* g10/keyring.c (keyring_get_keyblock): Whitelist allowed packet
types.
--
The keyring DB code did not reject packets which don't belong into a
keyring. If for example the keyblock contains a literal data packet
it is expected that the processing code stops at the data packet and
reads from the input stream which is referenced from the data packets.
Obviously the keyring processing code does not and cannot do that.
However, when exporting this messes up the IOBUF and leads to an
invalid read of sizeof (int).
We now skip all packets which are not allowed in a keyring.
Reported-by: Hanno Böck <hanno@hboeck.de>
(back ported from commit f0f71a721c)
* g10/build-packet.c (write_fake_data): Take care of a NULL stored as
opaque MPI.
--
Reported-by: Hanno Böck <hanno@hboeck.de>
(back ported from commit 0835d2f44e)
* kbx/keybox-update.c (blob_filecopy): Fix resource leak. On error
return, 'fp' and 'newfp' was never closed.
--
Signed-off-by: Joshua Rogers <git@internot.info>
[Log entry reformatted, and added more fixes - gniibe]
(cherry picked from commit 7db6c82cec)
* g10/parse-packet.c (MAX_KEY_PACKET_LENGTH): New.
(MAX_UID_PACKET_LENGTH): New.
(MAX_COMMENT_PACKET_LENGTH): New.
(MAX_ATTR_PACKET_LENGTH): New.
(parse_key): Limit the size of a key packet to 256k.
(parse_user_id): Use macro for the packet size limit.
(parse_attribute): Ditto.
(parse_comment): Ditto.
--
Without that it is possible to force gpg to allocate large amounts of
memory by using a bad encoded MPI. This would be an too easy DoS.
Another way to mitigate would be to change the MPI read function to
allocate memory dynamically while reading the MPI. However, that
complicates and possibly slows down the code. A too large key packet
is in any case a sign for broken data and thus gpg should not use it.
Reported-by: Hanno Böck
GnuPG-bug-id: 1823
Signed-off-by: Werner Koch <wk@gnupg.org>
(back ported from commit 382ba4b137)
* jnlib/dotlock.c (create_dotlock): Avoid double close due to EINTR.
--
close(2) says:
close() should not be retried after an EINTR since this may cause a
reused descriptor from another thread to be closed.
(backported from commit 628b111fa6)
Debian-Bug-Id: 773423
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/keyedit.c (subkey_expire_warning): New.
keyedit_menu): Call it when needed.
--
GnuPG-bug-id: 1715
The heuristic to detect a problem is not very advanced but it should
catch the most common cases.
(backported from commit ae3d1bbb65)
* scd/ccid-driver.c (send_escape_cmd): Fix setting of 'rc'.
--
Variable 'rc' in send_escape_cmd was overwritten before it was
returned, leading to incorrect computation.
Signed-off-by: Joshua Rogers <git@internot.info>
[Log entry reformatted - wk]
* tools/gpgconf-comp.c (option_check_validity): Enable check for
UINT32.
--
Reported-by: Günther Noack <gnoack@google.com>
This is actually a bug which inhibited the checking of values of type
UINT32.
Signed-off-by: Werner Koch <wk@gnupg.org>
(cherry picked from commit 3f6abb57a7)
* sm/minip12.c: (p12_build) if jnlib_iconv_open fails, avoid
double-free of pwbuf.
--
Observed by Joshua Rogers <honey@internot.info>, who proposed a
slightly different fix.
Debian-Bug-Id: 773472
Added fix at a second place - wk.
* scd/command.c (cmd_readkey): avoid double-free of cert
--
When ksba_cert_new() fails, cert will be double-freed.
Debian-Bug-Id: 773471
Original patch changed by wk to do the free only at leave.
* 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
* tools/gpgkey2ssh.c (key_to_blob) : ensure that va_end is called.
--
stdarg(3) says:
Each invocation of va_start() must be matched by a
corresponding invocation of va_end() in the same function.
Observed by Joshua Rogers <honey@internot.info>
Debian-Bug-Id: 773415
* sm/gpgsm.c (parse_keyserver_line): Set SERVER to NULL.
--
Cherry-pick of abd5f6752d.
