]>
Commit | Line | Data |
---|---|---|
81a6c781 | 1 | |
f1c236f8 | 2 | OpenSSL CHANGES |
651d0aff RE |
3 | _______________ |
4 | ||
9dcf5311 RS |
5 | This is a high-level summary of the most important changes. |
6 | For a full list of changes, see the git commit log; for example, | |
7 | https://github.com/openssl/openssl/commits/ and pick the appropriate | |
8 | release branch. | |
9 | ||
ec36b329 RL |
10 | Changes between 1.0.2s and 1.0.2t [xx XXX xxxx] |
11 | ||
12 | *) | |
13 | ||
cf9143f9 | 14 | Changes between 1.0.2r and 1.0.2s [28 May 2019] |
56ff0f64 | 15 | |
0f283c9a KR |
16 | *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024. |
17 | This changes the size when using the genpkey app when no size is given. It | |
18 | fixes an omission in earlier changes that changed all RSA, DSA and DH | |
19 | generation apps to use 2048 bits by default. | |
20 | [Kurt Roeckx] | |
21 | ||
f937540e DMSP |
22 | *) Add FIPS support for Android Arm 64-bit |
23 | ||
bb36ec5f RL |
24 | Support for Android Arm 64-bit was added to the OpenSSL FIPS Object |
25 | Module in Version 2.0.10. For some reason, the corresponding target | |
26 | 'android64-aarch64' was missing OpenSSL 1.0.2, whence it could not be | |
27 | built with FIPS support on Android Arm 64-bit. This omission has been | |
28 | fixed. | |
f937540e | 29 | [Matthias St. Pierre] |
56ff0f64 | 30 | |
b34cf4eb | 31 | Changes between 1.0.2q and 1.0.2r [26 Feb 2019] |
49d07eb3 | 32 | |
28c43df9 MC |
33 | *) 0-byte record padding oracle |
34 | ||
35 | If an application encounters a fatal protocol error and then calls | |
36 | SSL_shutdown() twice (once to send a close_notify, and once to receive one) | |
37 | then OpenSSL can respond differently to the calling application if a 0 byte | |
38 | record is received with invalid padding compared to if a 0 byte record is | |
39 | received with an invalid MAC. If the application then behaves differently | |
40 | based on that in a way that is detectable to the remote peer, then this | |
41 | amounts to a padding oracle that could be used to decrypt data. | |
42 | ||
43 | In order for this to be exploitable "non-stitched" ciphersuites must be in | |
44 | use. Stitched ciphersuites are optimised implementations of certain | |
45 | commonly used ciphersuites. Also the application must call SSL_shutdown() | |
46 | twice even if a protocol error has occurred (applications should not do | |
47 | this but some do anyway). | |
48 | ||
49 | This issue was discovered by Juraj Somorovsky, Robert Merget and Nimrod | |
50 | Aviram, with additional investigation by Steven Collison and Andrew | |
51 | Hourselt. It was reported to OpenSSL on 10th December 2018. | |
52 | (CVE-2019-1559) | |
53 | [Matt Caswell] | |
54 | ||
e42e5318 RL |
55 | *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0(). |
56 | [Richard Levitte] | |
49d07eb3 | 57 | |
5707219a | 58 | Changes between 1.0.2p and 1.0.2q [20 Nov 2018] |
8297ab58 | 59 | |
b18162a7 BB |
60 | *) Microarchitecture timing vulnerability in ECC scalar multiplication |
61 | ||
62 | OpenSSL ECC scalar multiplication, used in e.g. ECDSA and ECDH, has been | |
63 | shown to be vulnerable to a microarchitecture timing side channel attack. | |
64 | An attacker with sufficient access to mount local timing attacks during | |
65 | ECDSA signature generation could recover the private key. | |
66 | ||
67 | This issue was reported to OpenSSL on 26th October 2018 by Alejandro | |
68 | Cabrera Aldaya, Billy Brumley, Sohaib ul Hassan, Cesar Pereida Garcia and | |
69 | Nicola Tuveri. | |
70 | (CVE-2018-5407) | |
71 | [Billy Brumley] | |
72 | ||
548cce63 MC |
73 | *) Timing vulnerability in DSA signature generation |
74 | ||
75 | The OpenSSL DSA signature algorithm has been shown to be vulnerable to a | |
76 | timing side channel attack. An attacker could use variations in the signing | |
77 | algorithm to recover the private key. | |
78 | ||
79 | This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser. | |
80 | (CVE-2018-0734) | |
81 | [Paul Dale] | |
82 | ||
fff1da43 NT |
83 | *) Resolve a compatibility issue in EC_GROUP handling with the FIPS Object |
84 | Module, accidentally introduced while backporting security fixes from the | |
85 | development branch and hindering the use of ECC in FIPS mode. | |
86 | [Nicola Tuveri] | |
8297ab58 | 87 | |
e71ebf27 | 88 | Changes between 1.0.2o and 1.0.2p [14 Aug 2018] |
69a61c26 | 89 | |
0698c33a MC |
90 | *) Client DoS due to large DH parameter |
91 | ||
92 | During key agreement in a TLS handshake using a DH(E) based ciphersuite a | |
93 | malicious server can send a very large prime value to the client. This will | |
94 | cause the client to spend an unreasonably long period of time generating a | |
95 | key for this prime resulting in a hang until the client has finished. This | |
96 | could be exploited in a Denial Of Service attack. | |
97 | ||
98 | This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken | |
99 | (CVE-2018-0732) | |
100 | [Guido Vranken] | |
101 | ||
102 | *) Cache timing vulnerability in RSA Key Generation | |
103 | ||
104 | The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to | |
105 | a cache timing side channel attack. An attacker with sufficient access to | |
106 | mount cache timing attacks during the RSA key generation process could | |
107 | recover the private key. | |
108 | ||
109 | This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera | |
110 | Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia. | |
111 | (CVE-2018-0737) | |
112 | [Billy Brumley] | |
113 | ||
f72a7ce8 RL |
114 | *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str |
115 | parameter is no longer accepted, as it leads to a corrupt table. NULL | |
116 | pem_str is reserved for alias entries only. | |
117 | [Richard Levitte] | |
118 | ||
29d8bda9 AP |
119 | *) Revert blinding in ECDSA sign and instead make problematic addition |
120 | length-invariant. Switch even to fixed-length Montgomery multiplication. | |
121 | [Andy Polyakov] | |
122 | ||
be4e1f79 KR |
123 | *) Change generating and checking of primes so that the error rate of not |
124 | being prime depends on the intended use based on the size of the input. | |
125 | For larger primes this will result in more rounds of Miller-Rabin. | |
126 | The maximal error rate for primes with more than 1080 bits is lowered | |
127 | to 2^-128. | |
128 | [Kurt Roeckx, Annie Yousar] | |
129 | ||
7a23bff9 KR |
130 | *) Increase the number of Miller-Rabin rounds for DSA key generating to 64. |
131 | [Kurt Roeckx] | |
132 | ||
41d23d43 MC |
133 | *) Add blinding to ECDSA and DSA signatures to protect against side channel |
134 | attacks discovered by Keegan Ryan (NCC Group). | |
949ff366 MC |
135 | [Matt Caswell] |
136 | ||
18026c02 RL |
137 | *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we |
138 | now allow empty (zero character) pass phrases. | |
139 | [Richard Levitte] | |
140 | ||
7b6cfcd6 EK |
141 | *) Certificate time validation (X509_cmp_time) enforces stricter |
142 | compliance with RFC 5280. Fractional seconds and timezone offsets | |
143 | are no longer allowed. | |
144 | [Emilia Käsper] | |
69a61c26 | 145 | |
3ce7bc40 | 146 | Changes between 1.0.2n and 1.0.2o [27 Mar 2018] |
ebe18302 | 147 | |
b621f604 MC |
148 | *) Constructed ASN.1 types with a recursive definition could exceed the stack |
149 | ||
150 | Constructed ASN.1 types with a recursive definition (such as can be found | |
151 | in PKCS7) could eventually exceed the stack given malicious input with | |
152 | excessive recursion. This could result in a Denial Of Service attack. There | |
153 | are no such structures used within SSL/TLS that come from untrusted sources | |
154 | so this is considered safe. | |
155 | ||
156 | This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz | |
157 | project. | |
158 | (CVE-2018-0739) | |
159 | [Matt Caswell] | |
ebe18302 | 160 | |
e5bba24c | 161 | Changes between 1.0.2m and 1.0.2n [7 Dec 2017] |
95aec441 | 162 | |
f3b6b413 MC |
163 | *) Read/write after SSL object in error state |
164 | ||
165 | OpenSSL 1.0.2 (starting from version 1.0.2b) introduced an "error state" | |
166 | mechanism. The intent was that if a fatal error occurred during a handshake | |
167 | then OpenSSL would move into the error state and would immediately fail if | |
168 | you attempted to continue the handshake. This works as designed for the | |
169 | explicit handshake functions (SSL_do_handshake(), SSL_accept() and | |
170 | SSL_connect()), however due to a bug it does not work correctly if | |
171 | SSL_read() or SSL_write() is called directly. In that scenario, if the | |
172 | handshake fails then a fatal error will be returned in the initial function | |
173 | call. If SSL_read()/SSL_write() is subsequently called by the application | |
174 | for the same SSL object then it will succeed and the data is passed without | |
175 | being decrypted/encrypted directly from the SSL/TLS record layer. | |
176 | ||
177 | In order to exploit this issue an application bug would have to be present | |
178 | that resulted in a call to SSL_read()/SSL_write() being issued after having | |
179 | already received a fatal error. | |
180 | ||
181 | This issue was reported to OpenSSL by David Benjamin (Google). | |
182 | (CVE-2017-3737) | |
183 | [Matt Caswell] | |
184 | ||
185 | *) rsaz_1024_mul_avx2 overflow bug on x86_64 | |
186 | ||
187 | There is an overflow bug in the AVX2 Montgomery multiplication procedure | |
188 | used in exponentiation with 1024-bit moduli. No EC algorithms are affected. | |
