boringssl/ssl/internal.h

1552 lines
65 KiB
C
Raw Normal View History

/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
* All rights reserved.
*
* This package is an SSL implementation written
* by Eric Young (eay@cryptsoft.com).
* The implementation was written so as to conform with Netscapes SSL.
*
* This library is free for commercial and non-commercial use as long as
* the following conditions are aheared to. The following conditions
* apply to all code found in this distribution, be it the RC4, RSA,
* lhash, DES, etc., code; not just the SSL code. The SSL documentation
* included with this distribution is covered by the same copyright terms
* except that the holder is Tim Hudson (tjh@cryptsoft.com).
*
* Copyright remains Eric Young's, and as such any Copyright notices in
* the code are not to be removed.
* If this package is used in a product, Eric Young should be given attribution
* as the author of the parts of the library used.
* This can be in the form of a textual message at program startup or
* in documentation (online or textual) provided with the package.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* "This product includes cryptographic software written by
* Eric Young (eay@cryptsoft.com)"
* The word 'cryptographic' can be left out if the rouines from the library
* being used are not cryptographic related :-).
* 4. If you include any Windows specific code (or a derivative thereof) from
* the apps directory (application code) you must include an acknowledgement:
* "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
*
* THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* The licence and distribution terms for any publically available version or
* derivative of this code cannot be changed. i.e. this code cannot simply be
* copied and put under another distribution licence
* [including the GNU Public Licence.]
*/
/* ====================================================================
* Copyright (c) 1998-2007 The OpenSSL Project. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
*
* 3. All advertising materials mentioning features or use of this
* software must display the following acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
*
* 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
* endorse or promote products derived from this software without
* prior written permission. For written permission, please contact
* openssl-core@openssl.org.
*
* 5. Products derived from this software may not be called "OpenSSL"
* nor may "OpenSSL" appear in their names without prior written
* permission of the OpenSSL Project.
*
* 6. Redistributions of any form whatsoever must retain the following
* acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit (http://www.openssl.org/)"
*
* THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
* EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE OpenSSL PROJECT OR
* ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
* STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
* OF THE POSSIBILITY OF SUCH DAMAGE.
* ====================================================================
*
* This product includes cryptographic software written by Eric Young
* (eay@cryptsoft.com). This product includes software written by Tim
* Hudson (tjh@cryptsoft.com).
*
*/
/* ====================================================================
* Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
* ECC cipher suite support in OpenSSL originally developed by
* SUN MICROSYSTEMS, INC., and contributed to the OpenSSL project.
*/
/* ====================================================================
* Copyright 2005 Nokia. All rights reserved.
*
* The portions of the attached software ("Contribution") is developed by
* Nokia Corporation and is licensed pursuant to the OpenSSL open source
* license.
*
* The Contribution, originally written by Mika Kousa and Pasi Eronen of
* Nokia Corporation, consists of the "PSK" (Pre-Shared Key) ciphersuites
* support (see RFC 4279) to OpenSSL.
*
* No patent licenses or other rights except those expressly stated in
* the OpenSSL open source license shall be deemed granted or received
* expressly, by implication, estoppel, or otherwise.
*
* No assurances are provided by Nokia that the Contribution does not
* infringe the patent or other intellectual property rights of any third
* party or that the license provides you with all the necessary rights
* to make use of the Contribution.
*
* THE SOFTWARE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND. IN
* ADDITION TO THE DISCLAIMERS INCLUDED IN THE LICENSE, NOKIA
* SPECIFICALLY DISCLAIMS ANY LIABILITY FOR CLAIMS BROUGHT BY YOU OR ANY
* OTHER ENTITY BASED ON INFRINGEMENT OF INTELLECTUAL PROPERTY RIGHTS OR
* OTHERWISE.
*/
#ifndef OPENSSL_HEADER_SSL_INTERNAL_H
#define OPENSSL_HEADER_SSL_INTERNAL_H
#include <openssl/base.h>
#include <openssl/aead.h>
#include <openssl/ssl.h>
#include <openssl/stack.h>
#if defined(OPENSSL_WINDOWS)
/* Windows defines struct timeval in winsock2.h. */
OPENSSL_MSVC_PRAGMA(warning(push, 3))
#include <winsock2.h>
OPENSSL_MSVC_PRAGMA(warning(pop))
#else
#include <sys/time.h>
#endif
#if defined(__cplusplus)
extern "C" {
#endif
/* Cipher suites. */
/* Bits for |algorithm_mkey| (key exchange algorithm). */
#define SSL_kRSA 0x00000001L
#define SSL_kDHE 0x00000002L
#define SSL_kECDHE 0x00000004L
/* SSL_kPSK is only set for plain PSK, not ECDHE_PSK. */
#define SSL_kPSK 0x00000008L
#define SSL_kCECPQ1 0x00000010L
/* Bits for |algorithm_auth| (server authentication). */
#define SSL_aRSA 0x00000001L
#define SSL_aECDSA 0x00000002L
/* SSL_aPSK is set for both PSK and ECDHE_PSK. */
#define SSL_aPSK 0x00000004L
#define SSL_aCERT (SSL_aRSA | SSL_aECDSA)
/* Bits for |algorithm_enc| (symmetric encryption). */
#define SSL_3DES 0x00000001L
#define SSL_AES128 0x00000002L
#define SSL_AES256 0x00000004L
#define SSL_AES128GCM 0x00000008L
#define SSL_AES256GCM 0x00000010L
#define SSL_CHACHA20POLY1305_OLD 0x00000020L
#define SSL_eNULL 0x00000040L
#define SSL_CHACHA20POLY1305 0x00000080L
#define SSL_AES (SSL_AES128 | SSL_AES256 | SSL_AES128GCM | SSL_AES256GCM)
/* Bits for |algorithm_mac| (symmetric authentication). */
#define SSL_MD5 0x00000001L
#define SSL_SHA1 0x00000002L
#define SSL_SHA256 0x00000004L
#define SSL_SHA384 0x00000008L
/* SSL_AEAD is set for all AEADs. */
#define SSL_AEAD 0x00000010L
/* Bits for |algorithm_prf| (handshake digest). */
#define SSL_HANDSHAKE_MAC_DEFAULT 0x1
#define SSL_HANDSHAKE_MAC_SHA256 0x2
#define SSL_HANDSHAKE_MAC_SHA384 0x4
/* SSL_MAX_DIGEST is the number of digest types which exist. When adding a new
* one, update the table in ssl_cipher.c. */
#define SSL_MAX_DIGEST 4
/* ssl_cipher_get_evp_aead sets |*out_aead| to point to the correct EVP_AEAD
* object for |cipher| protocol version |version|. It sets |*out_mac_secret_len|
* and |*out_fixed_iv_len| to the MAC key length and fixed IV length,
* respectively. The MAC key length is zero except for legacy block and stream
* ciphers. It returns 1 on success and 0 on error. */
int ssl_cipher_get_evp_aead(const EVP_AEAD **out_aead,
size_t *out_mac_secret_len,
size_t *out_fixed_iv_len,
const SSL_CIPHER *cipher, uint16_t version);
/* ssl_get_handshake_digest returns the |EVP_MD| corresponding to
* |algorithm_prf|. It returns SHA-1 for |SSL_HANDSHAKE_DEFAULT|. The caller is
* responsible for maintaining the additional MD5 digest and switching to
* SHA-256 in TLS 1.2. */
const EVP_MD *ssl_get_handshake_digest(uint32_t algorithm_prf);
/* ssl_create_cipher_list evaluates |rule_str| according to the ciphers in
* |ssl_method|. It sets |*out_cipher_list| to a newly-allocated
* |ssl_cipher_preference_list_st| containing the result.
* |*out_cipher_list_by_id| is set to a list of selected ciphers sorted by
* id. It returns |(*out_cipher_list)->ciphers| on success and NULL on
* failure. */
STACK_OF(SSL_CIPHER) *
ssl_create_cipher_list(const SSL_PROTOCOL_METHOD *ssl_method,
struct ssl_cipher_preference_list_st **out_cipher_list,
STACK_OF(SSL_CIPHER) **out_cipher_list_by_id,
const char *rule_str);
/* ssl_cipher_get_value returns the cipher suite id of |cipher|. */
uint16_t ssl_cipher_get_value(const SSL_CIPHER *cipher);
/* ssl_cipher_get_resumption_cipher returns the cipher suite id of the cipher
* matching |cipher| with PSK enabled. */
int ssl_cipher_get_ecdhe_psk_cipher(const SSL_CIPHER *cipher,
uint16_t *out_cipher);
/* ssl_cipher_get_key_type returns the |EVP_PKEY_*| value corresponding to the
* server key used in |cipher| or |EVP_PKEY_NONE| if there is none. */
int ssl_cipher_get_key_type(const SSL_CIPHER *cipher);
/* ssl_cipher_uses_certificate_auth returns one if |cipher| authenticates the
* server and, optionally, the client with a certificate. Otherwise it returns
* zero. */
int ssl_cipher_uses_certificate_auth(const SSL_CIPHER *cipher);
/* ssl_cipher_requires_server_key_exchange returns 1 if |cipher| requires a
* ServerKeyExchange message. Otherwise it returns 0.
