch.
These options still have an effect on the 2.0.x branch, and the
duplicate declaration also causes the gpg argument parser to fail when
shortened versions of the option are present, like:
gpg: option "--compress-k" is ambiguous
* g10/gpg.c: Cleanup argument parsing.
--
With c76117f8b0165fe5cec5e7f234f55f5a4cd7f0ab, the GnuPG 2.0.x branch
accidentally introduced a second (identical) argument parser for both
--sk-comments, and for --no-sk-comments.
This caused short versions (e.g. omitting the trailing "s", as gpgme
does) of either command to fail with:
gpg: option "--sk-comment" is ambiguous
* g10/main.h (DEFAULT_DIGEST_ALGO): Use SHA256 in --gnupg and SHA1 in
strict RFC or PGP modes.
* g10/sign.c (make_keysig_packet): Use DEFAULT_DIGEST_ALGO also for
RSA key signatures.
--
(Backported from commit d33246700578cddd1cb8ed8164cfbba50aba4ef3)
* g10/gpg.c: Add config options that should belong in scdaemon.conf
* g10/main.h, g10/misc.c (obsolete_scdaemon_option): New.
--
In gpg2, the following options are only relevant for scdaemon:
reader-port
ctapi-driver
pcsc-driver
disable-ccid
but in gpg1, they are options for gpg itself.
Some users of gpg1 might have these options in their
~/.gnupg/gpg.conf, which causes gpg2 to fail hard if it reads that
config file.
gpg2 should not fail hard, though giving a warning (and suggesting a
move to scdaemon.conf) seems OK.
This patch does *not* reintroduce any documentation for these options
in gpg.texi, even to indicate that they are "dummy" options, since
scdaemon.texi contains the appropriate documentation.
Debian-bug-id: 762844
Program names factored out from obsolete_scdaemon_option to make
reuse without new translations easier. -wk
This is a backport of commit 371c2b14b0347209efd23b4e54e1981a12d7aeab
with parts of 20c6da50d4f6264d26d113d7de606971f719a0ca but without those
which would change existing translated strings. -wk
* g10/trustdb.c (clear_ownertrusts): Init trustdb.
--
This was fixed in 1.4 branch in commit
23191d7851eae2217ecdac6484349849a24fd94a but was not applied to the
2.0 branch that exhibits the same problem. This is actually a hack
to fix a bug introduced with commit 2528178.
GnuPG-bug-id: 1622
* g10/passphrase.c (passphrase_get): Replace sprintf by xasprintf.
--
Without that at least the French translation does not always work
because it requires positional parameters. Windows for example does
not support them as they are not defined by C99 but by POSIX.
* g10/parse-packet.c (parse_attribute): Avoid xmalloc failure and cap
size of packet.
--
Tavis Ormandy reported a fatal error for attribute packets with a zero
length payload. This is due to a check in Libgcrypt's xmalloc which
rejects a malloc(0) instead of silently allocating 1 byte. The fix is
obvious.
In addition we cap the size of attribute packets similar to what we do
with user id packets. OpenPGP keys are not the proper way to store
movies.
Resolved conflicts:
g10/parse-packet.c - indentation. Use plain fprintf.
* g10/keygen.c (gen_elg): Enforce keysize 1024 to 4096.
(gen_rsa): Enforce keysize 1024 to 4096.
(gen_dsa): Enforce keysize 768 to 3072.
--
It was possible to create 16k RSA keys in batch mode. In addition to the
silliness of such keys, they have the major drawback that under GnuPG
and Libgcrypt, with their limited amount of specially secured memory
areas, the use of such keys may lead to an "out of secure memory"
condition.
* g10/keyserver.c (ks_retrieval_filter_arg_s): new.
(keyserver_retrieval_filter): Use new struct and check all
descriptions.
(keyserver_spawn): Pass filter arg suing the new struct.
--
This is a fix for commit 5e933008.
The old code did only work for a single key. It failed as soon as
several keys are specified ("gpg --refresh-keys" or "gpg --recv-key A
B C").
--
GnuPG-bug-id: 1561
Note that this is not a complete solution. The libgpg-error include
directory has now a higher preference but ld may not pick up the right
library if another one is installed. The problem is that the -L
option and the -l options are not emitted separately by
gpg-error-config.
* g10/pkclist.c (build_pk_list): Use more specific reasons codes for
INV_RECP.
