1
0
mirror of git://git.gnupg.org/gnupg.git synced 2024-12-23 10:29:58 +01:00

1176 Commits

Author SHA1 Message Date
Neal H. Walfield
cab581c486 g10: Improve documentation and comments for getkey.c.
* g10/getkey.c: Improve documentation and comments for most
functions.  Move documentation for public functions from here...
* g10/keydb.h: ... to here.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:28 +02:00
Neal H. Walfield
7333e704ef g10: Remove unused function have_any_secret_key.
* g10/getkey.c (have_any_secret_key): Remove function.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:27 +02:00
Neal H. Walfield
5e233e12f5 g10: Bring cache semantics closer to non-cache semantics.
* g10/getkey.c (get_pubkey_fast): When reading from the cache, only
consider primary keys.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:26 +02:00
Neal H. Walfield
3940f10af7 g10: Break out of the loop earlier.
* g10/getkey.c (have_secret_key_with_kid): Once we find the relevent
key or subkey, stop searching.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.

Only a single key or subkey will ever be selected per keyblock.
2015-09-16 15:15:26 +02:00
Neal H. Walfield
50affffe93 g10: Don't skip legacy keys if the search mode is KEYDB_SEARCH_MODE_NEXT
* g10/getkey.c (lookup): Also don't skip legacy keys if the search
mode is KEYDB_SEARCH_MODE_NEXT.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.

We currently don't skip keys if the search mode is
KEYDB_SEARCH_MODE_FIRST.  Since we change KEYDB_SEARCH_MODE_FIRST to
KEYDB_SEARCH_MODE_NEXT (to avoid a reset), it only makes sense to have
the same semantics for KEYDB_SEARCH_MODE_NEXT.
2015-09-16 15:15:25 +02:00
Neal H. Walfield
efbaa8f891 g10: Remove unused function get_seckeyblock_byfprint.
* g10/keydb.h (get_seckeyblock_byfprint): Remove prototype.
* g10/getkey.c (get_seckeyblock_byfprint): Remove function.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:25 +02:00
Neal H. Walfield
e2b300801e g10: Remove unused function get_seckey_byfprint.
* g10/keydb.h (get_seckey_byfprint): Remove prototype.
* g10/getkey.c (get_seckey_byfprint): Remove function.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:24 +02:00
Neal H. Walfield
80dbf8006f g10: Simplify get_seckey_byname: it was never called with NAME not NULL.
* g10/keydb.h (get_seckey_byname): Rename from this...
(get_seckey_default): ... to this.  Drop the parameter name.  Update
users.
* g10/getkey.c (get_seckey_byname): Rename from this...
(get_seckey_default): ... to this.  Drop the parameter name.  Drop the
code which assumed that NAME is not NULL.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:23 +02:00
Neal H. Walfield
dc69804ab0 g10: Eliminate the redundant function get_keyblock_byfprint.
* g10/keydb.h (get_keyblock_byfprint): Remove prototype.  Replace use
of this function with get_pubkey_byfprint.
* g10/getkey.c (get_pubkey_byname): Remove function.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:23 +02:00
Neal H. Walfield
911fcca36d g10: Simplify semantics of get_pubkey_byname.
* g10/getkey.c (get_pubkey_byname): If R_KEYBLOCK is not NULL, return
the keyblock in R_KEYBLOCK independent of whether PK is set or not.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.