189 | Analysis suggests that attacks against RSA and DSA as a result of this | |
190 | defect would be very difficult to perform and are not believed likely. | |
191 | Attacks against DH1024 are considered just feasible, because most of the | |
192 | work necessary to deduce information about a private key may be performed | |
193 | offline. The amount of resources required for such an attack would be | |
194 | significant. However, for an attack on TLS to be meaningful, the server | |
195 | would have to share the DH1024 private key among multiple clients, which is | |
196 | no longer an option since CVE-2016-0701. | |
197 | ||
198 | This only affects processors that support the AVX2 but not ADX extensions | |
199 | like Intel Haswell (4th generation). | |
200 | ||
201 | This issue was reported to OpenSSL by David Benjamin (Google). The issue | |
202 | was originally found via the OSS-Fuzz project. | |
203 | (CVE-2017-3738) | |
204 | [Andy Polyakov] | |
95aec441 | 205 | |
8b1549a1 | 206 | Changes between 1.0.2l and 1.0.2m [2 Nov 2017] |
22d41cd3 | 207 | |
64c46a98 MC |
208 | *) bn_sqrx8x_internal carry bug on x86_64 |
209 | ||
210 | There is a carry propagating bug in the x86_64 Montgomery squaring | |
211 | procedure. No EC algorithms are affected. Analysis suggests that attacks | |
212 | against RSA and DSA as a result of this defect would be very difficult to | |
213 | perform and are not believed likely. Attacks against DH are considered just | |
214 | feasible (although very difficult) because most of the work necessary to | |
215 | deduce information about a private key may be performed offline. The amount | |
216 | of resources required for such an attack would be very significant and | |
217 | likely only accessible to a limited number of attackers. An attacker would | |
218 | additionally need online access to an unpatched system using the target | |
219 | private key in a scenario with persistent DH parameters and a private | |
220 | key that is shared between multiple clients. | |
221 | ||
222 | This only affects processors that support the BMI1, BMI2 and ADX extensions | |
223 | like Intel Broadwell (5th generation) and later or AMD Ryzen. | |
224 | ||
225 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
226 | (CVE-2017-3736) | |
227 | [Andy Polyakov] | |
228 | ||
229 | *) Malformed X.509 IPAddressFamily could cause OOB read | |
230 | ||
231 | If an X.509 certificate has a malformed IPAddressFamily extension, | |
232 | OpenSSL could do a one-byte buffer overread. The most likely result | |
233 | would be an erroneous display of the certificate in text format. | |
234 | ||
235 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
236 | (CVE-2017-3735) | |
237 | [Rich Salz] | |
22d41cd3 | 238 | |
b3a3bab0 | 239 | Changes between 1.0.2k and 1.0.2l [25 May 2017] |
f24fcf29 | 240 | |
7b64c79b RL |
241 | *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target |
242 | platform rather than 'mingw'. | |
243 | [Richard Levitte] | |
244 | ||
081314d0 | 245 | Changes between 1.0.2j and 1.0.2k [26 Jan 2017] |
f6e43fee | 246 | |
06f87e96 MC |
247 | *) Truncated packet could crash via OOB read |
248 | ||
249 | If one side of an SSL/TLS path is running on a 32-bit host and a specific | |
250 | cipher is being used, then a truncated packet can cause that host to | |
251 | perform an out-of-bounds read, usually resulting in a crash. | |
252 | ||
253 | This issue was reported to OpenSSL by Robert Święcki of Google. | |
254 | (CVE-2017-3731) | |
255 | [Andy Polyakov] | |
256 | ||
257 | *) BN_mod_exp may produce incorrect results on x86_64 | |
258 | ||
259 | There is a carry propagating bug in the x86_64 Montgomery squaring | |
260 | procedure. No EC algorithms are affected. Analysis suggests that attacks | |
261 | against RSA and DSA as a result of this defect would be very difficult to | |
262 | perform and are not believed likely. Attacks against DH are considered just | |
263 | feasible (although very difficult) because most of the work necessary to | |
264 | deduce information about a private key may be performed offline. The amount | |
265 | of resources required for such an attack would be very significant and | |
266 | likely only accessible to a limited number of attackers. An attacker would | |
267 | additionally need online access to an unpatched system using the target | |
268 | private key in a scenario with persistent DH parameters and a private | |
269 | key that is shared between multiple clients. For example this can occur by | |
270 | default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very | |
271 | similar to CVE-2015-3193 but must be treated as a separate problem. | |
272 | ||
273 | This issue was reported to OpenSSL by the OSS-Fuzz project. | |
274 | (CVE-2017-3732) | |
275 | [Andy Polyakov] | |
276 | ||
19e1de54 MC |
277 | *) Montgomery multiplication may produce incorrect results |
278 | ||
279 | There is a carry propagating bug in the Broadwell-specific Montgomery | |
280 | multiplication procedure that handles input lengths divisible by, but | |
281 | longer than 256 bits. Analysis suggests that attacks against RSA, DSA | |
282 | and DH private keys are impossible. This is because the subroutine in | |
283 | question is not used in operations with the private key itself and an input | |
284 | of the attacker's direct choice. Otherwise the bug can manifest itself as | |
285 | transient authentication and key negotiation failures or reproducible | |
286 | erroneous outcome of public-key operations with specially crafted input. | |
287 | Among EC algorithms only Brainpool P-512 curves are affected and one | |
288 | presumably can attack ECDH key negotiation. Impact was not analyzed in | |
289 | detail, because pre-requisites for attack are considered unlikely. Namely | |
290 | multiple clients have to choose the curve in question and the server has to | |
291 | share the private key among them, neither of which is default behaviour. | |
292 | Even then only clients that chose the curve will be affected. | |
293 | ||
294 | This issue was publicly reported as transient failures and was not | |
295 | initially recognized as a security issue. Thanks to Richard Morgan for | |
296 | providing reproducible case. | |
297 | (CVE-2016-7055) | |
298 | [Andy Polyakov] | |
299 | ||
563a34e1 MC |
300 | *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0 |
301 | or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to | |
302 | prevent issues where no progress is being made and the peer continually | |
303 | sends unrecognised record types, using up resources processing them. | |
304 | [Matt Caswell] | |
f6e43fee | 305 | |
e216bf9d | 306 | Changes between 1.0.2i and 1.0.2j [26 Sep 2016] |
9d264d11 | 307 | |
ca430ece MC |
308 | *) Missing CRL sanity check |
309 | ||
310 | A bug fix which included a CRL sanity check was added to OpenSSL 1.1.0 | |
311 | but was omitted from OpenSSL 1.0.2i. As a result any attempt to use | |
312 | CRLs in OpenSSL 1.0.2i will crash with a null pointer exception. | |
313 | ||
314 | This issue only affects the OpenSSL 1.0.2i | |
315 | (CVE-2016-7052) | |
316 | [Matt Caswell] | |
9d264d11 | 317 | |
32c13016 | 318 | Changes between 1.0.2h and 1.0.2i [22 Sep 2016] |
5c694459 | 319 | |
35aede1c | 320 | *) OCSP Status Request extension unbounded memory growth |
5c694459 | 321 | |
35aede1c MC |
322 | A malicious client can send an excessively large OCSP Status Request |
323 | extension. If that client continually requests renegotiation, sending a | |
324 | large OCSP Status Request extension each time, then there will be unbounded | |
325 | memory growth on the server. This will eventually lead to a Denial Of | |
326 | Service attack through memory exhaustion. Servers with a default | |
327 | configuration are vulnerable even if they do not support OCSP. Builds using | |
328 | the "no-ocsp" build time option are not affected. | |
329 | ||
330 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
331 | (CVE-2016-6304) | |
332 | [Matt Caswell] | |
333 | ||
334 | *) In order to mitigate the SWEET32 attack, the DES ciphers were moved from | |
335 | HIGH to MEDIUM. | |
336 | ||
337 | This issue was reported to OpenSSL Karthikeyan Bhargavan and Gaetan | |
338 | Leurent (INRIA) | |
339 | (CVE-2016-2183) | |
0fff5065 RS |
340 | [Rich Salz] |
341 | ||
35aede1c MC |
342 | *) OOB write in MDC2_Update() |
343 | ||
344 | An overflow can occur in MDC2_Update() either if called directly or | |
345 | through the EVP_DigestUpdate() function using MDC2. If an attacker | |
346 | is able to supply very large amounts of input data after a previous | |
347 | call to EVP_EncryptUpdate() with a partial block then a length check | |
348 | can overflow resulting in a heap corruption. | |
349 | ||
350 | The amount of data needed is comparable to SIZE_MAX which is impractical | |
351 | on most platforms. | |
352 | ||
353 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
354 | (CVE-2016-6303) | |
355 | [Stephen Henson] | |
356 | ||
357 | *) Malformed SHA512 ticket DoS | |
358 | ||
359 | If a server uses SHA512 for TLS session ticket HMAC it is vulnerable to a | |
360 | DoS attack where a malformed ticket will result in an OOB read which will | |
361 | ultimately crash. | |
362 | ||
363 | The use of SHA512 in TLS session tickets is comparatively rare as it requires | |
364 | a custom server callback and ticket lookup mechanism. | |
365 | ||
366 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
367 | (CVE-2016-6302) | |
368 | [Stephen Henson] | |
369 | ||
370 | *) OOB write in BN_bn2dec() | |
371 | ||
