summaryrefslogtreecommitdiff
path: root/cipher/pubkey-internal.h
AgeCommit message (Collapse)AuthorFilesLines
2014-03-04Add a simple (raw) PKCS#1 padding modeDmitry Eremin-Solenikov1-0/+4
* src/cipher.h (PUBKEY_ENC_PKCS1_RAW): New. * cipher/pubkey-util.c (_gcry_pk_util_parse_flaglist): Handle pkcs1-raw flag. * cipher/pubkey-util.c (_gcry_pk_util_data_to_mpi): Handle s-exp like (data (flags pkcs1-raw) (value xxxxx)) * cipher/rsa-common.c (_gcry_rsa_pkcs1_encode_raw_for_sig): PKCS#1-encode data with embedded hash OID for signature verification. * tests/basic.c (check_pubkey_sign): Add tests for s-exps with pkcs1-raw flag. -- Allow user to specify (flags pkcs1-raw) to enable pkcs1 padding of raw value (no hash algorithm is specified). It is up to the user to verify that the passed value is properly formatted and includes DER-encoded ASN OID of the used hash function. Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
2014-01-13Truncate hash values for ECDSA signature schemeDmitry Eremin-Solenikov1-0/+3
* cipher/dsa-common (_gcry_dsa_normalize_hash): New. Truncate opaque mpis as required for DSA and ECDSA signature schemas. * cipher/dsa.c (verify): Return gpg_err_code_t value from verify() to behave like the rest of internal sign/verify functions. * cipher/dsa.c (sign, verify, dsa_verify): Factor out hash truncation. * cipher/ecc-ecdsa.c (_gcry_ecc_ecdsa_sign): Factor out hash truncation. * cipher/ecc-ecdsa.c (_gcry_ecc_ecdsa_verify): as required by ECDSA scheme, truncate hash values to bitlength of used curve. * tests/pubkey.c (check_ecc_sample_key): add a testcase for hash truncation. Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
2013-10-16sexp: Add function gcry_sexp_extract_param.Werner Koch1-3/+0
* src/gcrypt.h.in (_GCRY_GCC_ATTR_SENTINEL): New. (gcry_sexp_extract_param): New. * src/visibility.c (gcry_sexp_extract_param): New. * src/visibility.h (gcry_sexp_extract_param): Add hack to detect internal use. * cipher/pubkey-util.c (_gcry_pk_util_extract_mpis): Move and split into ... * src/sexp.c (_gcry_sexp_vextract_param) (_gcry_sexp_extract_param): this. Change all callers. Add support for buffer descriptors and a path option/ * tests/tsexp.c (die, hex2buffer, hex2mpi, hex2mpiopa): New. (cmp_mpihex, cmp_bufhex): New. (check_extract_param): New. Signed-off-by: Werner Koch <wk@gnupg.org>
2013-10-14pubkey: Support flags list in gcry_pk_genkey.Werner Koch1-0/+3
* src/cipher.h (PUBKEY_FLAG_TRANSIENT_KEY): New. (PUBKEY_FLAG_USE_X931): New. (PUBKEY_FLAG_USE_FIPS186): New. (PUBKEY_FLAG_USE_FIPS186_2): New. * cipher/pubkey-util.c (_gcry_pk_util_parse_flaglist): Rename from parse_flags_list. Parse new flags. * cipher/dsa.c (dsa_generate): Support flag list. * cipher/ecc.c (ecc_generate): Ditto. * cipher/rsa.c (rsa_generate): Ditto. Signed-off-by: Werner Koch <wk@gnupg.org>
2013-10-11pubkey: Move sexp parsing for gcry_pk_decrypt to the modules.Werner Koch1-0/+4
* cipher/rsa.c (rsa_decrypt): Revamp. * cipher/elgamal.c (elg_decrypt): Revamp. * cipher/ecc.c (ecc_decrypt_raw): Revamp. * cipher/pubkey.c (gcry_pk_decrypt): Simplify. (sexp_to_enc): Remove. * cipher/pubkey-util.c (_gcry_pk_util_preparse_encval): New. -- Note that we do not have a regression test for ecc_decrypt_raw. Even GnuPG does not use it. we also better check whether the interface is really usable; for example GnuPG implements way to much low-level ECC code. Maybe we should move the OpenPGP ECC encryption code into Libgcrypt. Signed-off-by: Werner Koch <wk@gnupg.org>
2013-10-09pubkey: Move sexp parsing for gcry_pk_verify to the modules.Werner Koch1-0/+14
* cipher/rsa.c (rsa_verify): Revamp. * cipher/dsa.c (dsa_verify): Revamp. * cipher/elgamal.c (elg_verify): Revamp. * cipher/ecc.c (ecc_verify): Revamp. * cipher/pubkey.c (sexp_to_sig): Remove. (pss_verify_cmp): Move to pubkey-util.c (sexp_data_to_mpi): Ditto. (init_encoding_ctx): Ditto. (gcry_pk_verify): Simplify. * cipher/pubkey-util.c (_gcry_pk_util_init_encoding_ctx): Add. Take from pubkey.c (get_hash_algo): Ditto. (_gcry_pk_util_data_to_mpi): Ditto. (pss_verify_cmp): Ditto. (_gcry_pk_util_extract_mpis): New. (_gcry_pk_util_preparse_sigval): New. (_gcry_pk_util_free_encoding_ctx): New. * cipher/ecc-curves.c (_gcry_ecc_fill_in_curve): Make curve init optional. * src/g10lib.h (GCC_ATTR_SENTINEL): New. * tests/basic.c (check_pubkey_sign): Print the algo name. (main): Add option --pubkey. Signed-off-by: Werner Koch <wk@gnupg.org>