Currently, no caller invokes get_pubkey_byname with PK==NULL and
R_KEYBLOCK != NULL.  Thus, this change does not change any behavior.
2015-09-16 15:15:22 +02:00
Neal H. Walfield
b4672e4d48 g10: Eliminate the redundant function get_pubkey_byname.
* g10/getkey.c (get_pubkey_byname): Remove function.
(lookup): Replace use of get_pubkey_byname by get_pubkey_byfprint.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:22 +02:00
Neal H. Walfield
65e58ae674 g10: Eliminate the redundant function get_pubkey_end.
* g10/keydb.h (get_pubkey_end): Remove declaration.  Replace use of
function with getkey_end.
* g10/getkey.c (get_pubkey_byname): Remove function.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:21 +02:00
Neal H. Walfield
be6743b2e1 g10: Eliminate the redundant function get_pubkey_next.
* g10/keydb.h (get_pubkey_next): Remove prototype.
* g10/getkey.c (get_pubkey_next): Remove function.
* g10/keylist.c (locate_one): Use getkey_next instead of
get_pubkey_next.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:20 +02:00
Neal H. Walfield
9acbeac236 kbx: Change skipfnc's prototype so that we can provide all information.
* kbx/keybox-search-desc.h (struct keydb_search_desc.skipfnc): Change
third parameter to be the index of the user id packet in the keyblock
rather than the packet itself.  Update users.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.

The keybox code doesn't work directly with keyblocks.  As such, the
matched user packet is not readily available to pass to
DESC[n].SKIPFNC.  But, we do know the index of the user id packet that
matched.  Thus, pass that instead.  If the skip function needs the
user id packet, it can use the key id to look up the key block and
find the appropriate packet.
2015-09-16 15:15:20 +02:00
Neal H. Walfield
83e17ab1b4 g10: Remove unused prototype (get_pubkey_byfpr).
* g10/keydb.h (get_pubkey_byfpr): Remove unused prototype.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:19 +02:00
Neal H. Walfield
b06f96ba4f g10: Remove unused function (get_pubkey_bynames).
* g10/keydb.h (get_pubkey_bynames): Remove prototype.
* g10/getkey.c (get_pubkey_bynames): Remove function.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:19 +02:00
Neal H. Walfield
d47e84946e g10: Simplify code. Turn struct getkey_ctx_s.found_key into an argument
* g10/getkey.c (struct getkey_ctx_s): Remove field found_key.
(lookup): Add argument ret_found_key.  If not NULL, set it to the
found key.  Update callers.
(pk_from_block): Add argument found_key.  Use it instead of
CTX->FOUND_KEY.  Update callers.
(finish_lookup): Return a KBNODE (the found key) instead of an int.
Don't set CTX->FOUND_KEY.  Return the found key instead.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:18 +02:00
Neal H. Walfield
c110e186e0 g10: Remove unused field struct getkey_ctx_s.kbpos.
* g10/getkey.c (struct getkey_ctx_s): Remove field kbpos.
(getkey_end): Don't clear CTX->KBPOS.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:17 +02:00
Neal H. Walfield
3798f73c07 g10: Simplify code: remove field struct getkey_ctx_s.keyblock.
* g10/getkey.c (struct getkey_ctx_s): Remove field keyblock.
(finish_lookup): Add parameter keyblock.  Update caller to pass this.
(lookup): Add new local variable keyblock.  Use this instead of
ctx->keyblock for referencing the keyblock.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-16 15:15:17 +02:00
Werner Koch
e92a8ab021
g10: Fix make distcheck problem.
* g10/test.c: Include string.h.
(prepend_srcdir): New.  Taken from Libgcrypt.
(test_free): New.
* g10/t-keydb.c (do_test): Malloc the filename.
* g10/Makefile.am (AM_CPPFLAGS): Remove -DSOURCE_DIR
(EXTRA_DIST): Add t-keydb-keyring.kbx.
--

Using SOURCE_DIR should in general work but we have seen problems when
doing this in Libgcrypt.  Using the srcdir variable gives us anyway
more flexibility and aligns with the way we do it in tests/openpgp.

Signed-off-by: Werner Koch <wk@gnupg.org>
2015-09-10 18:12:20 +02:00
Werner Koch
fbf24cd09a
g10: Improve portability of the new test driver.
* g10/test.c: Include stdio.h and stdlib.h.
(verbose): New.
(print_results): Rename to exit_tests.
(main): Remove atexit and call exit_tests.  Set verbose.
(ASSERT, ABORT): Call exit_tests instead of exit.
--

Calling exit from an exit handler is undefined behaviour.  It works on
Linux but other systems will hit an endless loop.  That is indeed
unfortunate but we can't do anything about it.  Calling _exit() would
be possible but that may lead to other problems.  Thus we change to
call a custom exit function :-(.