372 | The function BN_bn2dec() does not check the return value of BN_div_word(). | |
373 | This can cause an OOB write if an application uses this function with an | |
374 | overly large BIGNUM. This could be a problem if an overly large certificate | |
375 | or CRL is printed out from an untrusted source. TLS is not affected because | |
376 | record limits will reject an oversized certificate before it is parsed. | |
377 | ||
378 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
379 | (CVE-2016-2182) | |
380 | [Stephen Henson] | |
381 | ||
382 | *) OOB read in TS_OBJ_print_bio() | |
383 | ||
384 | The function TS_OBJ_print_bio() misuses OBJ_obj2txt(): the return value is | |
385 | the total length the OID text representation would use and not the amount | |
386 | of data written. This will result in OOB reads when large OIDs are | |
387 | presented. | |
388 | ||
389 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
390 | (CVE-2016-2180) | |
391 | [Stephen Henson] | |
392 | ||
393 | *) Pointer arithmetic undefined behaviour | |
394 | ||
395 | Avoid some undefined pointer arithmetic | |
396 | ||
397 | A common idiom in the codebase is to check limits in the following manner: | |
398 | "p + len > limit" | |
399 | ||
400 | Where "p" points to some malloc'd data of SIZE bytes and | |
401 | limit == p + SIZE | |
402 | ||
403 | "len" here could be from some externally supplied data (e.g. from a TLS | |
404 | message). | |
405 | ||
406 | The rules of C pointer arithmetic are such that "p + len" is only well | |
407 | defined where len <= SIZE. Therefore the above idiom is actually | |
408 | undefined behaviour. | |
409 | ||
410 | For example this could cause problems if some malloc implementation | |
411 | provides an address for "p" such that "p + len" actually overflows for | |
412 | values of len that are too big and therefore p + len < limit. | |
413 | ||
414 | This issue was reported to OpenSSL by Guido Vranken | |
415 | (CVE-2016-2177) | |
416 | [Matt Caswell] | |
417 | ||
418 | *) Constant time flag not preserved in DSA signing | |
419 | ||
420 | Operations in the DSA signing algorithm should run in constant time in | |
421 | order to avoid side channel attacks. A flaw in the OpenSSL DSA | |
422 | implementation means that a non-constant time codepath is followed for | |
423 | certain operations. This has been demonstrated through a cache-timing | |
424 | attack to be sufficient for an attacker to recover the private DSA key. | |
425 | ||
426 | This issue was reported by César Pereida (Aalto University), Billy Brumley | |
427 | (Tampere University of Technology), and Yuval Yarom (The University of | |
428 | Adelaide and NICTA). | |
429 | (CVE-2016-2178) | |
430 | [César Pereida] | |
431 | ||
432 | *) DTLS buffered message DoS | |
433 | ||
434 | In a DTLS connection where handshake messages are delivered out-of-order | |
435 | those messages that OpenSSL is not yet ready to process will be buffered | |
436 | for later use. Under certain circumstances, a flaw in the logic means that | |
437 | those messages do not get removed from the buffer even though the handshake | |
438 | has been completed. An attacker could force up to approx. 15 messages to | |
439 | remain in the buffer when they are no longer required. These messages will | |
440 | be cleared when the DTLS connection is closed. The default maximum size for | |
441 | a message is 100k. Therefore the attacker could force an additional 1500k | |
442 | to be consumed per connection. By opening many simulataneous connections an | |
443 | attacker could cause a DoS attack through memory exhaustion. | |
444 | ||
445 | This issue was reported to OpenSSL by Quan Luo. | |
446 | (CVE-2016-2179) | |
447 | [Matt Caswell] | |
448 | ||
449 | *) DTLS replay protection DoS | |
450 | ||
451 | A flaw in the DTLS replay attack protection mechanism means that records | |
452 | that arrive for future epochs update the replay protection "window" before | |
453 | the MAC for the record has been validated. This could be exploited by an | |
454 | attacker by sending a record for the next epoch (which does not have to | |
455 | decrypt or have a valid MAC), with a very large sequence number. This means | |
456 | that all subsequent legitimate packets are dropped causing a denial of | |
457 | service for a specific DTLS connection. | |
458 | ||
459 | This issue was reported to OpenSSL by the OCAP audit team. | |
460 | (CVE-2016-2181) | |
461 | [Matt Caswell] | |
462 | ||
463 | *) Certificate message OOB reads | |
464 | ||
465 | In OpenSSL 1.0.2 and earlier some missing message length checks can result | |
466 | in OOB reads of up to 2 bytes beyond an allocated buffer. There is a | |
467 | theoretical DoS risk but this has not been observed in practice on common | |
468 | platforms. | |
469 | ||
470 | The messages affected are client certificate, client certificate request | |
471 | and server certificate. As a result the attack can only be performed | |
472 | against a client or a server which enables client authentication. | |
473 | ||
474 | This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.) | |
475 | (CVE-2016-6306) | |
476 | [Stephen Henson] | |
477 | ||
5dd94f18 | 478 | Changes between 1.0.2g and 1.0.2h [3 May 2016] |
a5006916 | 479 | |
b4d56b8e MC |
480 | *) Prevent padding oracle in AES-NI CBC MAC check |
481 | ||
482 | A MITM attacker can use a padding oracle attack to decrypt traffic | |
483 | when the connection uses an AES CBC cipher and the server support | |
484 | AES-NI. | |
485 | ||
486 | This issue was introduced as part of the fix for Lucky 13 padding | |
487 | attack (CVE-2013-0169). The padding check was rewritten to be in | |
488 | constant time by making sure that always the same bytes are read and | |
489 | compared against either the MAC or padding bytes. But it no longer | |
490 | checked that there was enough data to have both the MAC and padding | |
491 | bytes. | |
492 | ||
493 | This issue was reported by Juraj Somorovsky using TLS-Attacker. | |
494 | (CVE-2016-2107) | |
495 | [Kurt Roeckx] | |
496 | ||
497 | *) Fix EVP_EncodeUpdate overflow | |
498 | ||
499 | An overflow can occur in the EVP_EncodeUpdate() function which is used for | |
500 | Base64 encoding of binary data. If an attacker is able to supply very large | |
501 | amounts of input data then a length check can overflow resulting in a heap | |
502 | corruption. | |
503 | ||
504 | Internally to OpenSSL the EVP_EncodeUpdate() function is primarly used by | |
505 | the PEM_write_bio* family of functions. These are mainly used within the | |
506 | OpenSSL command line applications, so any application which processes data | |
507 | from an untrusted source and outputs it as a PEM file should be considered | |
508 | vulnerable to this issue. User applications that call these APIs directly | |
509 | with large amounts of untrusted data may also be vulnerable. | |
510 | ||
511 | This issue was reported by Guido Vranken. | |
512 | (CVE-2016-2105) | |
513 | [Matt Caswell] | |
514 | ||
515 | *) Fix EVP_EncryptUpdate overflow | |
516 | ||
517 | An overflow can occur in the EVP_EncryptUpdate() function. If an attacker | |
518 | is able to supply very large amounts of input data after a previous call to | |
519 | EVP_EncryptUpdate() with a partial block then a length check can overflow | |
520 | resulting in a heap corruption. Following an analysis of all OpenSSL | |
521 | internal usage of the EVP_EncryptUpdate() function all usage is one of two | |
522 | forms. The first form is where the EVP_EncryptUpdate() call is known to be | |
523 | the first called function after an EVP_EncryptInit(), and therefore that | |
524 | specific call must be safe. The second form is where the length passed to | |
525 | EVP_EncryptUpdate() can be seen from the code to be some small value and | |
526 | therefore there is no possibility of an overflow. Since all instances are | |
527 | one of these two forms, it is believed that there can be no overflows in | |
528 | internal code due to this problem. It should be noted that | |
529 | EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths. | |
530 | Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances | |
531 | of these calls have also been analysed too and it is believed there are no | |
532 | instances in internal usage where an overflow could occur. | |
533 | ||
534 | This issue was reported by Guido Vranken. | |
535 | (CVE-2016-2106) | |
536 | [Matt Caswell] | |
537 | ||
538 | *) Prevent ASN.1 BIO excessive memory allocation | |
539 | ||
540 | When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio() | |
541 | a short invalid encoding can casuse allocation of large amounts of memory | |
542 | potentially consuming excessive resources or exhausting memory. | |
543 | ||
544 | Any application parsing untrusted data through d2i BIO functions is | |
545 | affected. The memory based functions such as d2i_X509() are *not* affected. | |
546 | Since the memory based functions are used by the TLS library, TLS | |
547 | applications are not affected. | |
548 | ||
549 | This issue was reported by Brian Carpenter. | |
550 | (CVE-2016-2109) | |
551 | [Stephen Henson] | |
552 | ||
553 | *) EBCDIC overread | |
554 | ||
555 | ASN1 Strings that are over 1024 bytes can cause an overread in applications | |
556 | using the X509_NAME_oneline() function on EBCDIC systems. This could result | |
557 | in arbitrary stack data being returned in the buffer. | |
558 | ||
559 | This issue was reported by Guido Vranken. | |
560 | (CVE-2016-2176) | |
561 | [Matt Caswell] | |