*
* This function may return zero while still allowing |cipher| an optional
* ServerKeyExchange. This is the case for plain PSK ciphers. */
int ssl_cipher_requires_server_key_exchange(const SSL_CIPHER *cipher);
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
/* ssl_cipher_get_record_split_len, for TLS 1.0 CBC mode ciphers, returns the
* length of an encrypted 1-byte record, for use in record-splitting. Otherwise
* it returns zero. */
size_t ssl_cipher_get_record_split_len(const SSL_CIPHER *cipher);
/* Encryption layer. */
/* SSL_AEAD_CTX contains information about an AEAD that is being used to encrypt
* an SSL connection. */
struct ssl_aead_ctx_st {
const SSL_CIPHER *cipher;
EVP_AEAD_CTX ctx;
/* fixed_nonce contains any bytes of the nonce that are fixed for all
* records. */
uint8_t fixed_nonce[12];
uint8_t fixed_nonce_len, variable_nonce_len;
/* variable_nonce_included_in_record is non-zero if the variable nonce
* for a record is included as a prefix before the ciphertext. */
char variable_nonce_included_in_record;
/* random_variable_nonce is non-zero if the variable nonce is
* randomly generated, rather than derived from the sequence
* number. */
char random_variable_nonce;
/* omit_length_in_ad is non-zero if the length should be omitted in the
* AEAD's ad parameter. */
char omit_length_in_ad;
/* omit_version_in_ad is non-zero if the version should be omitted
* in the AEAD's ad parameter. */
char omit_version_in_ad;
/* omit_ad is non-zero if the AEAD's ad parameter should be omitted. */
char omit_ad;
/* xor_fixed_nonce is non-zero if the fixed nonce should be XOR'd into the
* variable nonce rather than prepended. */
char xor_fixed_nonce;
} /* SSL_AEAD_CTX */;
/* SSL_AEAD_CTX_new creates a newly-allocated |SSL_AEAD_CTX| using the supplied
* key material. It returns NULL on error. Only one of |SSL_AEAD_CTX_open| or
* |SSL_AEAD_CTX_seal| may be used with the resulting object, depending on
* |direction|. |version| is the normalized protocol version, so DTLS 1.0 is
* represented as 0x0301, not 0xffef. */
SSL_AEAD_CTX *SSL_AEAD_CTX_new(enum evp_aead_direction_t direction,
uint16_t version, const SSL_CIPHER *cipher,
const uint8_t *enc_key, size_t enc_key_len,
const uint8_t *mac_key, size_t mac_key_len,
const uint8_t *fixed_iv, size_t fixed_iv_len);
/* SSL_AEAD_CTX_free frees |ctx|. */
void SSL_AEAD_CTX_free(SSL_AEAD_CTX *ctx);
/* SSL_AEAD_CTX_explicit_nonce_len returns the length of the explicit nonce for
* |ctx|, if any. |ctx| may be NULL to denote the null cipher. */
size_t SSL_AEAD_CTX_explicit_nonce_len(SSL_AEAD_CTX *ctx);
/* SSL_AEAD_CTX_max_overhead returns the maximum overhead of calling
* |SSL_AEAD_CTX_seal|. |ctx| may be NULL to denote the null cipher. */
size_t SSL_AEAD_CTX_max_overhead(SSL_AEAD_CTX *ctx);
/* SSL_AEAD_CTX_open authenticates and decrypts |in_len| bytes from |in|
* in-place. On success, it sets |*out| to the plaintext in |in| and returns
* one. Otherwise, it returns zero. |ctx| may be NULL to denote the null cipher.
* The output will always be |explicit_nonce_len| bytes ahead of |in|. */
int SSL_AEAD_CTX_open(SSL_AEAD_CTX *ctx, CBS *out, uint8_t type,
uint16_t wire_version, const uint8_t seqnum[8],
uint8_t *in, size_t in_len);
/* SSL_AEAD_CTX_seal encrypts and authenticates |in_len| bytes from |in| and
* writes the result to |out|. It returns one on success and zero on
* error. |ctx| may be NULL to denote the null cipher.
*
* If |in| and |out| alias then |out| + |explicit_nonce_len| must be == |in|. */
int SSL_AEAD_CTX_seal(SSL_AEAD_CTX *ctx, uint8_t *out, size_t *out_len,
size_t max_out, uint8_t type, uint16_t wire_version,
const uint8_t seqnum[8], const uint8_t *in,
size_t in_len);
/* DTLS replay bitmap. */
/* DTLS1_BITMAP maintains a sliding window of 64 sequence numbers to detect
* replayed packets. It should be initialized by zeroing every field. */
typedef struct dtls1_bitmap_st {
/* map is a bit mask of the last 64 sequence numbers. Bit
* |1<<i| corresponds to |max_seq_num - i|. */
uint64_t map;
/* max_seq_num is the largest sequence number seen so far as a 64-bit
* integer. */
uint64_t max_seq_num;
} DTLS1_BITMAP;
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
/* Record layer. */
/* ssl_record_sequence_update increments the sequence number in |seq|. It
* returns one on success and zero on wraparound. */
int ssl_record_sequence_update(uint8_t *seq, size_t seq_len);
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
/* ssl_record_prefix_len returns the length of the prefix before the ciphertext
* of a record for |ssl|.
*
* TODO(davidben): Expose this as part of public API once the high-level
* buffer-free APIs are available. */
size_t ssl_record_prefix_len(const SSL *ssl);
enum ssl_open_record_t {
ssl_open_record_success,
ssl_open_record_discard,
ssl_open_record_partial,
ssl_open_record_close_notify,
ssl_open_record_fatal_alert,
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
ssl_open_record_error,
};
/* tls_open_record decrypts a record from |in| in-place.
*
* If the input did not contain a complete record, it returns
* |ssl_open_record_partial|. It sets |*out_consumed| to the total number of
* bytes necessary. It is guaranteed that a successful call to |tls_open_record|
* will consume at least that many bytes.
*
* Otherwise, it sets |*out_consumed| to the number of bytes of input
* consumed. Note that input may be consumed on all return codes if a record was
* decrypted.
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
*
* On success, it returns |ssl_open_record_success|. It sets |*out_type| to the
* record type and |*out| to the record body in |in|. Note that |*out| may be
* empty.
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
*
* If a record was successfully processed but should be discarded, it returns
* |ssl_open_record_discard|.
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
*
* If a record was successfully processed but is a close_notify or fatal alert,
* it returns |ssl_open_record_close_notify| or |ssl_open_record_fatal_alert|.
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
*
* On failure, it returns |ssl_open_record_error| and sets |*out_alert| to an
* alert to emit. */
enum ssl_open_record_t tls_open_record(SSL *ssl, uint8_t *out_type, CBS *out,
size_t *out_consumed, uint8_t *out_alert,
uint8_t *in, size_t in_len);
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
/* dtls_open_record implements |tls_open_record| for DTLS. It never returns
* |ssl_open_record_partial| but otherwise behaves analogously. */
enum ssl_open_record_t dtls_open_record(SSL *ssl, uint8_t *out_type, CBS *out,
size_t *out_consumed,
uint8_t *out_alert, uint8_t *in,
size_t in_len);
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
/* ssl_seal_align_prefix_len returns the length of the prefix before the start
* of the bulk of the ciphertext when sealing a record with |ssl|. Callers may
* use this to align buffers.
*
* Note when TLS 1.0 CBC record-splitting is enabled, this includes the one byte
* record and is the offset into second record's ciphertext. Thus this value may
* differ from |ssl_record_prefix_len| and sealing a small record may result in
* a smaller output than this value.
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
*
* TODO(davidben): Expose this as part of public API once the high-level
* buffer-free APIs are available. */
size_t ssl_seal_align_prefix_len(const SSL *ssl);
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
/* ssl_max_seal_overhead returns the maximum overhead of sealing a record with
* |ssl|.
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
*
* TODO(davidben): Expose this as part of public API once the high-level
* buffer-free APIs are available. */
size_t ssl_max_seal_overhead(const SSL *ssl);
/* tls_seal_record seals a new record of type |type| and body |in| and writes it
* to |out|. At most |max_out| bytes will be written. It returns one on success
* and zero on error. If enabled, |tls_seal_record| implements TLS 1.0 CBC 1/n-1
* record splitting and may write two records concatenated.
*
* For a large record, the bulk of the ciphertext will begin
* |ssl_seal_align_prefix_len| bytes into out. Aligning |out| appropriately may
* improve performance. It writes at most |in_len| + |ssl_max_seal_overhead|
* bytes to |out|.