Using "make check verbose=1" is supported by tests/openpgp and thus
we add the same mechanism here.

Signed-off-by: Werner Koch <wk@gnupg.org>
2015-09-10 17:43:37 +02:00
Werner Koch
56bf1061f8
doc: Minor comment fixes.
--
2015-09-09 15:14:35 +02:00
Neal H. Walfield
bd0c902f1d g10: Remove unused field req_algo.
* g10/packet.h (PKT_public_key): Remove unused field req_algo.  Remove
users.
* g10/getkey.c (struct getkey_ctx_s): Remove unused field req_algo.
Remove users.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-02 15:08:58 +02:00
Neal H. Walfield
1f03d4cd94 g10: Use a symbolic constant instead of a literal.
* g10/trustdb.c (KEY_HASH_TABLE_SIZE): Define.
(new_key_hash_table): Use KEY_HASH_TABLE_SIZE instead of a literal.
(release_key_hash_table): Likewise.
(test_key_hash_table): Likewise.
(add_key_hash_table): Likewise.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-02 15:08:57 +02:00
Neal H. Walfield
ee7ec1256b g10: Add test for keydb as well as new testing infrastructure.
* g10/Makefile.am (EXTRA_DIST): Add test.c.
(AM_CPPFLAGS): Add -DSOURCE_DIR="\"$(srcdir)\"".
(module_tests): Add t-keydb.
(t_keydb_SOURCES): New variable.
(t_keydb_LDADD): Likewise.
* g10/t-keydb.c: New file.
* g10/t-keydb-keyring.kbx: New file.
* g10/test-stubs.c: New file.
* g10/test.c: New file.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-02 15:08:57 +02:00
Neal H. Walfield
60bc518645 g10: Make the keyblock cache per-handle rather than global.
* g10/keydb.c (keyblock_cache): Don't declare this variable.  Instead...
(struct keyblock_cache): ... turn its type into this first class
object...
(struct keydb_handle): ... and instantiate it once per database
handle.  Update all users.
(keydb_rebuild_caches): Don't invalidate the keyblock cache.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-02 10:52:01 +02:00
Neal H. Walfield
f076fa190e g10: If iobuf_seek fails when reading from the cache, do a hard read.
* g10/keydb.c (keydb_get_keyblock): If the iobuf_seek fails when
reading from the cache, then simply clear the cache and try reading
from the database.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-09-02 10:38:42 +02:00
Werner Koch
f9c83d84e7
gpg: Remove option --no-sig-create-check.
* g10/gpg.c (opts): Remove --no-sig-create-check.
* g10/options.h (struct opt): Remove field no_sig_create_check.
* g10/sign.c (do_sign): Always check unless it is RSA and we are using
Libgcrypt 1.7.

Signed-off-by: Werner Koch <wk@gnupg.org>
2015-09-01 07:37:12 +02:00
Neal H. Walfield
04a6b903d0 g10: Don't leak memory if we fail to initialize a new database handle.
* g10/keydb.c (keydb_new): If we fail to open a keyring or keybox
correctly release all resources.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-31 11:30:57 +02:00
Neal H. Walfield
360b699e9b g10: Improve interface documentation of the keydb API.
* g10/keydb.c: Improve code comments and documentation of internal
interfaces.  Improve documentation of public APIs and move that to...
* g10/keydb.h: ... this file.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-31 11:30:56 +02:00
Neal H. Walfield
efd1ead9e7 g10: Don't cache search results if the search didn't scan the whole DB.
* g10/keydb.c (struct keydb_handle): Add new field is_reset.
(keydb_new): Initialize hd->is_reset to 1.
(keydb_locate_writable): Set hd->is_reset to 1.
(keydb_search): Set hd->is_reset to 0.  Don't cache a key not found if
the search started from the beginning of the database.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-31 11:30:55 +02:00
Neal H. Walfield
11d8ffc939 g10: Have keydb_search_first call keydb_search_reset before searching.
* g10/keydb.c (keydb_search_first): Reset the handle before starting
the search.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.