--
GnuPG-bug-id: 1650
Note that this patch is a bit more limited than the one in 2.1.
* g10/main.h (import_filter_t): New.
* g10/import.c (import): Add filter callbacks to param list.
(import_one): Ditto.
(import_secret_one): Ditto.
(import_keys_internal): Ditto.
(import_keys_stream): Ditto.
* g10/keyserver.c (keyserver_retrieval_filter): New.
(keyserver_spawn): Pass filter to import_keys_stream()
--
These changes introduces import functions that apply a constraining
filter to imported keys. These filters can verify the fingerprints of
the keys returned before importing them into the keyring, ensuring that
the keys fetched from the keyserver are in fact those selected by the
user beforehand.
Signed-off-by: Stefan Tomanek <tomanek@internet-sicherheit.de>
Re-indention and minor changes by wk.
Resolved conflicts:
g10/import.c
g10/keyserver.c
g10/main.h
* g10/compress.c (do_uncompress): Limit the number of extra FF bytes.
--
A packet like (a3 01 5b ff) leads to an infinite loop. Using
--max-output won't help if it is a partial packet. This patch
actually fixes a regression introduced on 1999-05-31 (c34c6769).
Actually it would be sufficient to stuff just one extra 0xff byte.
Given that this problem popped up only after 15 years, I feel safer to
allow for a very few FF bytes.
Thanks to Olivier Levillain and Florian Maury for their detailed
report.
* g10/photoid.c (show_photos): Set namehash.
* g10/misc.c (pct_expando): Add "%U" expando.
--
This makes is possible to extract all photos ids from a key to
different files.
(cherry picked from commit e184a11f94e2d41cd9266484542631bec23628b5)
Resolved conflicts:
g10/photoid.c - whitespaces
* g10/decrypt-data.c (decrypt_data): Do not distinguish between a bad
MDC packer header and a bad MDC.
--
The separate diagnostic was introduced for debugging a problems. For
explaining an MDC error a single error message is easier to
understand.
* g10/keygen.c (ask_expire_interval): Get the current time after the
prompt.
--
This almost avoid that an entered full ISO timestamp is not used as
given but off by the time the user required to enter the timestamp.
GnuPG-bug-id: 1639
* g10/parse-packet.c (mpi_read): Change error message on overflow.
--
This gets gpg 2.x in sync to what gpg 1.4 does. No need to die for a
broken MPI.
GnuPG-bug-id: 1593
* g10/mainproc.c (list_node): Rework.
--
The old code still merged the first user id into the key packet line
which resulted in all kind of complexity. --fixed-list-mode is
meanwhile the default and thus we also change this part of the code.
GnuPG-bug-id: 1640
* 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.
(cherry picked from commit 99972bd6e9abea71f270284f49997de5f00208af)
Resolved conflicts:
g10/getkey.c - ignore whitespace changes.
* g10/call-agent.c (check_hijacking): New.
(start_agent): Call it.
(membuf_data_cb, default_inq_cb): Move more to the top.
--
Note that GUIs may use the gpg status line
[GNUPG:] ERROR check_hijacking 33554509
to detect this and print an appropriate warning.
* g10/misc.c (pubkey_get_npkey): Map RSA_E and RSA_S to RSA.
(pubkey_get_nskey): Ditto.
(pubkey_get_nsig): Ditto.
(pubkey_get_nenc): Ditto.
(pubkey_nbits): Take care of RSA_E and RSA_S.
--
The problem was that parse_key did not know about RSA_S and thus used
an opaque MPI which later crashed Libgcrypt. It is possible to fix
that also in Libgcrypt but we better do it here as well.
A test key using RSA_S is 0x98EEB6F7D87171CF.
Reported-by: Hanno Böck
* g10/cpr.c (write_status_strings): New.
(write_status_text): Replace code by a call to write_status_strings.
* g10/mainproc.c (proc_encrypted): Remove show_session_key code.
* g10/decrypt-data.c (decrypt_data): Add new show_session_key code.
--
This feature can be used to return the session key for just a part of
a file. For example to downloading just the first 32k of a huge file,
decrypting that incomplete part and while ignoring all the errors
break out the session key. The session key may then be used on the
server to decrypt the entire file without the need to have the private
key on the server.
GnuPG-bug-id: 1389
Signed-off-by: Werner Koch <wk@gnupg.org>
(cherry picked from commit 101a54add351ff62793cbfbf3877787c4791f833)
Resolved Conflicts:
doc/DETAILS - removed
g10/cpr.c - replace estream fucntion by stdio.
g10/mainproc.c - Adjust for changed calling convention.