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
*
* |in| and |out| may not alias. */
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
int tls_seal_record(SSL *ssl, uint8_t *out, size_t *out_len, size_t max_out,
uint8_t type, const uint8_t *in, size_t in_len);
enum dtls1_use_epoch_t {
dtls1_use_previous_epoch,
dtls1_use_current_epoch,
};
/* dtls_seal_record implements |tls_seal_record| for DTLS. |use_epoch| selects
* which epoch's cipher state to use. */
int dtls_seal_record(SSL *ssl, uint8_t *out, size_t *out_len, size_t max_out,
uint8_t type, const uint8_t *in, size_t in_len,
enum dtls1_use_epoch_t use_epoch);
/* ssl_process_alert processes |in| as an alert and updates |ssl|'s shutdown
* state. It returns one of |ssl_open_record_discard|, |ssl_open_record_error|,
* |ssl_open_record_close_notify|, or |ssl_open_record_fatal_alert| as
* appropriate. */
enum ssl_open_record_t ssl_process_alert(SSL *ssl, uint8_t *out_alert,
const uint8_t *in, size_t in_len);
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
/* Private key operations. */
/* ssl_has_private_key returns one if |ssl| has a private key
* configured and zero otherwise. */
int ssl_has_private_key(const SSL *ssl);
/* ssl_is_ecdsa_key_type returns one if |type| is an ECDSA key type and zero
* otherwise. */
int ssl_is_ecdsa_key_type(int type);
/* ssl_private_key_* call the corresponding function on the
* |SSL_PRIVATE_KEY_METHOD| for |ssl|, if configured. Otherwise, they implement
* the operation with |EVP_PKEY|. */
int ssl_private_key_type(SSL *ssl);
size_t ssl_private_key_max_signature_len(SSL *ssl);
enum ssl_private_key_result_t ssl_private_key_sign(
SSL *ssl, uint8_t *out, size_t *out_len, size_t max_out,
uint16_t signature_algorithm, const uint8_t *in, size_t in_len);
enum ssl_private_key_result_t ssl_private_key_decrypt(
SSL *ssl, uint8_t *out, size_t *out_len, size_t max_out,
const uint8_t *in, size_t in_len);
enum ssl_private_key_result_t ssl_private_key_complete(SSL *ssl, uint8_t *out,
size_t *out_len,
size_t max_out);
/* ssl_private_key_supports_signature_algorithm returns one if |ssl|'s private
* key supports |signature_algorithm| and zero otherwise. */
int ssl_private_key_supports_signature_algorithm(SSL *ssl,
uint16_t signature_algorithm);
/* ssl_public_key_verify verifies that the |signature| is valid for the public
* key |pkey| and input |in|, using the |signature_algorithm| specified. */
int ssl_public_key_verify(
SSL *ssl, const uint8_t *signature, size_t signature_len,
uint16_t signature_algorithm, EVP_PKEY *pkey,
const uint8_t *in, size_t in_len);
/* Custom extensions */
/* ssl_custom_extension (a.k.a. SSL_CUSTOM_EXTENSION) is a structure that
* contains information about custom-extension callbacks. */
struct ssl_custom_extension {
SSL_custom_ext_add_cb add_callback;
void *add_arg;
SSL_custom_ext_free_cb free_callback;
SSL_custom_ext_parse_cb parse_callback;
void *parse_arg;
uint16_t value;
};
void SSL_CUSTOM_EXTENSION_free(SSL_CUSTOM_EXTENSION *custom_extension);
int custom_ext_add_clienthello(SSL *ssl, CBB *extensions);
int custom_ext_parse_serverhello(SSL *ssl, int *out_alert, uint16_t value,
const CBS *extension);
int custom_ext_parse_clienthello(SSL *ssl, int *out_alert, uint16_t value,
const CBS *extension);
int custom_ext_add_serverhello(SSL *ssl, CBB *extensions);
/* Handshake hash.
*
* The TLS handshake maintains a transcript of all handshake messages. At
* various points in the protocol, this is either a handshake buffer, a rolling
* hash (selected by cipher suite) or both. */
/* ssl3_init_handshake_buffer initializes the handshake buffer and resets the
* handshake hash. It returns one success and zero on failure. */
int ssl3_init_handshake_buffer(SSL *ssl);
/* ssl3_init_handshake_hash initializes the handshake hash based on the pending
* cipher and the contents of the handshake buffer. Subsequent calls to
* |ssl3_update_handshake_hash| will update the rolling hash. It returns one on
* success and zero on failure. It is an error to call this function after the
* handshake buffer is released. */
int ssl3_init_handshake_hash(SSL *ssl);
/* ssl3_free_handshake_buffer releases the handshake buffer. Subsequent calls
* to |ssl3_update_handshake_hash| will not update the handshake buffer. */
void ssl3_free_handshake_buffer(SSL *ssl);
/* ssl3_free_handshake_hash releases the handshake hash. */
void ssl3_free_handshake_hash(SSL *ssl);
/* ssl3_update_handshake_hash adds |in| to the handshake buffer and handshake
* hash, whichever is enabled. It returns one on success and zero on failure. */
int ssl3_update_handshake_hash(SSL *ssl, const uint8_t *in, size_t in_len);
/* ECDH groups. */
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
/* An SSL_ECDH_METHOD is an implementation of ECDH-like key exchanges for
* TLS. */
struct ssl_ecdh_method_st {
int nid;
uint16_t group_id;
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
const char name[8];
/* cleanup releases state in |ctx|. */
void (*cleanup)(SSL_ECDH_CTX *ctx);
/* offer generates a keypair and writes the public value to
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
* |out_public_key|. It returns one on success and zero on error. */
int (*offer)(SSL_ECDH_CTX *ctx, CBB *out_public_key);
/* accept performs a key exchange against the |peer_key| generated by |offer|.
* On success, it returns one, writes the public value to |out_public_key|,
* and sets |*out_secret| and |*out_secret_len| to a newly-allocated buffer
* containing the shared secret. The caller must release this buffer with
* |OPENSSL_free|. On failure, it returns zero and sets |*out_alert| to an
* alert to send to the peer. */
int (*accept)(SSL_ECDH_CTX *ctx, CBB *out_public_key, uint8_t **out_secret,
size_t *out_secret_len, uint8_t *out_alert,
const uint8_t *peer_key, size_t peer_key_len);
/* finish performs a key exchange against the |peer_key| generated by
* |accept|. On success, it returns one and sets |*out_secret| and
* |*out_secret_len| to a newly-allocated buffer containing the shared
* secret. The caller must release this buffer with |OPENSSL_free|. On
* failure, it returns zero and sets |*out_alert| to an alert to send to the
* peer. */
int (*finish)(SSL_ECDH_CTX *ctx, uint8_t **out_secret, size_t *out_secret_len,
uint8_t *out_alert, const uint8_t *peer_key,
size_t peer_key_len);
/* get_key initializes |out| with a length-prefixed key from |cbs|. It returns
* one on success and zero on error. */
int (*get_key)(CBS *cbs, CBS *out);
/* add_key initializes |out_contents| to receive a key. Typically it will then
* be passed to |offer| or |accept|. It returns one on success and zero on
* error. */
int (*add_key)(CBB *cbb, CBB *out_contents);
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
} /* SSL_ECDH_METHOD */;
/* ssl_nid_to_group_id looks up the group corresponding to |nid|. On success, it
* sets |*out_group_id| to the group ID and returns one. Otherwise, it returns
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
* zero. */
int ssl_nid_to_group_id(uint16_t *out_group_id, int nid);
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
/* SSL_ECDH_CTX_init sets up |ctx| for use with curve |group_id|. It returns one
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
* on success and zero on error. */
int SSL_ECDH_CTX_init(SSL_ECDH_CTX *ctx, uint16_t group_id);
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
/* SSL_ECDH_CTX_init_for_dhe sets up |ctx| for use with legacy DHE-based ciphers
* where the server specifies a group. It takes ownership of |params|. */
void SSL_ECDH_CTX_init_for_dhe(SSL_ECDH_CTX *ctx, DH *params);
/* SSL_ECDH_CTX_init_for_cecpq1 sets up |ctx| for use with CECPQ1. */
void SSL_ECDH_CTX_init_for_cecpq1(SSL_ECDH_CTX *ctx);
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
/* SSL_ECDH_CTX_cleanup releases memory associated with |ctx|. It is legal to
* call it in the zero state. */
void SSL_ECDH_CTX_cleanup(SSL_ECDH_CTX *ctx);
/* SSL_ECDH_CTX_get_id returns the group ID for |ctx|. */
uint16_t SSL_ECDH_CTX_get_id(const SSL_ECDH_CTX *ctx);
/* SSL_ECDH_CTX_get_key calls the |get_key| method of |SSL_ECDH_METHOD|. */
int SSL_ECDH_CTX_get_key(SSL_ECDH_CTX *ctx, CBS *cbs, CBS *out);
/* SSL_ECDH_CTX_add_key calls the |add_key| method of |SSL_ECDH_METHOD|. */
int SSL_ECDH_CTX_add_key(SSL_ECDH_CTX *ctx, CBB *cbb, CBB *out_contents);
/* SSL_ECDH_CTX_offer calls the |offer| method of |SSL_ECDH_METHOD|. */
int SSL_ECDH_CTX_offer(SSL_ECDH_CTX *ctx, CBB *out_public_key);
/* SSL_ECDH_CTX_accept calls the |accept| method of |SSL_ECDH_METHOD|. */
int SSL_ECDH_CTX_accept(SSL_ECDH_CTX *ctx, CBB *out_public_key,
uint8_t **out_secret, size_t *out_secret_len,
uint8_t *out_alert, const uint8_t *peer_key,
size_t peer_key_len);
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