This bug hasn't shown up yet in practice, because keydb_search_first
is always called immediately after a keydb_new.  This changes cleans
up the semantics and will hopefully prevent future bugs.
2015-08-31 11:30:55 +02:00
Neal H. Walfield
0377db4b35 g10: Remove unused parameter.
* g10/keydb.h (keydb_locate_writable): Remove unused parameter
reserved.  Update users.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-31 11:30:54 +02:00
Neal H. Walfield
fad91071ca g10: Simplify cache. Only include data that is actually used.
* g10/keydb.c (struct kid_list_s): Rename from this...
(struct kid_not_found_cache_bucket): ... to this.  Update users.
Remove field state.
(kid_list_t): Remove type.
(KID_NOT_FOUND_CACHE_BUCKETS): Define.  Use this instead of a literal.
(kid_found_table): Rename from this...
(kid_not_found_cache_bucket): ... to this.  Update users.
(kid_found_table_count): Rename from this...
(kid_not_found_cache_count): ... to this.  Update users.
(kid_not_found_p): Only return whether a key with the specified key id
is definitely not in the database.
(kid_not_found_insert): Remove parameter found.  Update callers.
(keydb_search): Only insert a key id in the not found cache if it is
not found.  Rename local variable once_found to already_in_cache.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.

Commit e0873a33 started tracking whether key ids where definitely in
the database.  This information is, however, never used and thus just
unnecessarily inflates the cache.  This patch effectively reverts that
change (however, e0873a33 contains two separate changes and this only
reverts that change).
2015-08-26 12:25:53 +02:00
Werner Koch
51b9b8fba4
gpg: Emit ERROR status for key signing failures.
* g10/keyedit.c (sign_uids): Write an ERROR status for a signing
failure.
(menu_adduid, menu_addrevoker, menu_revsig): Ditto.
(menu_revuid, menu_revkey, menu_revsubkey): Ditto.
--

This change helps GPA to show better error messages.

Signed-off-by: Werner Koch <wk@gnupg.org>
2015-08-25 15:26:50 +02:00
Werner Koch
9cdff09743
gpg: Print a new FAILURE status after most commands.
* common/status.h (STATUS_FAILURE): New.
* g10/cpr.c (write_status_failure): New.
* g10/gpg.c (main): Call write_status_failure for all commands which
print an error message here.
* g10/call-agent.c (start_agent): Print an STATUS_ERROR if we can't
set the pinentry mode.
--

This status line can be used similar to the error code returned by
commands send over the Assuan interface in gpgsm.  We don't emit them
in gpgsm because there we already have that Assuan interface to return
proper error code.  This change helps GPGME to return better error
codes.

Signed-off-by: Werner Koch <wk@gnupg.org>
2015-08-25 15:26:33 +02:00
Werner Koch
88317ae8c8
gpg: Fix regression in packet parser from Aug 19.
* g10/parse-packet.c (parse): Use an int to compare to -1.  Use
buf32_to_ulong.
--

Regression-due-to: 0add91ae1ca3718e8140af09294c595f47c958d3
Signed-off-by: Werner Koch <wk@gnupg.org>
2015-08-23 12:17:43 +02:00
Werner Koch
00def10d36
gpg: Show not found keys with --locate-key --verbose.
* g10/keylist.c (locate_one): Print a diagnostic for a not-found key.