562 | ||
af2db04c TS |
563 | *) Modify behavior of ALPN to invoke callback after SNI/servername |
564 | callback, such that updates to the SSL_CTX affect ALPN. | |
565 | [Todd Short] | |
566 | ||
29cce508 KR |
567 | *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the |
568 | default. | |
569 | [Kurt Roeckx] | |
a5006916 | 570 | |
42569575 KR |
571 | *) Only remove the SSLv2 methods with the no-ssl2-method option. When the |
572 | methods are enabled and ssl2 is disabled the methods return NULL. | |
573 | [Kurt Roeckx] | |
574 | ||
902f3f50 | 575 | Changes between 1.0.2f and 1.0.2g [1 Mar 2016] |
22d192f1 | 576 | |
bc38a7d2 VD |
577 | * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL. |
578 | Builds that are not configured with "enable-weak-ssl-ciphers" will not | |
579 | provide any "EXPORT" or "LOW" strength ciphers. | |
580 | [Viktor Dukhovni] | |
581 | ||
9dfd2be8 VD |
582 | * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2 |
583 | is by default disabled at build-time. Builds that are not configured with | |
584 | "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used, | |
585 | users who want to negotiate SSLv2 via the version-flexible SSLv23_method() | |
586 | will need to explicitly call either of: | |
587 | ||
588 | SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2); | |
589 | or | |
590 | SSL_clear_options(ssl, SSL_OP_NO_SSLv2); | |
591 | ||
592 | as appropriate. Even if either of those is used, or the application | |
593 | explicitly uses the version-specific SSLv2_method() or its client and | |
594 | server variants, SSLv2 ciphers vulnerable to exhaustive search key | |
595 | recovery have been removed. Specifically, the SSLv2 40-bit EXPORT | |
596 | ciphers, and SSLv2 56-bit DES are no longer available. | |
248808c8 | 597 | (CVE-2016-0800) |
9dfd2be8 | 598 | [Viktor Dukhovni] |
248808c8 MC |
599 | |
600 | *) Fix a double-free in DSA code | |
601 | ||
602 | A double free bug was discovered when OpenSSL parses malformed DSA private | |
603 | keys and could lead to a DoS attack or memory corruption for applications | |
604 | that receive DSA private keys from untrusted sources. This scenario is | |
605 | considered rare. | |
606 | ||
607 | This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using | |
608 | libFuzzer. | |
609 | (CVE-2016-0705) | |
610 | [Stephen Henson] | |
611 | ||
259b664f EK |
612 | *) Disable SRP fake user seed to address a server memory leak. |
613 | ||
614 | Add a new method SRP_VBASE_get1_by_user that handles the seed properly. | |
615 | ||
616 | SRP_VBASE_get_by_user had inconsistent memory management behaviour. | |
617 | In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user | |
618 | was changed to ignore the "fake user" SRP seed, even if the seed | |
619 | is configured. | |
620 | ||
621 | Users should use SRP_VBASE_get1_by_user instead. Note that in | |
622 | SRP_VBASE_get1_by_user, caller must free the returned value. Note | |
623 | also that even though configuring the SRP seed attempts to hide | |
624 | invalid usernames by continuing the handshake with fake | |
625 | credentials, this behaviour is not constant time and no strong | |
626 | guarantees are made that the handshake is indistinguishable from | |
627 | that of a valid user. | |
628 | (CVE-2016-0798) | |
629 | [Emilia Käsper] | |
630 | ||
248808c8 MC |
631 | *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption |
632 | ||
633 | In the BN_hex2bn function the number of hex digits is calculated using an | |
634 | int value |i|. Later |bn_expand| is called with a value of |i * 4|. For | |
635 | large values of |i| this can result in |bn_expand| not allocating any | |
636 | memory because |i * 4| is negative. This can leave the internal BIGNUM data | |
637 | field as NULL leading to a subsequent NULL ptr deref. For very large values | |
638 | of |i|, the calculation |i * 4| could be a positive value smaller than |i|. | |
639 | In this case memory is allocated to the internal BIGNUM data field, but it | |
640 | is insufficiently sized leading to heap corruption. A similar issue exists | |
641 | in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn | |
642 | is ever called by user applications with very large untrusted hex/dec data. | |
643 | This is anticipated to be a rare occurrence. | |
644 | ||
645 | All OpenSSL internal usage of these functions use data that is not expected | |
646 | to be untrusted, e.g. config file data or application command line | |
647 | arguments. If user developed applications generate config file data based | |
648 | on untrusted data then it is possible that this could also lead to security | |
649 | consequences. This is also anticipated to be rare. | |
650 | ||
651 | This issue was reported to OpenSSL by Guido Vranken. | |
652 | (CVE-2016-0797) | |
653 | [Matt Caswell] | |
654 | ||
655 | *) Fix memory issues in BIO_*printf functions | |
656 | ||
657 | The internal |fmtstr| function used in processing a "%s" format string in | |
658 | the BIO_*printf functions could overflow while calculating the length of a | |
659 | string and cause an OOB read when printing very long strings. | |
660 | ||
661 | Additionally the internal |doapr_outch| function can attempt to write to an | |
662 | OOB memory location (at an offset from the NULL pointer) in the event of a | |
663 | memory allocation failure. In 1.0.2 and below this could be caused where | |
664 | the size of a buffer to be allocated is greater than INT_MAX. E.g. this | |
665 | could be in processing a very long "%s" format string. Memory leaks can | |
666 | also occur. | |
667 | ||
668 | The first issue may mask the second issue dependent on compiler behaviour. | |
669 | These problems could enable attacks where large amounts of untrusted data | |
670 | is passed to the BIO_*printf functions. If applications use these functions | |
671 | in this way then they could be vulnerable. OpenSSL itself uses these | |
672 | functions when printing out human-readable dumps of ASN.1 data. Therefore | |
673 | applications that print this data could be vulnerable if the data is from | |
674 | untrusted sources. OpenSSL command line applications could also be | |
675 | vulnerable where they print out ASN.1 data, or if untrusted data is passed | |
676 | as command line arguments. | |
677 | ||
678 | Libssl is not considered directly vulnerable. Additionally certificates etc | |
679 | received via remote connections via libssl are also unlikely to be able to | |
680 | trigger these issues because of message size limits enforced within libssl. | |
681 | ||
682 | This issue was reported to OpenSSL Guido Vranken. | |
683 | (CVE-2016-0799) | |
684 | [Matt Caswell] | |
685 | ||
686 | *) Side channel attack on modular exponentiation | |
687 | ||
688 | A side-channel attack was found which makes use of cache-bank conflicts on | |
689 | the Intel Sandy-Bridge microarchitecture which could lead to the recovery | |
690 | of RSA keys. The ability to exploit this issue is limited as it relies on | |
691 | an attacker who has control of code in a thread running on the same | |
692 | hyper-threaded core as the victim thread which is performing decryptions. | |
693 | ||
694 | This issue was reported to OpenSSL by Yuval Yarom, The University of | |
695 | Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and | |
696 | Nadia Heninger, University of Pennsylvania with more information at | |
697 | http://cachebleed.info. | |
698 | (CVE-2016-0702) | |
699 | [Andy Polyakov] | |
700 | ||
b4b23d05 EK |
701 | *) Change the req app to generate a 2048-bit RSA/DSA key by default, |
702 | if no keysize is specified with default_bits. This fixes an | |
703 | omission in an earlier change that changed all RSA/DSA key generation | |
704 | apps to use 2048 bits by default. | |
705 | [Emilia Käsper] | |
22d192f1 | 706 | |
95605f3a | 707 | Changes between 1.0.2e and 1.0.2f [28 Jan 2016] |
8a27243c | 708 | |
f26a179a MC |
709 | *) DH small subgroups |
710 | ||
711 | Historically OpenSSL only ever generated DH parameters based on "safe" | |
712 | primes. More recently (in version 1.0.2) support was provided for | |
713 | generating X9.42 style parameter files such as those required for RFC 5114 | |
714 | support. The primes used in such files may not be "safe". Where an | |
715 | application is using DH configured with parameters based on primes that are | |
716 | not "safe" then an attacker could use this fact to find a peer's private | |
717 | DH exponent. This attack requires that the attacker complete multiple | |
718 | handshakes in which the peer uses the same private DH exponent. For example | |
719 | this could be used to discover a TLS server's private DH exponent if it's | |
720 | reusing the private DH exponent or it's using a static DH ciphersuite. | |
721 | ||
722 | OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in | |
723 | TLS. It is not on by default. If the option is not set then the server | |
724 | reuses the same private DH exponent for the life of the server process and | |
725 | would be vulnerable to this attack. It is believed that many popular | |
726 | applications do set this option and would therefore not be at risk. | |
727 | ||
728 | The fix for this issue adds an additional check where a "q" parameter is | |
729 | available (as is the case in X9.42 based parameters). This detects the | |
730 | only known attack, and is the only possible defense for static DH | |
731 | ciphersuites. This could have some performance impact. | |
732 | ||
733 | Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by | |