/* SSL_ECDH_CTX_finish the |finish| method of |SSL_ECDH_METHOD|. */
int SSL_ECDH_CTX_finish(SSL_ECDH_CTX *ctx, uint8_t **out_secret,
size_t *out_secret_len, uint8_t *out_alert,
const uint8_t *peer_key, size_t peer_key_len);
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
Simplify handshake message size limits. A handshake message can go up to 2^24 bytes = 16MB which is a little large for the peer to force us to buffer. Accordingly, we bound the size of a handshake message. Rather than have a global limit, the existing logic uses a different limit at each state in the handshake state machine and, for certificates, allows configuring the maximum certificate size. This is nice in that we engage larger limits iff the relevant state is reachable from the handshake. Servers without client auth get a tighter limit "for free". However, this doesn't work for DTLS due to out-of-order messages and we use a simpler scheme for DTLS. This scheme also is tricky on optional messages and makes the handshake <-> message layer communication complex. Apart from an ignored 20,000 byte limit on ServerHello, the largest non-certificate limit is the common 16k limit on ClientHello. So this complexity wasn't buying us anything. Unify everything on the DTLS scheme except, so as not to regress bounds on client-auth-less servers, also correctly check for whether client auth is configured. The value of 16k was chosen based on this value. (The 20,000 byte ServerHello limit makes no sense. We can easily bound the ServerHello because servers may not send extensions we don't implement. But it gets overshadowed by the certificate anyway.) Change-Id: I00309b16d809a3c2a1543f99fd29c4163e3add81 Reviewed-on: https://boringssl-review.googlesource.com/7941 Reviewed-by: David Benjamin <davidben@google.com>
2016-05-12 05:43:05 +01:00
/* Handshake messages. */
/* SSL_MAX_HANDSHAKE_FLIGHT is the number of messages, including
* ChangeCipherSpec, in the longest handshake flight. Currently this is the
* client's second leg in a full handshake when client certificates, NPN, and
* Channel ID, are all enabled. */
#define SSL_MAX_HANDSHAKE_FLIGHT 7
Simplify handshake message size limits. A handshake message can go up to 2^24 bytes = 16MB which is a little large for the peer to force us to buffer. Accordingly, we bound the size of a handshake message. Rather than have a global limit, the existing logic uses a different limit at each state in the handshake state machine and, for certificates, allows configuring the maximum certificate size. This is nice in that we engage larger limits iff the relevant state is reachable from the handshake. Servers without client auth get a tighter limit "for free". However, this doesn't work for DTLS due to out-of-order messages and we use a simpler scheme for DTLS. This scheme also is tricky on optional messages and makes the handshake <-> message layer communication complex. Apart from an ignored 20,000 byte limit on ServerHello, the largest non-certificate limit is the common 16k limit on ClientHello. So this complexity wasn't buying us anything. Unify everything on the DTLS scheme except, so as not to regress bounds on client-auth-less servers, also correctly check for whether client auth is configured. The value of 16k was chosen based on this value. (The 20,000 byte ServerHello limit makes no sense. We can easily bound the ServerHello because servers may not send extensions we don't implement. But it gets overshadowed by the certificate anyway.) Change-Id: I00309b16d809a3c2a1543f99fd29c4163e3add81 Reviewed-on: https://boringssl-review.googlesource.com/7941 Reviewed-by: David Benjamin <davidben@google.com>
2016-05-12 05:43:05 +01:00
/* ssl_max_handshake_message_len returns the maximum number of bytes permitted
* in a handshake message for |ssl|. */
size_t ssl_max_handshake_message_len(const SSL *ssl);
/* dtls_clear_incoming_messages releases all buffered incoming messages. */
void dtls_clear_incoming_messages(SSL *ssl);
/* dtls_has_incoming_messages returns one if there are buffered incoming
* messages ahead of the current message and zero otherwise. */
int dtls_has_incoming_messages(const SSL *ssl);
typedef struct dtls_outgoing_message_st {
uint8_t *data;
uint32_t len;
uint16_t epoch;
char is_ccs;
} DTLS_OUTGOING_MESSAGE;
/* dtls_clear_outgoing_messages releases all buffered outgoing messages. */
void dtls_clear_outgoing_messages(SSL *ssl);
Simplify handshake message size limits. A handshake message can go up to 2^24 bytes = 16MB which is a little large for the peer to force us to buffer. Accordingly, we bound the size of a handshake message. Rather than have a global limit, the existing logic uses a different limit at each state in the handshake state machine and, for certificates, allows configuring the maximum certificate size. This is nice in that we engage larger limits iff the relevant state is reachable from the handshake. Servers without client auth get a tighter limit "for free". However, this doesn't work for DTLS due to out-of-order messages and we use a simpler scheme for DTLS. This scheme also is tricky on optional messages and makes the handshake <-> message layer communication complex. Apart from an ignored 20,000 byte limit on ServerHello, the largest non-certificate limit is the common 16k limit on ClientHello. So this complexity wasn't buying us anything. Unify everything on the DTLS scheme except, so as not to regress bounds on client-auth-less servers, also correctly check for whether client auth is configured. The value of 16k was chosen based on this value. (The 20,000 byte ServerHello limit makes no sense. We can easily bound the ServerHello because servers may not send extensions we don't implement. But it gets overshadowed by the certificate anyway.) Change-Id: I00309b16d809a3c2a1543f99fd29c4163e3add81 Reviewed-on: https://boringssl-review.googlesource.com/7941 Reviewed-by: David Benjamin <davidben@google.com>
2016-05-12 05:43:05 +01:00
/* Callbacks. */
/* ssl_do_info_callback calls |ssl|'s info callback, if set. */
void ssl_do_info_callback(const SSL *ssl, int type, int value);
/* ssl_do_msg_callback calls |ssl|'s message callback, if set. */
void ssl_do_msg_callback(SSL *ssl, int is_write, int content_type,
const void *buf, size_t len);
Factor out the buffering and low-level record code. This begins decoupling the transport from the SSL state machine. The buffering logic is hidden behind an opaque API. Fields like ssl->packet and ssl->packet_length are gone. ssl3_get_record and dtls1_get_record now call low-level tls_open_record and dtls_open_record functions that unpack a single record independent of who owns the buffer. Both may be called in-place. This removes ssl->rstate which was redundant with the buffer length. Future work will push the buffer up the stack until it is above the handshake. Then we can expose SSL_open and SSL_seal APIs which act like *_open_record but return a slightly larger enum due to other events being possible. Likewise the handshake state machine will be detached from its buffer. The existing SSL_read, SSL_write, etc., APIs will be implemented on top of SSL_open, etc., combined with ssl_read_buffer_* and ssl_write_buffer_*. (Which is why ssl_read_buffer_extend still tries to abstract between TLS's and DTLS's fairly different needs.) The new buffering logic does not support read-ahead (removed previously) since it lacks a memmove on ssl_read_buffer_discard for TLS, but this could be added if desired. The old buffering logic wasn't quite right anyway; it tried to avoid the memmove in some cases and could get stuck too far into the buffer and not accept records. (The only time the memmove is optional is in DTLS or if enough of the record header is available to know that the entire next record would fit in the buffer.) The new logic also now actually decrypts the ciphertext in-place again, rather than almost in-place when there's an explicit nonce/IV. (That accidentally switched in https://boringssl-review.googlesource.com/#/c/4792/; see 3d59e04bce96474099ba76786a2337e99ae14505.) BUG=468889 Change-Id: I403c1626253c46897f47c7ae93aeab1064b767b2 Reviewed-on: https://boringssl-review.googlesource.com/5715 Reviewed-by: Adam Langley <agl@google.com>
2015-07-29 02:34:45 +01:00
/* Transport buffers. */
/* ssl_read_buffer returns a pointer to contents of the read buffer. */
uint8_t *ssl_read_buffer(SSL *ssl);
/* ssl_read_buffer_len returns the length of the read buffer. */
size_t ssl_read_buffer_len(const SSL *ssl);
/* ssl_read_buffer_extend_to extends the read buffer to the desired length. For
* TLS, it reads to the end of the buffer until the buffer is |len| bytes
* long. For DTLS, it reads a new packet and ignores |len|. It returns one on
* success, zero on EOF, and a negative number on error.