* g10/tdbio.c (tdbio_set_dbname): Add arg R_NOFILE.
* g10/trustdb.c (trustdb_args): Add field no_trustdb.
(init_trustdb): Set that field.
(revalidation_mark): Take care of a nonexistent trustdb file.
(read_trust_options): Ditto.
(get_ownertrust): Ditto.
(get_min_ownertrust): Ditto.
(update_ownertrust): Ditto.
(update_min_ownertrust): Ditto.
(clear_ownertrusts): Ditto.
(cache_disabled_value): Ditto.
(check_trustdb_stale): Ditto.
(get_validity): Ditto.
* g10/gpg.c (main): Do not create a trustdb with most commands for
trust-model always.
--
This slightly changes the semantics of most commands in that they
won't create a trustdb if --trust-model=always is used. It just does
not make sense to create a trustdb if there is no need for it.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/misc.c (openpgp_pk_algo_name): New. Replace all calls in g10/
to gcry_pk_algo_name by a call to this function.
(map_pk_openpgp_to_gcry): Map algo PUBKEY_ALGO_ELGAMAL_E to GCRY_PK_ELG.
(openpgp_pk_test_algo): Use PUBKEY_ALGO_ELGAMAL_E instead of
GCRY_PK_ELG_E. Return an error for ECC algos.
(openpgp_pk_test_algo2): Return an error for ECC algos.
* g10/gpg.c (build_list): Avoid printing ECC two times.
* include/cipher.h: Do not use GCRY_PK_* macros for PUBKEY_ALGO_*.
--
Due to recent changes to adjust for use with Libgcrypt 1.6, "gpg
--version" printed two question marks. This patches fixes that and
also make sure that gpg does advertise any ECC features. The patch in
build_list is not really needed.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/misc.c (print_pubkey_algo_note): Map the algo.
(openpgp_pk_test_algo, openpgp_pk_test_algo2): Ditto.
(pubkey_get_npkey, pubkey_get_nskey, pubkey_get_nsig)
(pubkey_get_nenc): Return 0 for ECC algorithms.
--
Libgcrypt 1.6 features algorithm 18 (generic ECC). Because of the
missing mapping and no real support for the OpenPGP ECC format, this
led to parsing errors of ECC packets. We better better explicitly
tell gpg that we ECC is not supported.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/options.h (IMPORT_NO_SECKEY): New.
* g10/keyserver.c (keyserver_spawn, keyserver_import_cert): Set new
flag.
* g10/import.c (import_secret_one): Deny import if flag is set.
--
By modifying a keyserver or a DNS record to send a secret key, an
attacker could trick a user into signing using a different key and
user id. The trust model should protect against such rogue keys but
we better make sure that secret keys are never received from remote
sources.
Suggested-by: Stefan Tomanek
Signed-off-by: Werner Koch <wk@gnupg.org>
* include/cipher.h (PUBKEY_USAGE_NONE): New.
* g10/getkey.c (parse_key_usage): Set new flag.
--
We do not want to use the default capabilities (derived from the
algorithm) if any key flags are given in a signature. Thus if key
flags are used in any way, the default key capabilities are never
used.
This allows to create a key with key flags set to all zero so it can't
be used. This better reflects common sense.
(cherry picked from commit 4bde12206c5bf199dc6e12a74af8da4558ba41bf)
* common/iobuf.c (MAX_NESTING_FILTER): New.
(iobuf_push_filter2): Limit the nesting level.
* g10/mainproc.c (mainproc_context): New field ANY. Change HAVE_DATA
and ANY_SIG_SIGN to bit fields of ANY. Add bit field
UNCOMPRESS_FAILED.
(proc_compressed): Avoid printing multiple Bad Data messages.
(check_nesting): Return GPG_ERR_BAD_DATA instead of UNEXPECTED_DATA.
--
This is a more general fix for the nested compression packet bug. In
particular this helps g10/import.c:read_block to stop pushing
compression filters onto an iobuf stream. This patch also reduces the
number of error messages for the non-import case.
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/gpg.c (main): Special case setup_trustdb for --encrypt.
--
Signed-off-by: Werner Koch <wk@gnupg.org>
(cherry picked from commit 498b9a95dc65c43240835d64cc92d8fb43014d53)