734 | default and cannot be disabled. This could have some performance impact. | |
735 | ||
736 | This issue was reported to OpenSSL by Antonio Sanso (Adobe). | |
737 | (CVE-2016-0701) | |
738 | [Matt Caswell] | |
739 | ||
740 | *) SSLv2 doesn't block disabled ciphers | |
741 | ||
742 | A malicious client can negotiate SSLv2 ciphers that have been disabled on | |
743 | the server and complete SSLv2 handshakes even if all SSLv2 ciphers have | |
744 | been disabled, provided that the SSLv2 protocol was not also disabled via | |
745 | SSL_OP_NO_SSLv2. | |
746 | ||
747 | This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram | |
748 | and Sebastian Schinzel. | |
749 | (CVE-2015-3197) | |
750 | [Viktor Dukhovni] | |
751 | ||
a4530ce0 KR |
752 | *) Reject DH handshakes with parameters shorter than 1024 bits. |
753 | [Kurt Roeckx] | |
8a27243c | 754 | |
bfe07df4 | 755 | Changes between 1.0.2d and 1.0.2e [3 Dec 2015] |
a7ef1e90 | 756 | |
9330fbd0 MC |
757 | *) BN_mod_exp may produce incorrect results on x86_64 |
758 | ||
759 | There is a carry propagating bug in the x86_64 Montgomery squaring | |
760 | procedure. No EC algorithms are affected. Analysis suggests that attacks | |
761 | against RSA and DSA as a result of this defect would be very difficult to | |
762 | perform and are not believed likely. Attacks against DH are considered just | |
763 | feasible (although very difficult) because most of the work necessary to | |
764 | deduce information about a private key may be performed offline. The amount | |
765 | of resources required for such an attack would be very significant and | |
766 | likely only accessible to a limited number of attackers. An attacker would | |
767 | additionally need online access to an unpatched system using the target | |
768 | private key in a scenario with persistent DH parameters and a private | |
769 | key that is shared between multiple clients. For example this can occur by | |
770 | default in OpenSSL DHE based SSL/TLS ciphersuites. | |
771 | ||
772 | This issue was reported to OpenSSL by Hanno Böck. | |
773 | (CVE-2015-3193) | |
774 | [Andy Polyakov] | |
775 | ||
776 | *) Certificate verify crash with missing PSS parameter | |
777 | ||
778 | The signature verification routines will crash with a NULL pointer | |
779 | dereference if presented with an ASN.1 signature using the RSA PSS | |
780 | algorithm and absent mask generation function parameter. Since these | |
781 | routines are used to verify certificate signature algorithms this can be | |
782 | used to crash any certificate verification operation and exploited in a | |
783 | DoS attack. Any application which performs certificate verification is | |
784 | vulnerable including OpenSSL clients and servers which enable client | |
785 | authentication. | |
786 | ||
787 | This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG). | |
788 | (CVE-2015-3194) | |
789 | [Stephen Henson] | |
790 | ||
791 | *) X509_ATTRIBUTE memory leak | |
792 | ||
793 | When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak | |
794 | memory. This structure is used by the PKCS#7 and CMS routines so any | |
795 | application which reads PKCS#7 or CMS data from untrusted sources is | |
796 | affected. SSL/TLS is not affected. | |
797 | ||
798 | This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using | |
799 | libFuzzer. | |
800 | (CVE-2015-3195) | |
801 | [Stephen Henson] | |
802 | ||
37faf117 EK |
803 | *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs. |
804 | This changes the decoding behaviour for some invalid messages, | |
805 | though the change is mostly in the more lenient direction, and | |
806 | legacy behaviour is preserved as much as possible. | |
807 | [Emilia Käsper] | |
808 | ||
1d7df236 | 809 | *) In DSA_generate_parameters_ex, if the provided seed is too short, |
2b0c11a6 | 810 | use a random seed, as already documented. |
1d7df236 IP |
811 | [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>] |
812 | ||
33dd0832 | 813 | Changes between 1.0.2c and 1.0.2d [9 Jul 2015] |
54ae378c | 814 | |
5627e0f7 MC |
815 | *) Alternate chains certificate forgery |
816 | ||
817 | During certificate verfification, OpenSSL will attempt to find an | |
818 | alternative certificate chain if the first attempt to build such a chain | |
819 | fails. An error in the implementation of this logic can mean that an | |
820 | attacker could cause certain checks on untrusted certificates to be | |
821 | bypassed, such as the CA flag, enabling them to use a valid leaf | |
822 | certificate to act as a CA and "issue" an invalid certificate. | |
823 | ||
824 | This issue was reported to OpenSSL by Adam Langley/David Benjamin | |
825 | (Google/BoringSSL). | |
9330fbd0 | 826 | (CVE-2015-1793) |
5627e0f7 | 827 | [Matt Caswell] |
54ae378c | 828 | |
9330fbd0 MC |
829 | *) Race condition handling PSK identify hint |
830 | ||
831 | If PSK identity hints are received by a multi-threaded client then | |
832 | the values are wrongly updated in the parent SSL_CTX structure. This can | |
833 | result in a race condition potentially leading to a double free of the | |
834 | identify hint data. | |
835 | (CVE-2015-3196) | |
836 | [Stephen Henson] | |
837 | ||
0ee5fcde | 838 | Changes between 1.0.2b and 1.0.2c [12 Jun 2015] |
b6ed9917 | 839 | |
d4c17638 MC |
840 | *) Fix HMAC ABI incompatibility. The previous version introduced an ABI |
841 | incompatibility in the handling of HMAC. The previous ABI has now been | |
842 | restored. | |
b6ed9917 | 843 | |
7b560c17 | 844 | Changes between 1.0.2a and 1.0.2b [11 Jun 2015] |
0d6d10d9 | 845 | |
ab17f6b7 MC |
846 | *) Malformed ECParameters causes infinite loop |
847 | ||
848 | When processing an ECParameters structure OpenSSL enters an infinite loop | |
849 | if the curve specified is over a specially malformed binary polynomial | |
850 | field. | |
851 | ||
852 | This can be used to perform denial of service against any | |
853 | system which processes public keys, certificate requests or | |
854 | certificates. This includes TLS clients and TLS servers with | |
855 | client authentication enabled. | |
856 | ||
857 | This issue was reported to OpenSSL by Joseph Barr-Pixton. | |
858 | (CVE-2015-1788) | |
859 | [Andy Polyakov] | |
860 | ||
861 | *) Exploitable out-of-bounds read in X509_cmp_time | |
862 | ||
863 | X509_cmp_time does not properly check the length of the ASN1_TIME | |
864 | string and can read a few bytes out of bounds. In addition, | |
865 | X509_cmp_time accepts an arbitrary number of fractional seconds in the | |
866 | time string. | |
867 | ||
868 | An attacker can use this to craft malformed certificates and CRLs of | |
869 | various sizes and potentially cause a segmentation fault, resulting in | |
870 | a DoS on applications that verify certificates or CRLs. TLS clients | |
871 | that verify CRLs are affected. TLS clients and servers with client | |
872 | authentication enabled may be affected if they use custom verification | |
873 | callbacks. | |
874 | ||
875 | This issue was reported to OpenSSL by Robert Swiecki (Google), and | |
9f0b86c6 | 876 | independently by Hanno Böck. |
ab17f6b7 | 877 | (CVE-2015-1789) |
9f0b86c6 | 878 | [Emilia Käsper] |
ab17f6b7 MC |
879 | |
880 | *) PKCS7 crash with missing EnvelopedContent | |
881 | ||
882 | The PKCS#7 parsing code does not handle missing inner EncryptedContent | |
883 | correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs | |
884 | with missing content and trigger a NULL pointer dereference on parsing. | |
885 | ||
886 | Applications that decrypt PKCS#7 data or otherwise parse PKCS#7 | |
887 | structures from untrusted sources are affected. OpenSSL clients and | |
888 | servers are not affected. | |
889 | ||
890 | This issue was reported to OpenSSL by Michal Zalewski (Google). | |
891 | (CVE-2015-1790) | |
9f0b86c6 | 892 | [Emilia Käsper] |
ab17f6b7 MC |
893 | |
894 | *) CMS verify infinite loop with unknown hash function | |
895 | ||
896 | When verifying a signedData message the CMS code can enter an infinite loop | |
897 | if presented with an unknown hash function OID. This can be used to perform | |
898 | denial of service against any system which verifies signedData messages using | |
899 | the CMS code. | |
900 | This issue was reported to OpenSSL by Johannes Bauer. | |
901 | (CVE-2015-1792) | |
902 | [Stephen Henson] | |
903 | ||
904 | *) Race condition handling NewSessionTicket | |
905 | ||
906 | If a NewSessionTicket is received by a multi-threaded client when attempting to | |
907 | reuse a previous ticket then a race condition can occur potentially leading to | |
908 | a double free of the ticket data. | |
909 | (CVE-2015-1791) | |
910 | [Matt Caswell] | |
911 | ||
595487ea MC |
912 | *) Removed support for the two export grade static DH ciphersuites |
913 | EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites | |
914 | were newly added (along with a number of other static DH ciphersuites) to | |
915 | 1.0.2. However the two export ones have *never* worked since they were | |
916 | introduced. It seems strange in any case to be adding new export | |
917 | ciphersuites, and given "logjam" it also does not seem correct to fix them. | |
918 | [Matt Caswell] | |
919 | ||
f4d1fb77 EK |
920 | *) Only support 256-bit or stronger elliptic curves with the |
921 | 'ecdh_auto' setting (server) or by default (client). Of supported | |
922 | curves, prefer P-256 (both). | |
923 | [Emilia Kasper] | |
924 | ||
10a70da7 EK |
925 | *) Reject DH handshakes with parameters shorter than 768 bits. |
926 | [Kurt Roeckx and Emilia Kasper] | |
0d6d10d9 | 927 | |