*
* It is an error to call |ssl_read_buffer_extend_to| in DTLS when the buffer is
* non-empty. */
int ssl_read_buffer_extend_to(SSL *ssl, size_t len);
/* ssl_read_buffer_consume consumes |len| bytes from the read buffer. It
* advances the data pointer and decrements the length. The memory consumed will
* remain valid until the next call to |ssl_read_buffer_extend| or it is
* discarded with |ssl_read_buffer_discard|. */
void ssl_read_buffer_consume(SSL *ssl, size_t len);
/* ssl_read_buffer_discard discards the consumed bytes from the read buffer. If
* the buffer is now empty, it releases memory used by it. */
void ssl_read_buffer_discard(SSL *ssl);
/* ssl_read_buffer_clear releases all memory associated with the read buffer and
* zero-initializes it. */
void ssl_read_buffer_clear(SSL *ssl);
/* ssl_write_buffer_is_pending returns one if the write buffer has pending data
* and zero if is empty. */
int ssl_write_buffer_is_pending(const SSL *ssl);
/* ssl_write_buffer_init initializes the write buffer. On success, it sets
* |*out_ptr| to the start of the write buffer with space for up to |max_len|
* bytes. It returns one on success and zero on failure. Call
* |ssl_write_buffer_set_len| to complete initialization. */
int ssl_write_buffer_init(SSL *ssl, uint8_t **out_ptr, size_t max_len);
/* ssl_write_buffer_set_len is called after |ssl_write_buffer_init| to complete
* initialization after |len| bytes are written to the buffer. */
void ssl_write_buffer_set_len(SSL *ssl, size_t len);
/* ssl_write_buffer_flush flushes the write buffer to the transport. It returns
* one on success and <= 0 on error. For DTLS, whether or not the write
* succeeds, the write buffer will be cleared. */
int ssl_write_buffer_flush(SSL *ssl);
/* ssl_write_buffer_clear releases all memory associated with the write buffer
* and zero-initializes it. */
void ssl_write_buffer_clear(SSL *ssl);
/* Certificate functions. */
/* ssl_has_certificate returns one if a certificate and private key are
* configured and zero otherwise. */
int ssl_has_certificate(const SSL *ssl);
/* ssl_parse_cert_chain parses a certificate list from |cbs| in the format used
* by a TLS Certificate message. On success, it returns a newly-allocated
* |X509| list and advances |cbs|. Otherwise, it returns NULL and sets
* |*out_alert| to an alert to send to the peer. If the list is non-empty and
* |out_leaf_sha256| is non-NULL, it writes the SHA-256 hash of the leaf to
* |out_leaf_sha256|. */
STACK_OF(X509) *ssl_parse_cert_chain(SSL *ssl, uint8_t *out_alert,
uint8_t *out_leaf_sha256, CBS *cbs);
/* ssl_add_cert_to_cbb adds |x509| to |cbb|. It returns one on success and zero
* on error. */
int ssl_add_cert_to_cbb(CBB *cbb, X509 *x509);
/* ssl_add_cert_chain adds |ssl|'s certificate chain to |cbb| in the format used
* by a TLS Certificate message. If there is no certificate chain, it emits an
* empty certificate list. It returns one on success and zero on error. */
int ssl_add_cert_chain(SSL *ssl, CBB *cbb);
/* ssl_parse_client_CA_list parses a CA list from |cbs| in the format used by a
* TLS CertificateRequest message. On success, it returns a newly-allocated
* |X509_NAME| list and advances |cbs|. Otherwise, it returns NULL and sets
* |*out_alert| to an alert to send to the peer. */
STACK_OF(X509_NAME) *
ssl_parse_client_CA_list(SSL *ssl, uint8_t *out_alert, CBS *cbs);
/* ssl_add_client_CA_list adds the configured CA list to |cbb| in the format
* used by a TLS CertificateRequest message. It returns one on success and zero
* on error. */
int ssl_add_client_CA_list(SSL *ssl, CBB *cbb);
/* ssl_check_leaf_certificate returns one if |leaf| is a suitable leaf server
* certificate for |ssl|. Otherwise, it returns zero and pushes an error on the
* error queue. */
int ssl_check_leaf_certificate(SSL *ssl, X509 *leaf);
/* ssl_do_client_cert_cb runs the client_cert_cb, if any, and returns one on
* success and zero on error. On error, it sets |*out_should_retry| to one if
* the callback failed and should be retried and zero otherwise. */
int ssl_do_client_cert_cb(SSL *ssl, int *out_should_retry);
/* TLS 1.3 key derivation. */
/* tls13_init_key_schedule initializes the handshake hash and key derivation
* state with the given resumption context. The cipher suite and PRF hash must
* have been selected at this point. It returns one on success and zero on
* error. */
int tls13_init_key_schedule(SSL *ssl, const uint8_t *resumption_ctx,
size_t resumption_ctx_len);
/* tls13_advance_key_schedule incorporates |in| into the key schedule with
* HKDF-Extract. It returns one on success and zero on error. */
int tls13_advance_key_schedule(SSL *ssl, const uint8_t *in, size_t len);
/* tls13_get_context_hashes writes Hash(Handshake Context) +
* Hash(resumption_context) to |out| which much have room for at least 2 *
* |EVP_MAX_MD_SIZE| bytes. On success, it returns one and sets |*out_len| to
* the number of bytes written. Otherwise, it returns zero. */
int tls13_get_context_hashes(SSL *ssl, uint8_t *out, size_t *out_len);
enum tls_record_type_t {
type_early_handshake,
type_early_data,
type_handshake,
type_data,
};
/* tls13_set_traffic_key sets the read or write traffic keys to |traffic_secret|
* for the given traffic phase |type|. It returns one on success and zero on
* error. */
int tls13_set_traffic_key(SSL *ssl, enum tls_record_type_t type,
enum evp_aead_direction_t direction,
const uint8_t *traffic_secret,
size_t traffic_secret_len);
/* tls13_set_handshake_traffic derives the handshake traffic secret and
* switches both read and write traffic to it. It returns one on success and
* zero on error. */
int tls13_set_handshake_traffic(SSL *ssl);
/* tls13_rotate_traffic_key derives the next read or write traffic secret. It
* returns one on success and zero on error. */
int tls13_rotate_traffic_key(SSL *ssl, enum evp_aead_direction_t direction);
/* tls13_derive_traffic_secret_0 derives the initial application data traffic
* secret based on the handshake transcripts and |master_secret|. It returns one
* on success and zero on error. */
int tls13_derive_traffic_secret_0(SSL *ssl);
/* tls13_finalize_keys derives the |exporter_secret| and |resumption_secret|. */
int tls13_finalize_keys(SSL *ssl);
/* tls13_export_keying_material provides and exporter interface to use the
* |exporter_secret|. */
int tls13_export_keying_material(SSL *ssl, uint8_t *out, size_t out_len,
const char *label, size_t label_len,
const uint8_t *context, size_t context_len,
int use_context);
/* tls13_finished_mac calculates the MAC of the handshake transcript to verify
* the integrity of the Finished message, and stores the result in |out| and
* length in |out_len|. |is_server| is 1 if this is for the Server Finished and
* 0 for the Client Finished. */
int tls13_finished_mac(SSL *ssl, uint8_t *out, size_t *out_len, int is_server);
/* tls13_resumption_psk calculates the PSK to use for the resumption of
* |session| and stores the result in |out|. It returns one on success, and
* zero on failure. */
int tls13_resumption_psk(SSL *ssl, uint8_t *out, size_t out_len,
const SSL_SESSION *session);
/* tls13_resumption_context derives the context to be used for the handshake
* transcript on the resumption of |session|. It returns one on success, and
* zero on failure. */
int tls13_resumption_context(SSL *ssl, uint8_t *out, size_t out_len,
const SSL_SESSION *session);
/* Handshake functions. */
enum ssl_hs_wait_t {
ssl_hs_error,
ssl_hs_ok,
ssl_hs_read_message,
ssl_hs_write_message,
ssl_hs_flush,
ssl_hs_flush_and_read_message,
ssl_hs_x509_lookup,
ssl_hs_private_key_operation,
};
struct ssl_handshake_st {
/* wait contains the operation |do_handshake| is currently blocking on or
* |ssl_hs_ok| if none. */
enum ssl_hs_wait_t wait;
/* do_handshake runs the handshake. On completion, it returns |ssl_hs_ok|.
* Otherwise, it returns a value corresponding to what operation is needed to
* progress. */
enum ssl_hs_wait_t (*do_handshake)(SSL *ssl);
int state;
size_t hash_len;
uint8_t resumption_hash[EVP_MAX_MD_SIZE];
uint8_t secret[EVP_MAX_MD_SIZE];
uint8_t traffic_secret_0[EVP_MAX_MD_SIZE];
/* ecdh_ctx is the active client ECDH offer in TLS 1.3. */
SSL_ECDH_CTX ecdh_ctx;
/* retry_group is the group ID selected by the server in HelloRetryRequest in
* TLS 1.3. */
uint16_t retry_group;
/* key_share_bytes is the value of the previously sent KeyShare extension by
* the client in TLS 1.3. */
uint8_t *key_share_bytes;
size_t key_share_bytes_len;
/* public_key, for servers, is the key share to be sent to the client in TLS
* 1.3. */
uint8_t *public_key;
size_t public_key_len;
/* peer_sigalgs are the signature algorithms that the peer supports. These are
* taken from the contents of the signature algorithms extension for a server
* or from the CertificateRequest for a client. */
uint16_t *peer_sigalgs;