Reported-by: Joshua Rogers <git@internot.info>
"If something inside the ldapserver_parse_one function failed,
'server' would be freed, then returned, leading to a
use-after-free. This code is likely copied from sm/gpgsm.c, which
was also susceptible to this bug."
Signed-off-by: Werner Koch <wk@gnupg.org>
* g10/parse-packet.c (can_handle_critical): Check content length
before calling can_handle_critical_notation.
--
The problem was found by Jan Bee and gniibe proposed the used fix.
Thanks.
This bug can't be exploited: Only if the announced length of the
notation is 21 or 32 a memcmp against fixed strings using that length
would be done. The compared data is followed by the actual signature
and thus it is highly likely that not even read of unallocated memory
will happen. Nevertheless such a bug needs to be fixed.
Signed-off-by: Werner Koch <wk@gnupg.org>
* scd/app-openpgp.c (get_public_key): correctly close 'fp' upon use.
--
Inside the get_public_key function, 'fp' was opened using popen, but
incorrectly closed using fclose.
Debian-Bug-Id: 773474
* g10/keygen.c (generate_subkeypair): Release DEK soon.
--
This fixes the out_of_core error in the test case of adding
RSA-4096 subkey to RSA-4096 primary key with configuration:
s2k-cipher-algo S10
Debian-bug-id: 772780
Cherry-picked da66ad5bba from
STABLE-BRANCH-1-4 branch.
* g10/parse-packet.c (dump_sig_subpkt): Print regex subpacket
sanitized.
--
We may not use "%s" to print an arbitrary buffer. At least "%.*s"
should have been used. However, it is in general preferable to escape
control characters while printf user data.
Reported-by: Hanno Böck
Signed-off-by: Werner Koch <wk@gnupg.org>
(backported from commit 596ae9f543)
* g10/parse-packet.c (parse_attribute_subpkts): Check that the
attribute packet is large enough for the subpacket type.
--
Reported-by: Hanno Böck
Signed-off-by: Werner Koch <wk@gnupg.org>
(backported from commit 0988764397)
* g10/mainproc.c (proc_encrypted): Take care of canceled passpharse
entry.
--
GnuPG-bug-id: 1761
Signed-off-by: Werner Koch <wk@gnupg.org>
(backported from commit 32e85668b8)
* g10/openfile.c (open_sigfile): Factor some code out to ...
(get_matching_datafile): new function.
* g10/plaintext.c (hash_datafiles): Do not try to find matching file
in batch mode.
* g10/mainproc.c (check_sig_and_print): Print a warning if a possibly
matching data file is not used by a standard signatures.
--
Allowing to use the abbreviated form for detached signatures is a long
standing bug which has only been noticed by the public with the
release of 2.1.0. :-(
What we do is to remove the ability to check detached signature in
--batch using the one file abbreviated mode. This should exhibit
problems in scripts which use this insecure practice. We also print a
warning if a matching data file exists but was not considered because
the detached signature was actually a standard signature:
gpgv: Good signature from "Werner Koch (dist sig)"
gpgv: WARNING: not a detached signature; \
file 'gnupg-2.1.0.tar.bz2' was NOT verified!
We can only print a warning because it is possible that a standard
signature is indeed to be verified but by coincidence a file with a
matching name is stored alongside the standard signature.
Reported-by: Simon Nicolussi (to gnupg-users on Nov 7)
Signed-off-by: Werner Koch <wk@gnupg.org>
(backported from commit 69384568f6)
* g10/options.h (IMPORT_KEEP_OWNERTTRUST): New.
* g10/import.c (parse_import_options): Add "keep-ownertrust".
(import_one): Act upon new option.
--
This option is in particular useful to convert from a pubring.gpg to
the new pubring.kbx in GnuPG 2.1 or vice versa:
gpg1 --export | gpg2 --import-options keep-ownertrust --import
(cherry-picked from commit ffc2307843)
* g10/keyid.c (fingerprint_from_pk): Show v3 fingerprints as all zero.
--
MD5 is considered broken for a long time now. To make it easier for
users to notice that a listing shows a v3 key, the fingerprint is now
displayed as 16 zero bytes unless --allow-weak-digest-algos is active.
Signed-off-by: Werner Koch <wk@gnupg.org>
* 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 534e2876ac 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.
--
c76117f8b0 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 c76117f8b0, 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 d332467005)