3df69d3a | 928 | Changes between 1.0.2 and 1.0.2a [19 Mar 2015] |
06aab268 | 929 | |
da947c97 MC |
930 | *) ClientHello sigalgs DoS fix |
931 | ||
932 | If a client connects to an OpenSSL 1.0.2 server and renegotiates with an | |
933 | invalid signature algorithms extension a NULL pointer dereference will | |
934 | occur. This can be exploited in a DoS attack against the server. | |
935 | ||
936 | This issue was was reported to OpenSSL by David Ramos of Stanford | |
937 | University. | |
938 | (CVE-2015-0291) | |
939 | [Stephen Henson and Matt Caswell] | |
940 | ||
941 | *) Multiblock corrupted pointer fix | |
942 | ||
943 | OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This | |
944 | feature only applies on 64 bit x86 architecture platforms that support AES | |
945 | NI instructions. A defect in the implementation of "multiblock" can cause | |
946 | OpenSSL's internal write buffer to become incorrectly set to NULL when | |
947 | using non-blocking IO. Typically, when the user application is using a | |
948 | socket BIO for writing, this will only result in a failed connection. | |
949 | However if some other BIO is used then it is likely that a segmentation | |
950 | fault will be triggered, thus enabling a potential DoS attack. | |
951 | ||
952 | This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller. | |
953 | (CVE-2015-0290) | |
954 | [Matt Caswell] | |
955 | ||
956 | *) Segmentation fault in DTLSv1_listen fix | |
957 | ||
958 | The DTLSv1_listen function is intended to be stateless and processes the | |
959 | initial ClientHello from many peers. It is common for user code to loop | |
960 | over the call to DTLSv1_listen until a valid ClientHello is received with | |
961 | an associated cookie. A defect in the implementation of DTLSv1_listen means | |
962 | that state is preserved in the SSL object from one invocation to the next | |
963 | that can lead to a segmentation fault. Errors processing the initial | |
964 | ClientHello can trigger this scenario. An example of such an error could be | |
965 | that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only | |
966 | server. | |
967 | ||
968 | This issue was reported to OpenSSL by Per Allansson. | |
969 | (CVE-2015-0207) | |
970 | [Matt Caswell] | |
971 | ||
972 | *) Segmentation fault in ASN1_TYPE_cmp fix | |
973 | ||
974 | The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is | |
975 | made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check | |
976 | certificate signature algorithm consistency this can be used to crash any | |
977 | certificate verification operation and exploited in a DoS attack. Any | |
978 | application which performs certificate verification is vulnerable including | |
979 | OpenSSL clients and servers which enable client authentication. | |
980 | (CVE-2015-0286) | |
981 | [Stephen Henson] | |
982 | ||
983 | *) Segmentation fault for invalid PSS parameters fix | |
984 | ||
985 | The signature verification routines will crash with a NULL pointer | |
986 | dereference if presented with an ASN.1 signature using the RSA PSS | |
987 | algorithm and invalid parameters. Since these routines are used to verify | |
988 | certificate signature algorithms this can be used to crash any | |
989 | certificate verification operation and exploited in a DoS attack. Any | |
990 | application which performs certificate verification is vulnerable including | |
991 | OpenSSL clients and servers which enable client authentication. | |
992 | ||
993 | This issue was was reported to OpenSSL by Brian Carpenter. | |
994 | (CVE-2015-0208) | |
995 | [Stephen Henson] | |
996 | ||
997 | *) ASN.1 structure reuse memory corruption fix | |
998 | ||
999 | Reusing a structure in ASN.1 parsing may allow an attacker to cause | |
1000 | memory corruption via an invalid write. Such reuse is and has been | |
1001 | strongly discouraged and is believed to be rare. | |
1002 | ||
1003 | Applications that parse structures containing CHOICE or ANY DEFINED BY | |
1004 | components may be affected. Certificate parsing (d2i_X509 and related | |
1005 | functions) are however not affected. OpenSSL clients and servers are | |
1006 | not affected. | |
1007 | (CVE-2015-0287) | |
1008 | [Stephen Henson] | |
1009 | ||
1010 | *) PKCS7 NULL pointer dereferences fix | |
1011 | ||
1012 | The PKCS#7 parsing code does not handle missing outer ContentInfo | |
1013 | correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with | |
1014 | missing content and trigger a NULL pointer dereference on parsing. | |
1015 | ||
1016 | Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or | |
1017 | otherwise parse PKCS#7 structures from untrusted sources are | |
1018 | affected. OpenSSL clients and servers are not affected. | |
1019 | ||
1020 | This issue was reported to OpenSSL by Michal Zalewski (Google). | |
1021 | (CVE-2015-0289) | |
9f0b86c6 | 1022 | [Emilia Käsper] |
da947c97 MC |
1023 | |
1024 | *) DoS via reachable assert in SSLv2 servers fix | |
1025 | ||
1026 | A malicious client can trigger an OPENSSL_assert (i.e., an abort) in | |
1027 | servers that both support SSLv2 and enable export cipher suites by sending | |
1028 | a specially crafted SSLv2 CLIENT-MASTER-KEY message. | |
1029 | ||
9f0b86c6 | 1030 | This issue was discovered by Sean Burford (Google) and Emilia Käsper |
da947c97 MC |
1031 | (OpenSSL development team). |
1032 | (CVE-2015-0293) | |
9f0b86c6 | 1033 | [Emilia Käsper] |
da947c97 MC |
1034 | |
1035 | *) Empty CKE with client auth and DHE fix | |
1036 | ||
1037 | If client auth is used then a server can seg fault in the event of a DHE | |
1038 | ciphersuite being selected and a zero length ClientKeyExchange message | |
1039 | being sent by the client. This could be exploited in a DoS attack. | |
1040 | (CVE-2015-1787) | |
1041 | [Matt Caswell] | |
1042 | ||
1043 | *) Handshake with unseeded PRNG fix | |
1044 | ||
1045 | Under certain conditions an OpenSSL 1.0.2 client can complete a handshake | |
1046 | with an unseeded PRNG. The conditions are: | |
1047 | - The client is on a platform where the PRNG has not been seeded | |
1048 | automatically, and the user has not seeded manually | |
1049 | - A protocol specific client method version has been used (i.e. not | |
1050 | SSL_client_methodv23) | |
1051 | - A ciphersuite is used that does not require additional random data from | |
1052 | the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA). | |
1053 | ||
1054 | If the handshake succeeds then the client random that has been used will | |
1055 | have been generated from a PRNG with insufficient entropy and therefore the | |
1056 | output may be predictable. | |
1057 | ||
1058 | For example using the following command with an unseeded openssl will | |
1059 | succeed on an unpatched platform: | |
1060 | ||
1061 | openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA | |
1062 | (CVE-2015-0285) | |
1063 | [Matt Caswell] | |
1064 | ||
1065 | *) Use After Free following d2i_ECPrivatekey error fix | |
1066 | ||
1067 | A malformed EC private key file consumed via the d2i_ECPrivateKey function | |
1068 | could cause a use after free condition. This, in turn, could cause a double | |
1069 | free in several private key parsing functions (such as d2i_PrivateKey | |
1070 | or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption | |
1071 | for applications that receive EC private keys from untrusted | |
1072 | sources. This scenario is considered rare. | |
1073 | ||
1074 | This issue was discovered by the BoringSSL project and fixed in their | |
1075 | commit 517073cd4b. | |
1076 | (CVE-2015-0209) | |
1077 | [Matt Caswell] | |
1078 | ||
1079 | *) X509_to_X509_REQ NULL pointer deref fix | |
1080 | ||
1081 | The function X509_to_X509_REQ will crash with a NULL pointer dereference if | |
1082 | the certificate key is invalid. This function is rarely used in practice. | |
1083 | ||
1084 | This issue was discovered by Brian Carpenter. | |
1085 | (CVE-2015-0288) | |
1086 | [Stephen Henson] | |
1087 | ||
f417997a KR |
1088 | *) Removed the export ciphers from the DEFAULT ciphers |
1089 | [Kurt Roeckx] | |
06aab268 | 1090 | |
4ac03295 | 1091 | Changes between 1.0.1l and 1.0.2 [22 Jan 2015] |
0a9f7780 | 1092 | |
cc42e4af EK |
1093 | *) Change RSA and DH/DSA key generation apps to generate 2048-bit |
1094 | keys by default. | |
1095 | [Kurt Roeckx] | |
1096 | ||
1cfd7cf3 AP |
1097 | *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g. |
1098 | ARMv5 through ARMv8, as opposite to "locking" it to single one. | |
1099 | So far those who have to target multiple plaforms would compromise | |
1100 | and argue that binary targeting say ARMv5 would still execute on | |
1101 | ARMv8. "Universal" build resolves this compromise by providing | |
1102 | near-optimal performance even on newer platforms. | |
1103 | [Andy Polyakov] | |
1104 | ||
d2a1226b AP |
1105 | *) Accelerated NIST P-256 elliptic curve implementation for x86_64 |
1106 | (other platforms pending). | |
0ce2dbfb | 1107 | [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov] |
d2a1226b | 1108 | |
2102c53c DSH |
1109 | *) Add support for the SignedCertificateTimestampList certificate and |
1110 | OCSP response extensions from RFC6962. | |
1111 | [Rob Stradling] | |
1112 | ||
d5213519 BM |
1113 | *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) |
1114 | for corner cases. (Certain input points at infinity could lead to | |
1115 | bogus results, with non-infinity inputs mapped to infinity too.) | |
1116 | [Bodo Moeller] | |
1117 | ||
0ae6ba18 AP |
1118 | *) Initial support for PowerISA 2.0.7, first implemented in POWER8. |