/* num_peer_sigalgs is the number of entries in |peer_sigalgs|. */
size_t num_peer_sigalgs;
uint8_t session_tickets_sent;
/* peer_psk_identity_hint, on the client, is the psk_identity_hint sent by the
* server when using a TLS 1.2 PSK key exchange. */
char *peer_psk_identity_hint;
} /* SSL_HANDSHAKE */;
SSL_HANDSHAKE *ssl_handshake_new(enum ssl_hs_wait_t (*do_handshake)(SSL *ssl));
/* ssl_handshake_free releases all memory associated with |hs|. */
void ssl_handshake_free(SSL_HANDSHAKE *hs);
/* tls13_handshake runs the TLS 1.3 handshake. It returns one on success and <=
* 0 on error. */
int tls13_handshake(SSL *ssl);
/* The following are implementations of |do_handshake| for the client and
* server. */
enum ssl_hs_wait_t tls13_client_handshake(SSL *ssl);
enum ssl_hs_wait_t tls13_server_handshake(SSL *ssl);
/* tls13_post_handshake processes a post-handshake message. It returns one on
* success and zero on failure. */
int tls13_post_handshake(SSL *ssl);
/* tls13_check_message_type checks if the current message has type |type|. If so
* it returns one. Otherwise, it sends an alert and returns zero. */
int tls13_check_message_type(SSL *ssl, int type);
int tls13_process_certificate(SSL *ssl, int allow_anonymous);
int tls13_process_certificate_verify(SSL *ssl);
int tls13_process_finished(SSL *ssl);
int tls13_prepare_certificate(SSL *ssl);
enum ssl_private_key_result_t tls13_prepare_certificate_verify(
SSL *ssl, int is_first_run);
int tls13_prepare_finished(SSL *ssl);
int tls13_process_new_session_ticket(SSL *ssl);
int ssl_ext_key_share_parse_serverhello(SSL *ssl, uint8_t **out_secret,
size_t *out_secret_len,
uint8_t *out_alert, CBS *contents);
int ssl_ext_key_share_parse_clienthello(SSL *ssl, int *out_found,
uint8_t **out_secret,
size_t *out_secret_len,
uint8_t *out_alert, CBS *contents);
int ssl_ext_key_share_add_serverhello(SSL *ssl, CBB *out);
int ssl_ext_pre_shared_key_parse_serverhello(SSL *ssl, uint8_t *out_alert,
CBS *contents);
int ssl_ext_pre_shared_key_parse_clienthello(SSL *ssl,
SSL_SESSION **out_session,
uint8_t *out_alert, CBS *contents);
int ssl_ext_pre_shared_key_add_serverhello(SSL *ssl, CBB *out);
int ssl_add_client_hello_body(SSL *ssl, CBB *body);
/* ssl_clear_tls13_state releases client state only needed for TLS 1.3. It
* should be called once the version is known to be TLS 1.2 or earlier. */
void ssl_clear_tls13_state(SSL *ssl);
2016-07-19 06:26:49 +01:00
/* SSLKEYLOGFILE functions. */
/* ssl_log_rsa_client_key_exchange logs |premaster|, if logging is enabled for
* |ssl|. It returns one on success and zero on failure. The entry is identified
* by the first 8 bytes of |encrypted_premaster|. */
int ssl_log_rsa_client_key_exchange(const SSL *ssl,
const uint8_t *encrypted_premaster,
size_t encrypted_premaster_len,
const uint8_t *premaster,
size_t premaster_len);
/* ssl_log_secret logs |secret| with label |label|, if logging is enabled for
* |ssl|. It returns one on success and zero on failure. */
int ssl_log_secret(const SSL *ssl, const char *label, const uint8_t *secret,
size_t secret_len);
/* ClientHello functions. */
int ssl_early_callback_init(SSL *ssl, struct ssl_early_callback_ctx *ctx,
const uint8_t *in, size_t in_len);
int ssl_early_callback_get_extension(const struct ssl_early_callback_ctx *ctx,
CBS *out, uint16_t extension_type);
STACK_OF(SSL_CIPHER) *
ssl_parse_client_cipher_list(const struct ssl_early_callback_ctx *ctx);
int ssl_client_cipher_list_contains_cipher(
const struct ssl_early_callback_ctx *client_hello, uint16_t id);
/* GREASE. */
enum ssl_grease_index_t {
ssl_grease_cipher = 0,
ssl_grease_group,
ssl_grease_extension1,
ssl_grease_extension2,
ssl_grease_version,
};
/* ssl_get_grease_value returns a GREASE value for |ssl|. For a given
* connection, the values for each index will be deterministic. This allows the
* same ClientHello be sent twice for a HelloRetryRequest or the same group be
* advertised in both supported_groups and key_shares. */
uint16_t ssl_get_grease_value(const SSL *ssl, enum ssl_grease_index_t index);
/* Underdocumented functions.
*
* Functions below here haven't been touched up and may be underdocumented. */
#define TLSEXT_CHANNEL_ID_SIZE 128
/* From RFC4492, used in encoding the curve type in ECParameters */
#define NAMED_CURVE_TYPE 3
enum ssl_hash_message_t {
ssl_dont_hash_message,
ssl_hash_message,
};
typedef struct cert_st {
X509 *x509;
EVP_PKEY *privatekey;
/* Chain for this certificate */
STACK_OF(X509) *chain;
/* key_method, if non-NULL, is a set of callbacks to call for private key
* operations. */
const SSL_PRIVATE_KEY_METHOD *key_method;
/* For clients the following masks are of *disabled* key and auth algorithms
* based on the current configuration.
*
* TODO(davidben): Remove these. They get checked twice: when sending the
* ClientHello and when processing the ServerHello. */
uint32_t mask_k;
uint32_t mask_a;
DH *dh_tmp;
DH *(*dh_tmp_cb)(SSL *ssl, int is_export, int keysize);
/* sigalgs, if non-NULL, is the set of signature algorithms supported by
* |privatekey| in decreasing order of preference. */
uint16_t *sigalgs;
size_t num_sigalgs;
/* Certificate setup callback: if set is called whenever a
* certificate may be required (client or server). the callback
* can then examine any appropriate parameters and setup any
* certificates required. This allows advanced applications
* to select certificates on the fly: for example based on
* supported signature algorithms or curves. */
int (*cert_cb)(SSL *ssl, void *arg);
void *cert_cb_arg;
/* Optional X509_STORE for certificate validation. If NULL the parent SSL_CTX
* store is used instead. */
X509_STORE *verify_store;
} CERT;
/* SSL_METHOD is a compatibility structure to support the legacy version-locked
* methods. */
struct ssl_method_st {
/* version, if non-zero, is the only protocol version acceptable to an
* SSL_CTX initialized from this method. */
uint16_t version;
/* method is the underlying SSL_PROTOCOL_METHOD that initializes the
* SSL_CTX. */
const SSL_PROTOCOL_METHOD *method;
};
/* Used to hold functions for SSLv2 or SSLv3/TLSv1 functions */
struct ssl_protocol_method_st {
/* is_dtls is one if the protocol is DTLS and zero otherwise. */
char is_dtls;
/* min_version is the minimum implemented version. */
uint16_t min_version;
/* max_version is the maximum implemented version. */
uint16_t max_version;
/* version_from_wire maps |wire_version| to a protocol version. On success, it
* sets |*out_version| to the result and returns one. If the version is
* unknown, it returns zero. */
int (*version_from_wire)(uint16_t *out_version, uint16_t wire_version);
/* version_to_wire maps |version| to the wire representation. It is an error
* to call it with an invalid version. */
uint16_t (*version_to_wire)(uint16_t version);
int (*ssl_new)(SSL *ssl);
void (*ssl_free)(SSL *ssl);
/* ssl_get_message reads the next handshake message. If |msg_type| is not -1,
* the message must have the specified type. On success, it returns one and
* sets |ssl->s3->tmp.message_type|, |ssl->init_msg|, and |ssl->init_num|.
* Otherwise, it returns <= 0. */
int (*ssl_get_message)(SSL *ssl, int msg_type,
enum ssl_hash_message_t hash_message);
/* hash_current_message incorporates the current handshake message into the
* handshake hash. It returns one on success and zero on allocation
* failure. */
int (*hash_current_message)(SSL *ssl);
/* release_current_message is called to release the current handshake message.
* If |free_buffer| is one, buffers will also be released. */
void (*release_current_message)(SSL *ssl, int free_buffer);
/* read_app_data reads up to |len| bytes of application data into |buf|. On
* success, it returns the number of bytes read. Otherwise, it returns <= 0
* and sets |*out_got_handshake| to whether the failure was due to a
* post-handshake handshake message. If so, it fills in the current message as
* in |ssl_get_message|. */
int (*read_app_data)(SSL *ssl, int *out_got_handshake, uint8_t *buf, int len,
int peek);
int (*read_change_cipher_spec)(SSL *ssl);
void (*read_close_notify)(SSL *ssl);
int (*write_app_data)(SSL *ssl, const void *buf_, int len);
int (*dispatch_alert)(SSL *ssl);
/* supports_cipher returns one if |cipher| is supported by this protocol and
* zero otherwise. */
int (*supports_cipher)(const SSL_CIPHER *cipher);
/* init_message begins a new handshake message of type |type|. |cbb| is the
* root CBB to be passed into |finish_message|. |*body| is set to a child CBB
* the caller should write to. It returns one on success and zero on error. */
int (*init_message)(SSL *ssl, CBB *cbb, CBB *body, uint8_t type);
/* finish_message finishes a handshake message and prepares it to be
* written. It returns one on success and zero on error. */
int (*finish_message)(SSL *ssl, CBB *cbb);
/* write_message writes the next message to the transport. It returns one on