2013-10-08pubkey: Move sexp parsing for gcry_pk_getkey to the modules.Werner Koch1-2/+10
* cipher/pubkey-util.c: New. (_gcry_pk_util_get_nbits): New. Based on code from gcry_pk_genkey. (_gcry_pk_util_get_rsa_use_e): Ditto. * cipher/pubkey.c (gcry_pk_genkey): Strip most code and pass. * cipher/rsa.c (rsa_generate): Remove args ALGO, NBITS and EVALUE. Call new fucntions to get these values. * cipher/dsa.c (dsa_generate): Remove args ALGO, NBITS and EVALUE. Call _gcry_pk_util_get_nbits to get nbits. Always parse genparms. * cipher/elgamal.c (elg_generate): Ditto. * cipher/ecc.c (ecc_generate): Ditto. Signed-off-by: Werner Koch <wk@gnupg.org>
2013-09-19pk: Move RSA encoding functions to a new file.Werner Koch1-0/+33
* cipher/rsa-common: New. * cipher/pubkey.c (pkcs1_encode_for_encryption): Move to rsa-common.c and rename to _gcry_rsa_pkcs1_encode_for_enc. (pkcs1_decode_for_encryption): Move to rsa-common.c and rename to _gcry_rsa_pkcs1_decode_for_enc. (pkcs1_encode_for_signature): Move to rsa-common.c and rename to _gcry_rsa_pkcs1_encode_for_sig. (oaep_encode): Move to rsa-common.c and rename to _gcry_rsa_oaep_encode. (oaep_decode): Move to rsa-common.c and rename to _gcry_rsa_oaep_decode. (pss_encode): Move to rsa-common.c and rename to _gcry_rsa_pss_encode. (pss_verify): Move to rsa-common.c and rename to _gcry_rsa_pss_decode. (octet_string_from_mpi, mgf1): Move to rsa-common.c. Signed-off-by: Werner Koch <wk@gnupg.org>
2013-07-26Implement deterministic DSA as specified by rfc-6979.Werner Koch1-0/+6
* cipher/dsa.c (dsa_sign): Move opaque mpi extraction to sign. (sign): Add args FLAGS and HASHALGO. Implement deterministic DSA. Add code path for R==0 to comply with the standard. (dsa_verify): Left fill opaque mpi based hash values. * cipher/dsa-common.c (int2octets, bits2octets): New. (_gcry_dsa_gen_rfc6979_k): New. * tests/dsa-rfc6979.c: New. * tests/Makefile.am (TESTS): Add dsa-rfc6979. -- This patch also fixes a recent patch (37d0a1e) which allows to pass the hash in a (hash) element. Support for deterministic ECDSA will come soon. Signed-off-by: Werner Koch <wk@gnupg.org>
2013-05-24ecc: Fix a minor flaw in the generation of K.Werner Koch1-0/+3
* cipher/dsa.c (gen_k): Factor code out to .. * cipher/dsa-common.c (_gcry_dsa_gen_k): new file and function. Add arg security_level and re-indent a bit. * cipher/ecc.c (gen_k): Remove and change callers to _gcry_dsa_gen_k. * cipher/dsa.c: Include pubkey-internal. * cipher/Makefile.am (libcipher_la_SOURCES): Add dsa-common.c -- The ECDSA code used the simple $k = k \bmod p$ method which introduces a small bias. We now use the bias free method we have always used with DSA. Signed-off-by: Werner Koch <wk@gnupg.org>
2013-04-11Add gcry_pubkey_get_sexp.Werner Koch1-0/+29
* src/gcrypt.h.in (GCRY_PK_GET_PUBKEY): New. (GCRY_PK_GET_SECKEY): New. (gcry_pubkey_get_sexp): New. * src/visibility.c (gcry_pubkey_get_sexp): New. * src/visibility.h (gcry_pubkey_get_sexp): Mark visible. * src/libgcrypt.def, src/libgcrypt.vers: Add new function. * cipher/pubkey-internal.h: New. * cipher/Makefile.am (libcipher_la_SOURCES): Add new file. * cipher/ecc.c: Include pubkey-internal.h (_gcry_pk_ecc_get_sexp): New. * cipher/pubkey.c: Include pubkey-internal.h and context.h. (_gcry_pubkey_get_sexp): New. * src/context.c (_gcry_ctx_find_pointer): New. * src/cipher-proto.h: Add _gcry_pubkey_get_sexp. * tests/t-mpi-point.c (print_sexp): New. (context_param, basic_ec_math_simplified): Add tests for the new function. * configure.ac (NEED_GPG_ERROR_VERSION): Set to 1.11. (AH_BOTTOM) Add error codes from gpg-error 1.12 * src/g10lib.h (fips_not_operational): Use GPG_ERR_NOT_OPERATIONAL. * mpi/ec.c (_gcry_mpi_ec_get_mpi): Fix computation of Q. (_gcry_mpi_ec_get_point): Ditto. -- While checking the new code I figured that the auto-computation of Q must have led to a segv. It seems we had no test case for that. Signed-off-by: Werner Koch <wk@gnupg.org>