1119 | This covers AES, SHA256/512 and GHASH. "Initial" means that most | |
1120 | common cases are optimized and there still is room for further | |
1121 | improvements. Vector Permutation AES for Altivec is also added. | |
1122 | [Andy Polyakov] | |
1123 | ||
1124 | *) Add support for little-endian ppc64 Linux target. | |
1125 | [Marcelo Cerri (IBM)] | |
1126 | ||
1127 | *) Initial support for AMRv8 ISA crypto extensions. This covers AES, | |
1128 | SHA1, SHA256 and GHASH. "Initial" means that most common cases | |
1129 | are optimized and there still is room for further improvements. | |
1130 | Both 32- and 64-bit modes are supported. | |
1131 | [Andy Polyakov, Ard Biesheuvel (Linaro)] | |
1132 | ||
1133 | *) Improved ARMv7 NEON support. | |
1134 | [Andy Polyakov] | |
1135 | ||
1136 | *) Support for SPARC Architecture 2011 crypto extensions, first | |
1137 | implemented in SPARC T4. This covers AES, DES, Camellia, SHA1, | |
1138 | SHA256/512, MD5, GHASH and modular exponentiation. | |
1139 | [Andy Polyakov, David Miller] | |
1140 | ||
1141 | *) Accelerated modular exponentiation for Intel processors, a.k.a. | |
1142 | RSAZ. | |
0ce2dbfb | 1143 | [Shay Gueron & Vlad Krasnov (Intel Corp)] |
0ae6ba18 AP |
1144 | |
1145 | *) Support for new and upcoming Intel processors, including AVX2, | |
1146 | BMI and SHA ISA extensions. This includes additional "stitched" | |
1147 | implementations, AESNI-SHA256 and GCM, and multi-buffer support | |
1148 | for TLS encrypt. | |
1149 | ||
1150 | This work was sponsored by Intel Corp. | |
1151 | [Andy Polyakov] | |
1152 | ||
c578fe37 BM |
1153 | *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method() |
1154 | supports both DTLS 1.2 and 1.0 and should use whatever version the peer | |
1155 | supports and DTLSv1_2_*_method() which supports DTLS 1.2 only. | |
1156 | [Steve Henson] | |
1157 | ||
b9fa413a DSH |
1158 | *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file(): |
1159 | this fixes a limiation in previous versions of OpenSSL. | |
1160 | [Steve Henson] | |
1161 | ||
25f93585 DSH |
1162 | *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest, |
1163 | MGF1 digest and OAEP label. | |
1164 | [Steve Henson] | |
1165 | ||
c6f33865 DSH |
1166 | *) Add EVP support for key wrapping algorithms, to avoid problems with |
1167 | existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in | |
1168 | the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap | |
1169 | algorithms and include tests cases. | |
1170 | [Steve Henson] | |
1171 | ||
7c23127f DSH |
1172 | *) Add functions to allocate and set the fields of an ECDSA_METHOD |
1173 | structure. | |
1174 | [Douglas E. Engert, Steve Henson] | |
1175 | ||
904348a4 DSH |
1176 | *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the |
1177 | difference in days and seconds between two tm or ASN1_TIME structures. | |
1178 | [Steve Henson] | |
1179 | ||
171c4da5 DSH |
1180 | *) Add -rev test option to s_server to just reverse order of characters |
1181 | received by client and send back to server. Also prints an abbreviated | |
1182 | summary of the connection parameters. | |
1183 | [Steve Henson] | |
1184 | ||
04611fb0 DSH |
1185 | *) New option -brief for s_client and s_server to print out a brief summary |
1186 | of connection parameters. | |
1187 | [Steve Henson] | |
1188 | ||
e27711cf T |
1189 | *) Add callbacks for arbitrary TLS extensions. |
1190 | [Trevor Perrin <trevp@trevp.net> and Ben Laurie] | |
1191 | ||
57912ed3 DSH |
1192 | *) New option -crl_download in several openssl utilities to download CRLs |
1193 | from CRLDP extension in certificates. | |
1194 | [Steve Henson] | |
1195 | ||
e318431e DSH |
1196 | *) New options -CRL and -CRLform for s_client and s_server for CRLs. |
1197 | [Steve Henson] | |
1198 | ||
6a10f38d DSH |
1199 | *) New function X509_CRL_diff to generate a delta CRL from the difference |
1200 | of two full CRLs. Add support to "crl" utility. | |
1201 | [Steve Henson] | |
1202 | ||
75f53531 DSH |
1203 | *) New functions to set lookup_crls function and to retrieve |
1204 | X509_STORE from X509_STORE_CTX. | |
1205 | [Steve Henson] | |
1206 | ||
2aa3ef78 DSH |
1207 | *) Print out deprecated issuer and subject unique ID fields in |
1208 | certificates. | |
1209 | [Steve Henson] | |
1210 | ||
5c8d41be DSH |
1211 | *) Extend OCSP I/O functions so they can be used for simple general purpose |
1212 | HTTP as well as OCSP. New wrapper function which can be used to download | |
1213 | CRLs using the OCSP API. | |
1214 | [Steve Henson] | |
1215 | ||
15387e4c DSH |
1216 | *) Delegate command line handling in s_client/s_server to SSL_CONF APIs. |
1217 | [Steve Henson] | |
1218 | ||
49ef33fa DSH |
1219 | *) SSL_CONF* functions. These provide a common framework for application |
1220 | configuration using configuration files or command lines. | |
1221 | [Steve Henson] | |
1222 | ||
bc200e69 DSH |
1223 | *) SSL/TLS tracing code. This parses out SSL/TLS records using the |
1224 | message callback and prints the results. Needs compile time option | |
1225 | "enable-ssl-trace". New options to s_client and s_server to enable | |
1226 | tracing. | |
1227 | [Steve Henson] | |
1228 | ||
78b5d89d | 1229 | *) New ctrl and macro to retrieve supported points extensions. |
1b9a59c3 | 1230 | Print out extension in s_server and s_client. |
78b5d89d DSH |
1231 | [Steve Henson] |
1232 | ||
bd9fc1d6 DSH |
1233 | *) New functions to retrieve certificate signature and signature |
1234 | OID NID. | |
1235 | [Steve Henson] | |
1236 | ||
1520e6c0 DSH |
1237 | *) Add functions to retrieve and manipulate the raw cipherlist sent by a |
1238 | client to OpenSSL. | |
1239 | [Steve Henson] | |
1240 | ||
ccf6a19e DSH |
1241 | *) New Suite B modes for TLS code. These use and enforce the requirements |
1242 | of RFC6460: restrict ciphersuites, only permit Suite B algorithms and | |
1243 | only use Suite B curves. The Suite B modes can be set by using the | |
1244 | strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring. | |
1245 | [Steve Henson] | |
1246 | ||
ba8bdea7 DSH |
1247 | *) New chain verification flags for Suite B levels of security. Check |
1248 | algorithms are acceptable when flags are set in X509_verify_cert. | |
1249 | [Steve Henson] | |
1250 | ||
6660baee DSH |
1251 | *) Make tls1_check_chain return a set of flags indicating checks passed |
1252 | by a certificate chain. Add additional tests to handle client | |
1253 | certificates: checks for matching certificate type and issuer name | |
1254 | comparison. | |
1255 | [Steve Henson] | |
1256 | ||
25d4c925 DSH |
1257 | *) If an attempt is made to use a signature algorithm not in the peer |
1258 | preference list abort the handshake. If client has no suitable | |
1259 | signature algorithms in response to a certificate request do not | |
1260 | use the certificate. | |
1261 | [Steve Henson] | |
1262 | ||
44adfeb6 DSH |
1263 | *) If server EC tmp key is not in client preference list abort handshake. |
1264 | [Steve Henson] | |
1265 | ||
b762acad DSH |
1266 | *) Add support for certificate stores in CERT structure. This makes it |
1267 | possible to have different stores per SSL structure or one store in | |
1268 | the parent SSL_CTX. Include distint stores for certificate chain | |
1269 | verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN | |
1270 | to build and store a certificate chain in CERT structure: returing | |
1271 | an error if the chain cannot be built: this will allow applications | |
1272 | to test if a chain is correctly configured. | |
1273 | ||
1274 | Note: if the CERT based stores are not set then the parent SSL_CTX | |
1275 | store is used to retain compatibility with existing behaviour. | |
1276 | ||
1277 | [Steve Henson] | |
1278 | ||
b28fbdfa DSH |
1279 | *) New function ssl_set_client_disabled to set a ciphersuite disabled |
1280 | mask based on the current session, check mask when sending client | |
1281 | hello and checking the requested ciphersuite. | |
1282 | [Steve Henson] | |
1283 | ||
a897502c DSH |
1284 | *) New ctrls to retrieve and set certificate types in a certificate |
1285 | request message. Print out received values in s_client. If certificate | |
1286 | types is not set with custom values set sensible values based on | |
1287 | supported signature algorithms. | |
1288 | [Steve Henson] | |
1289 | ||
04c32cdd DSH |
1290 | *) Support for distinct client and server supported signature algorithms. |
1291 | [Steve Henson] | |
1292 | ||
623a5e24 DSH |
1293 | *) Add certificate callback. If set this is called whenever a certificate |
1294 | is required by client or server. An application can decide which | |
1295 | certificate chain to present based on arbitrary criteria: for example | |
1296 | supported signature algorithms. Add very simple example to s_server. | |
1297 | This fixes many of the problems and restrictions of the existing client | |
1298 | certificate callback: for example you can now clear an existing | |
1299 | certificate and specify the whole chain. | |
1300 | [Steve Henson] | |
1301 | ||
484f8762 DSH |
1302 | *) Add new "valid_flags" field to CERT_PKEY structure which determines what |
1303 | the certificate can be used for (if anything). Set valid_flags field | |
1304 | in new tls1_check_chain function. Simplify ssl_set_cert_masks which used | |
1305 | to have similar checks in it. | |
1306 | ||
1307 | Add new "cert_flags" field to CERT structure and include a "strict mode". | |
1308 | This enforces some TLS certificate requirements (such as only permitting | |