* success and <= 0 on error. */
int (*write_message)(SSL *ssl);
/* send_change_cipher_spec sends a ChangeCipherSpec message. */
int (*send_change_cipher_spec)(SSL *ssl);
/* expect_flight is called when the handshake expects a flight of messages from
* the peer. */
void (*expect_flight)(SSL *ssl);
/* received_flight is called when the handshake has received a flight of
* messages from the peer. */
void (*received_flight)(SSL *ssl);
/* set_read_state sets |ssl|'s read cipher state to |aead_ctx|. It takes
* ownership of |aead_ctx|. It returns one on success and zero if changing the
* read state is forbidden at this point. */
int (*set_read_state)(SSL *ssl, SSL_AEAD_CTX *aead_ctx);
/* set_write_state sets |ssl|'s write cipher state to |aead_ctx|. It takes
* ownership of |aead_ctx|. It returns one on success and zero if changing the
* write state is forbidden at this point. */
int (*set_write_state)(SSL *ssl, SSL_AEAD_CTX *aead_ctx);
};
/* This is for the SSLv3/TLSv1.0 differences in crypto/hash stuff It is a bit
* of a mess of functions, but hell, think of it as an opaque structure. */
struct ssl3_enc_method {
/* prf computes the PRF function for |ssl|. It writes |out_len| bytes to
* |out|, using |secret| as the secret and |label| as the label. |seed1| and
* |seed2| are concatenated to form the seed parameter. It returns one on
* success and zero on failure. */
int (*prf)(const SSL *ssl, uint8_t *out, size_t out_len,
const uint8_t *secret, size_t secret_len, const char *label,
size_t label_len, const uint8_t *seed1, size_t seed1_len,
const uint8_t *seed2, size_t seed2_len);
int (*final_finish_mac)(SSL *ssl, int from_server, uint8_t *out);
};
/* lengths of messages */
#define DTLS1_COOKIE_LENGTH 256
#define DTLS1_RT_HEADER_LENGTH 13
#define DTLS1_HM_HEADER_LENGTH 12
#define DTLS1_CCS_HEADER_LENGTH 1
#define DTLS1_AL_HEADER_LENGTH 2
struct hm_header_st {
uint8_t type;
uint32_t msg_len;
uint16_t seq;
uint32_t frag_off;
uint32_t frag_len;
};
/* An hm_fragment is an incoming DTLS message, possibly not yet assembled. */
typedef struct hm_fragment_st {
/* type is the type of the message. */
uint8_t type;
/* seq is the sequence number of this message. */
uint16_t seq;
/* msg_len is the length of the message body. */
uint32_t msg_len;
/* data is a pointer to the message, including message header. It has length
* |DTLS1_HM_HEADER_LENGTH| + |msg_len|. */
uint8_t *data;
/* reassembly is a bitmask of |msg_len| bits corresponding to which parts of
* the message have been received. It is NULL if the message is complete. */
uint8_t *reassembly;
} hm_fragment;
typedef struct dtls1_state_st {
/* send_cookie is true if we are resending the ClientHello
* with a cookie from a HelloVerifyRequest. */
unsigned int send_cookie;
uint8_t cookie[DTLS1_COOKIE_LENGTH];
size_t cookie_len;
/* The current data and handshake epoch. This is initially undefined, and
* starts at zero once the initial handshake is completed. */
uint16_t r_epoch;
uint16_t w_epoch;
/* records being received in the current epoch */
DTLS1_BITMAP bitmap;
uint16_t handshake_write_seq;
uint16_t handshake_read_seq;
/* save last sequence number for retransmissions */
uint8_t last_write_sequence[8];
/* incoming_messages is a ring buffer of incoming handshake messages that have
* yet to be processed. The front of the ring buffer is message number
* |handshake_read_seq|, at position |handshake_read_seq| %
* |SSL_MAX_HANDSHAKE_FLIGHT|. */
hm_fragment *incoming_messages[SSL_MAX_HANDSHAKE_FLIGHT];
/* outgoing_messages is the queue of outgoing messages from the last handshake
* flight. */
DTLS_OUTGOING_MESSAGE outgoing_messages[SSL_MAX_HANDSHAKE_FLIGHT];
uint8_t outgoing_messages_len;
unsigned int mtu; /* max DTLS packet size */
/* num_timeouts is the number of times the retransmit timer has fired since
* the last time it was reset. */
unsigned int num_timeouts;
/* Indicates when the last handshake msg or heartbeat sent will
* timeout. */
struct timeval next_timeout;
/* timeout_duration_ms is the timeout duration in milliseconds. */
unsigned timeout_duration_ms;
} DTLS1_STATE;
extern const SSL3_ENC_METHOD TLSv1_enc_data;
extern const SSL3_ENC_METHOD SSLv3_enc_data;
/* From draft-ietf-tls-tls13-14, used in determining ticket validity. */
#define SSL_TICKET_ALLOW_EARLY_DATA 1
#define SSL_TICKET_ALLOW_DHE_RESUMPTION 2
#define SSL_TICKET_ALLOW_PSK_RESUMPTION 4
CERT *ssl_cert_new(void);
CERT *ssl_cert_dup(CERT *cert);
void ssl_cert_clear_certs(CERT *c);
void ssl_cert_free(CERT *c);
int ssl_get_new_session(SSL *ssl, int is_server);
int ssl_encrypt_ticket(SSL *ssl, CBB *out, const SSL_SESSION *session);
/* ssl_session_is_context_valid returns one if |session|'s session ID context
* matches the one set on |ssl| and zero otherwise. */
int ssl_session_is_context_valid(const SSL *ssl, const SSL_SESSION *session);
/* ssl_session_is_time_valid returns one if |session| is still valid and zero if
* it has expired. */
int ssl_session_is_time_valid(const SSL *ssl, const SSL_SESSION *session);
void ssl_set_session(SSL *ssl, SSL_SESSION *session);
enum ssl_session_result_t {
ssl_session_success,
ssl_session_error,
ssl_session_retry,
};
/* ssl_get_prev_session looks up the previous session based on |ctx|. On
* success, it sets |*out_session| to the session or NULL if none was found. It
* sets |*out_send_ticket| to whether a ticket should be sent at the end of the
* handshake. If the session could not be looked up synchronously, it returns
* |ssl_session_retry| and should be called again. Otherwise, it returns
* |ssl_session_error|. */
enum ssl_session_result_t ssl_get_prev_session(
SSL *ssl, SSL_SESSION **out_session, int *out_send_ticket,
const struct ssl_early_callback_ctx *ctx);
/* The following flags determine which parts of the session are duplicated. */
#define SSL_SESSION_DUP_AUTH_ONLY 0x0
#define SSL_SESSION_INCLUDE_TICKET 0x1
#define SSL_SESSION_INCLUDE_NONAUTH 0x2
#define SSL_SESSION_DUP_ALL \
(SSL_SESSION_INCLUDE_TICKET | SSL_SESSION_INCLUDE_NONAUTH)
/* SSL_SESSION_dup returns a newly-allocated |SSL_SESSION| with a copy of the
* fields in |session| or NULL on error. The new session is non-resumable and
* must be explicitly marked resumable once it has been filled in. */
OPENSSL_EXPORT SSL_SESSION *SSL_SESSION_dup(SSL_SESSION *session,
int dup_flags);
void ssl_cipher_preference_list_free(
struct ssl_cipher_preference_list_st *cipher_list);
struct ssl_cipher_preference_list_st *ssl_get_cipher_preferences(SSL *ssl);
int ssl_cert_set0_chain(CERT *cert, STACK_OF(X509) *chain);
int ssl_cert_set1_chain(CERT *cert, STACK_OF(X509) *chain);
int ssl_cert_add0_chain_cert(CERT *cert, X509 *x509);
int ssl_cert_add1_chain_cert(CERT *cert, X509 *x509);
void ssl_cert_set_cert_cb(CERT *cert,
int (*cb)(SSL *ssl, void *arg), void *arg);
int ssl_verify_cert_chain(SSL *ssl, long *out_verify_result,
STACK_OF(X509) * cert_chain);
void ssl_update_cache(SSL *ssl, int mode);
/* ssl_get_compatible_server_ciphers determines the key exchange and
* authentication cipher suite masks compatible with the server configuration
* and current ClientHello parameters of |ssl|. It sets |*out_mask_k| to the key
* exchange mask and |*out_mask_a| to the authentication mask. */
void ssl_get_compatible_server_ciphers(SSL *ssl, uint32_t *out_mask_k,
uint32_t *out_mask_a);
STACK_OF(SSL_CIPHER) *ssl_get_ciphers_by_id(SSL *ssl);
int ssl_verify_alarm_type(long type);
int ssl3_get_finished(SSL *ssl);
int ssl3_send_change_cipher_spec(SSL *ssl);
void ssl3_cleanup_key_block(SSL *ssl);
int ssl3_send_alert(SSL *ssl, int level, int desc);
int ssl3_get_message(SSL *ssl, int msg_type,
enum ssl_hash_message_t hash_message);
int ssl3_hash_current_message(SSL *ssl);
void ssl3_release_current_message(SSL *ssl, int free_buffer);
/* ssl3_cert_verify_hash writes the SSL 3.0 CertificateVerify hash into the
* bytes pointed to by |out| and writes the number of bytes to |*out_len|. |out|
* must have room for |EVP_MAX_MD_SIZE| bytes. It sets |*out_md| to the hash
* function used. It returns one on success and zero on failure. */
int ssl3_cert_verify_hash(SSL *ssl, const EVP_MD **out_md, uint8_t *out,
size_t *out_len, uint16_t signature_algorithm);
int ssl3_send_finished(SSL *ssl, int a, int b);
int ssl3_supports_cipher(const SSL_CIPHER *cipher);
int ssl3_dispatch_alert(SSL *ssl);
int ssl3_read_app_data(SSL *ssl, int *out_got_handshake, uint8_t *buf, int len,
int peek);
int ssl3_read_change_cipher_spec(SSL *ssl);
void ssl3_read_close_notify(SSL *ssl);
int ssl3_read_handshake_bytes(SSL *ssl, uint8_t *buf, int len);
int ssl3_write_app_data(SSL *ssl, const void *buf, int len);
int ssl3_write_bytes(SSL *ssl, int type, const void *buf, int len);
int ssl3_output_cert_chain(SSL *ssl);
const SSL_CIPHER *ssl3_choose_cipher(
SSL *ssl, const struct ssl_early_callback_ctx *client_hello,