Signed-off-by: Werner Koch <wk@gnupg.org>
2015-08-23 11:56:17 +02:00
Neal H. Walfield
09f2a7bca6 common: Don't incorrectly reject 4 GB - 1 sized packets.
* g10/parse-packet.c (parse): Don't reject 4 GB - 1 sized packets.
Add the constraint that the type must be 63.
* kbx/keybox-openpgp.c (next_packet): Likewise.
* tests/openpgp/4gb-packet.asc: New file.
* tests/openpgp/4gb-packet.test: New file.
* tests/openpgp/Makefile.am (TESTS): Add 4gb-packet.test.
(TEST_FILES): Add 4gb-packet.asc.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-21 14:21:13 +02:00
Neal H. Walfield
4f37820334 common: Don't assume on-disk layout matches in-memory layout.
* g10/packet.h (PKT_signature): Change revkey's type from a struct
revocation_key ** to a struct revocation_key *.  Update users.

--
revkey was a pointer into the raw data.  But, C doesn't guarantee that
there is no padding.  Thus, we copy the data.

Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-21 14:21:08 +02:00
Neal H. Walfield
b3226cadf9 common: Don't incorrectly copy packets with partial lengths.
* g10/parse-packet.c (parse): We don't handle copying packets with a
partial body length to an output stream.  If this occurs, log an error
and abort.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-21 14:21:04 +02:00
Neal H. Walfield
0143d5c1ca common: Check parameters more rigorously.
* g10/parse-packet.c (dbg_copy_all_packets): Check that OUT is not
NULL.
(copy_all_packets): Likewise.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-21 14:21:00 +02:00
Neal H. Walfield
48e792cc95 common: Don't continuing processing on error.
* g10/parse-packet.c (dbg_parse_packet): Also return if parse returns
an error.
(parse_packet): Likewise.
(dbg_search_packet): Likewise.
(search_packet): Likewise.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-21 14:20:54 +02:00
Neal H. Walfield
73af66a0aa common: Better respect the packet's length when reading it.
* g10/parse-packet.c (parse_signature): Make sure PKTLEN doesn't
underflow.  Be more careful that a read doesn't read more data than
PKTLEN says is available.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-21 14:20:43 +02:00
Neal H. Walfield
0add91ae1c g10/parse-packet.c:parse: Try harder to not ignore an EOF.
* g10/parse-packet.c (parse): Be more robust: make sure to process any
EOF.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-20 14:16:31 +02:00
Neal H. Walfield
24a72dffa7 g10/parse-packet.c: Replace literal with symbolic expression.
* g10/parse-packet.c (dump_hex_line): Use sizeof rather than the
buffer's size.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-20 14:16:30 +02:00
Neal H. Walfield
026feff4a8 Add documentation for g10/parse-packet.c.
* g10/packet.h: Add documentation for functions defined in
parse-packet.c.
* g10/parse-packet.c: Improve comments for many functions.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-20 14:16:29 +02:00
Neal H. Walfield
c46e8bfe9a g10/packet.h: Remove unused argument from enum_sig_subpkt.
* g10/packet.h (enum_sig_subpkt): Remove argument RET_N.  Update
callers.
* g10/parse-packet.c (enum_sig_subpkt): Remove argument RET_N.

--
Remove the RET_N argument, because it is unused and because it is
meaningless: it's not clear whether it is an offset into SIG->HASHED
or SIG->UNHASHED.

Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-20 14:16:29 +02:00
Neal H. Walfield
c271feb536 g10/parse-packet.c:mpi_read: Detect EOF and correct boundary conditions.
* g10/parse-packet.c (mpi_read): Improve documentation.  Correctly
handle an EOF.  On overflow, correctly return the number of bytes read
from the pipeline.

--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-20 14:16:28 +02:00
Neal H. Walfield
8402815d8e common/iobuf.h: Remove iobuf_open_fd_or_name.
* common/iobuf.h (iobuf_open_fd_or_name): Remove prototype.  Replace
use with either iobuf_open or iobuf_fdopen_nc, as appropriate.
* common/iobuf.c (iobuf_open): Remove function.
--
Signed-off-by: Neal H. Walfield <neal@g10code.com>.
2015-08-20 14:16:22 +02:00