const struct ssl_cipher_preference_list_st *srvr);
int ssl3_new(SSL *ssl);
void ssl3_free(SSL *ssl);
int ssl3_accept(SSL *ssl);
int ssl3_connect(SSL *ssl);
int ssl3_init_message(SSL *ssl, CBB *cbb, CBB *body, uint8_t type);
int ssl3_finish_message(SSL *ssl, CBB *cbb);
int ssl3_write_message(SSL *ssl);
void ssl3_expect_flight(SSL *ssl);
void ssl3_received_flight(SSL *ssl);
int dtls1_init_message(SSL *ssl, CBB *cbb, CBB *body, uint8_t type);
int dtls1_finish_message(SSL *ssl, CBB *cbb);
int dtls1_write_message(SSL *ssl);
/* dtls1_get_record reads a new input record. On success, it places it in
* |ssl->s3->rrec| and returns one. Otherwise it returns <= 0 on error or if
* more data is needed. */
int dtls1_get_record(SSL *ssl);
int dtls1_read_app_data(SSL *ssl, int *out_got_handshake, uint8_t *buf, int len,
int peek);
int dtls1_read_change_cipher_spec(SSL *ssl);
void dtls1_read_close_notify(SSL *ssl);
int dtls1_write_app_data(SSL *ssl, const void *buf, int len);
/* dtls1_write_record sends a record. It returns one on success and <= 0 on
* error. */
int dtls1_write_record(SSL *ssl, int type, const uint8_t *buf, size_t len,
enum dtls1_use_epoch_t use_epoch);
int dtls1_send_change_cipher_spec(SSL *ssl);
int dtls1_send_finished(SSL *ssl, int a, int b, const char *sender, int slen);
int dtls1_retransmit_outgoing_messages(SSL *ssl);
void dtls1_clear_record_buffer(SSL *ssl);
int dtls1_parse_fragment(CBS *cbs, struct hm_header_st *out_hdr,
CBS *out_body);
int dtls1_check_timeout_num(SSL *ssl);
int dtls1_handshake_write(SSL *ssl);
void dtls1_expect_flight(SSL *ssl);
void dtls1_received_flight(SSL *ssl);
int dtls1_supports_cipher(const SSL_CIPHER *cipher);
void dtls1_start_timer(SSL *ssl);
void dtls1_stop_timer(SSL *ssl);
int dtls1_is_timer_expired(SSL *ssl);
void dtls1_double_timeout(SSL *ssl);
unsigned int dtls1_min_mtu(void);
int dtls1_new(SSL *ssl);
int dtls1_accept(SSL *ssl);
int dtls1_connect(SSL *ssl);
void dtls1_free(SSL *ssl);
int dtls1_get_message(SSL *ssl, int mt, enum ssl_hash_message_t hash_message);
int dtls1_hash_current_message(SSL *ssl);
void dtls1_release_current_message(SSL *ssl, int free_buffer);
int dtls1_dispatch_alert(SSL *ssl);
/* ssl_is_wbio_buffered returns one if |ssl|'s write BIO is buffered and zero
* otherwise. */
int ssl_is_wbio_buffered(const SSL *ssl);
int ssl_init_wbio_buffer(SSL *ssl);
void ssl_free_wbio_buffer(SSL *ssl);
int tls1_change_cipher_state(SSL *ssl, int which);
int tls1_setup_key_block(SSL *ssl);
int tls1_handshake_digest(SSL *ssl, uint8_t *out, size_t out_len);
int tls1_generate_master_secret(SSL *ssl, uint8_t *out, const uint8_t *premaster,
size_t premaster_len);
/* tls1_get_grouplist sets |*out_group_ids| and |*out_group_ids_len| to the
* list of allowed group IDs. If |get_peer_groups| is non-zero, return the
* peer's group list. Otherwise, return the preferred list. */
void tls1_get_grouplist(SSL *ssl, int get_peer_groups,
const uint16_t **out_group_ids,
size_t *out_group_ids_len);
/* tls1_check_group_id returns one if |group_id| is consistent with both our
* and the peer's group preferences. Note: if called as the client, only our
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
* preferences are checked; the peer (the server) does not send preferences. */
int tls1_check_group_id(SSL *ssl, uint16_t group_id);
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
/* tls1_get_shared_group sets |*out_group_id| to the first preferred shared
* group between client and server preferences and returns one. If none may be
Implement draft-ietf-tls-curve25519-01 in C. The new curve is not enabled by default. As EC_GROUP/EC_POINT is a bit too complex for X25519, this introduces an SSL_ECDH_METHOD abstraction which wraps just the raw ECDH operation. It also tidies up some of the curve code which kept converting back and force between NIDs and curve IDs. Now everything transits as curve IDs except for API entry points (SSL_set1_curves) which take NIDs. Those convert immediately and act on curve IDs from then on. Note that, like the Go implementation, this slightly tweaks the order of operations. The client sees the server public key before sending its own. To keep the abstraction simple, SSL_ECDH_METHOD expects to generate a keypair before consuming the peer's public key. Instead, the client handshake stashes the serialized peer public value and defers parsing it until it comes time to send ClientKeyExchange. (This is analogous to what it was doing before where it stashed the parsed peer public value instead.) It still uses TLS 1.2 terminology everywhere, but this abstraction should also be compatible with TLS 1.3 which unifies (EC)DH-style key exchanges. (Accordingly, this abstraction intentionally does not handle parsing the ClientKeyExchange/ServerKeyExchange framing or attempt to handle asynchronous plain RSA or the authentication bits.) BUG=571231 Change-Id: Iba09dddee5bcdfeb2b70185308e8ab0632717932 Reviewed-on: https://boringssl-review.googlesource.com/6780 Reviewed-by: Adam Langley <agl@google.com>
2015-12-19 05:18:25 +00:00
* found, it returns zero. */
int tls1_get_shared_group(SSL *ssl, uint16_t *out_group_id);
/* tls1_set_curves converts the array of |ncurves| NIDs pointed to by |curves|
* into a newly allocated array of TLS group IDs. On success, the function
* returns one and writes the array to |*out_group_ids| and its size to
* |*out_group_ids_len|. Otherwise, it returns zero. */
int tls1_set_curves(uint16_t **out_group_ids, size_t *out_group_ids_len,
const int *curves, size_t ncurves);
/* tls1_check_ec_cert returns one if |x| is an ECC certificate with curve and
* point format compatible with the client's preferences. Otherwise it returns
* zero. */
int tls1_check_ec_cert(SSL *ssl, X509 *x);
/* ssl_add_clienthello_tlsext writes ClientHello extensions to |out|. It
* returns one on success and zero on failure. The |header_len| argument is the
* length of the ClientHello written so far and is used to compute the padding
* length. (It does not include the record header.) */
int ssl_add_clienthello_tlsext(SSL *ssl, CBB *out, size_t header_len);
int ssl_add_serverhello_tlsext(SSL *ssl, CBB *out);
int ssl_parse_clienthello_tlsext(
SSL *ssl, const struct ssl_early_callback_ctx *client_hello);
int ssl_parse_serverhello_tlsext(SSL *ssl, CBS *cbs);
#define tlsext_tick_md EVP_sha256
/* tls_process_ticket processes a session ticket from the client. On success,
* it sets |*out_session| to the decrypted session or NULL if the ticket was
* rejected. If the ticket was valid, it sets |*out_renew_ticket| to whether
* the ticket should be renewed. It returns one on success and zero on fatal
* error. */
int tls_process_ticket(SSL *ssl, SSL_SESSION **out_session,
int *out_renew_ticket, const uint8_t *ticket,
size_t ticket_len, const uint8_t *session_id,
size_t session_id_len);
/* tls1_channel_id_hash computes the hash to be signed by Channel ID and writes
* it to |out|, which must contain at least |EVP_MAX_MD_SIZE| bytes. It returns
* one on success and zero on failure. */
int tls1_channel_id_hash(SSL *ssl, uint8_t *out, size_t *out_len);
int tls1_record_handshake_hashes_for_channel_id(SSL *ssl);
/* ssl3_can_false_start returns one if |ssl| is allowed to False Start and zero
* otherwise. */
int ssl3_can_false_start(const SSL *ssl);
/* ssl3_get_enc_method returns the SSL3_ENC_METHOD corresponding to
* |version|. */
const SSL3_ENC_METHOD *ssl3_get_enc_method(uint16_t version);
/* ssl_get_version_range sets |*out_min_version| and |*out_max_version| to the
* minimum and maximum enabled protocol versions, respectively. */
int ssl_get_version_range(const SSL *ssl, uint16_t *out_min_version,
uint16_t *out_max_version);
/* ssl3_protocol_version returns |ssl|'s protocol version. It is an error to
* call this function before the version is determined. */
uint16_t ssl3_protocol_version(const SSL *ssl);
uint32_t ssl_get_algorithm_prf(const SSL *ssl);
int tls1_parse_peer_sigalgs(SSL *ssl, const CBS *sigalgs);
/* tls1_choose_signature_algorithm sets |*out| to a signature algorithm for use
* with |ssl|'s private key based on the peer's preferences and the digests
* supported. It returns one on success and zero on error. */
int tls1_choose_signature_algorithm(SSL *ssl, uint16_t *out);
size_t tls12_get_psigalgs(SSL *ssl, const uint16_t **psigs);
/* tls12_check_peer_sigalg checks that |signature_algorithm| is consistent with
* |ssl|'s sent, supported signature algorithms and returns 1. Otherwise it
* returns 0 and writes an alert into |*out_alert|. */
int tls12_check_peer_sigalg(SSL *ssl, int *out_alert,
uint16_t signature_algorithm);
void ssl_set_client_disabled(SSL *ssl);
void ssl_get_current_time(const SSL *ssl, struct timeval *out_clock);
#if defined(__cplusplus)
} /* extern C */
#endif
#endif /* OPENSSL_HEADER_SSL_INTERNAL_H */