1309 | certificate signature algorithms contained in the supported algorithms | |
1310 | extension) which some implementations ignore: this option should be used | |
1311 | with caution as it could cause interoperability issues. | |
1312 | [Steve Henson] | |
1313 | ||
c70a1fee DSH |
1314 | *) Update and tidy signature algorithm extension processing. Work out |
1315 | shared signature algorithms based on preferences and peer algorithms | |
1316 | and print them out in s_client and s_server. Abort handshake if no | |
1317 | shared signature algorithms. | |
1318 | [Steve Henson] | |
1319 | ||
0b362de5 DSH |
1320 | *) Add new functions to allow customised supported signature algorithms |
1321 | for SSL and SSL_CTX structures. Add options to s_client and s_server | |
1322 | to support them. | |
1323 | [Steve Henson] | |
1324 | ||
d312f7be DSH |
1325 | *) New function SSL_certs_clear() to delete all references to certificates |
1326 | from an SSL structure. Before this once a certificate had been added | |
1327 | it couldn't be removed. | |
1328 | [Steve Henson] | |
1329 | ||
70cd3c6b DSH |
1330 | *) Integrate hostname, email address and IP address checking with certificate |
1331 | verification. New verify options supporting checking in opensl utility. | |
1332 | [Steve Henson] | |
1333 | ||
45da1efc DSH |
1334 | *) Fixes and wildcard matching support to hostname and email checking |
1335 | functions. Add manual page. | |
1336 | [Florian Weimer (Red Hat Product Security Team)] | |
1337 | ||
1338 | *) New functions to check a hostname email or IP address against a | |
1339 | certificate. Add options x509 utility to print results of checks against | |
1340 | a certificate. | |
1341 | [Steve Henson] | |
1342 | ||
d65b8b21 BL |
1343 | *) Fix OCSP checking. |
1344 | [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie] | |
1345 | ||
8c149cfd BM |
1346 | *) Initial experimental support for explicitly trusted non-root CAs. |
1347 | OpenSSL still tries to build a complete chain to a root but if an | |
1348 | intermediate CA has a trust setting included that is used. The first | |
1349 | setting is used: whether to trust (e.g., -addtrust option to the x509 | |
1350 | utility) or reject. | |
1351 | [Steve Henson] | |
d65b8b21 | 1352 | |
9d2006d8 DSH |
1353 | *) Add -trusted_first option which attempts to find certificates in the |
1354 | trusted store even if an untrusted chain is also supplied. | |
1355 | [Steve Henson] | |
1356 | ||
988037fe AP |
1357 | *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE, |
1358 | platform support for Linux and Android. | |
1359 | [Andy Polyakov] | |
1360 | ||
0e05b51f AP |
1361 | *) Support for linux-x32, ILP32 environment in x86_64 framework. |
1362 | [Andy Polyakov] | |
1363 | ||
1dded7f7 DSH |
1364 | *) Experimental multi-implementation support for FIPS capable OpenSSL. |
1365 | When in FIPS mode the approved implementations are used as normal, | |
1366 | when not in FIPS mode the internal unapproved versions are used instead. | |
1367 | This means that the FIPS capable OpenSSL isn't forced to use the | |
1368 | (often lower perfomance) FIPS implementations outside FIPS mode. | |
1369 | [Steve Henson] | |
1370 | ||
c3cb0691 DSH |
1371 | *) Transparently support X9.42 DH parameters when calling |
1372 | PEM_read_bio_DHparameters. This means existing applications can handle | |
1373 | the new parameter format automatically. | |
1374 | [Steve Henson] | |
1375 | ||
491734eb DSH |
1376 | *) Initial experimental support for X9.42 DH parameter format: mainly |
1377 | to support use of 'q' parameter for RFC5114 parameters. | |
1378 | [Steve Henson] | |
1379 | ||
e811eff5 DSH |
1380 | *) Add DH parameters from RFC5114 including test data to dhtest. |
1381 | [Steve Henson] | |
1382 | ||
e46c807e DSH |
1383 | *) Support for automatic EC temporary key parameter selection. If enabled |
1384 | the most preferred EC parameters are automatically used instead of | |
1385 | hardcoded fixed parameters. Now a server just has to call: | |
1386 | SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically | |
1387 | support ECDH and use the most appropriate parameters. | |
1388 | [Steve Henson] | |
1389 | ||
6b870763 DSH |
1390 | *) Enhance and tidy EC curve and point format TLS extension code. Use |
1391 | static structures instead of allocation if default values are used. | |
1392 | New ctrls to set curves we wish to support and to retrieve shared curves. | |
1393 | Print out shared curves in s_server. New options to s_server and s_client | |
1394 | to set list of supported curves. | |
1395 | [Steve Henson] | |
1396 | ||
55058181 DSH |
1397 | *) New ctrls to retrieve supported signature algorithms and |
1398 | supported curve values as an array of NIDs. Extend openssl utility | |
1399 | to print out received values. | |
1400 | [Steve Henson] | |
1401 | ||
a068a1d0 DSH |
1402 | *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert |
1403 | between NIDs and the more common NIST names such as "P-256". Enhance | |
1404 | ecparam utility and ECC method to recognise the NIST names for curves. | |
1405 | [Steve Henson] | |
1406 | ||
37b16c84 DSH |
1407 | *) Enhance SSL/TLS certificate chain handling to support different |
1408 | chains for each certificate instead of one chain in the parent SSL_CTX. | |
1409 | [Steve Henson] | |
1410 | ||
c523eb98 DSH |
1411 | *) Support for fixed DH ciphersuite client authentication: where both |
1412 | server and client use DH certificates with common parameters. | |
1413 | [Steve Henson] | |
1414 | ||
0ffa4997 DSH |
1415 | *) Support for fixed DH ciphersuites: those requiring DH server |
1416 | certificates. | |
1417 | [Steve Henson] | |
b9115239 | 1418 | |
e9128d94 EK |
1419 | *) New function i2d_re_X509_tbs for re-encoding the TBS portion of |
1420 | the certificate. | |
1421 | Note: Related 1.0.2-beta specific macros X509_get_cert_info, | |
1422 | X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and | |
1423 | X509_CINF_get_signature were reverted post internal team review. | |
1424 | ||
ba7e998d MC |
1425 | Changes between 1.0.1k and 1.0.1l [15 Jan 2015] |
1426 | ||
1427 | *) Build fixes for the Windows and OpenVMS platforms | |
1428 | [Matt Caswell and Richard Levitte] | |
1429 | ||
1430 | Changes between 1.0.1j and 1.0.1k [8 Jan 2015] | |
4c75f4e5 | 1431 | |
a8b1e52f MC |
1432 | *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS |
1433 | message can cause a segmentation fault in OpenSSL due to a NULL pointer | |
1434 | dereference. This could lead to a Denial Of Service attack. Thanks to | |
1435 | Markus Stenberg of Cisco Systems, Inc. for reporting this issue. | |
1436 | (CVE-2014-3571) | |
1437 | [Steve Henson] | |
1438 | ||
1439 | *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the | |
1440 | dtls1_buffer_record function under certain conditions. In particular this | |
1441 | could occur if an attacker sent repeated DTLS records with the same | |
1442 | sequence number but for the next epoch. The memory leak could be exploited | |
1443 | by an attacker in a Denial of Service attack through memory exhaustion. | |
1444 | Thanks to Chris Mueller for reporting this issue. | |
1445 | (CVE-2015-0206) | |
1446 | [Matt Caswell] | |
1447 | ||
1448 | *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is | |
1449 | built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl | |
1450 | method would be set to NULL which could later result in a NULL pointer | |
1451 | dereference. Thanks to Frank Schmirler for reporting this issue. | |
1452 | (CVE-2014-3569) | |
1453 | [Kurt Roeckx] | |
1454 | ||
4aaf1e49 DSH |
1455 | *) Abort handshake if server key exchange message is omitted for ephemeral |
1456 | ECDH ciphersuites. | |
1457 | ||
a936ba11 DSH |
1458 | Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for |
1459 | reporting this issue. | |
4aaf1e49 DSH |
1460 | (CVE-2014-3572) |
1461 | [Steve Henson] | |
1462 | ||
4b4c1fcc DSH |
1463 | *) Remove non-export ephemeral RSA code on client and server. This code |
1464 | violated the TLS standard by allowing the use of temporary RSA keys in | |
1465 | non-export ciphersuites and could be used by a server to effectively | |
1466 | downgrade the RSA key length used to a value smaller than the server | |
a936ba11 DSH |
1467 | certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at |
1468 | INRIA or reporting this issue. | |
4b4c1fcc DSH |
1469 | (CVE-2015-0204) |
1470 | [Steve Henson] | |
1471 | ||
a8b1e52f MC |
1472 | *) Fixed issue where DH client certificates are accepted without verification. |
1473 | An OpenSSL server will accept a DH certificate for client authentication | |
1474 | without the certificate verify message. This effectively allows a client to | |
1475 | authenticate without the use of a private key. This only affects servers | |
1476 | which trust a client certificate authority which issues certificates | |
1477 | containing DH keys: these are extremely rare and hardly ever encountered. | |
1478 | Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting | |
1479 | this issue. | |
1480 | (CVE-2015-0205) | |
1481 | [Steve Henson] | |
1482 | ||
d9b277e0 AL |
1483 | *) Ensure that the session ID context of an SSL is updated when its |
1484 | SSL_CTX is updated via SSL_set_SSL_CTX. | |
1485 | ||
1486 | The session ID context is typically set from the parent SSL_CTX, | |
1487 | and can vary with the CTX. | |
1488 | [Adam Langley] | |
1489 | ||
85cfc188 DSH |
1490 | *) Fix various certificate fingerprint issues. |
1491 | ||
1492 | By using non-DER or invalid encodings outside the signed portion of a | |
1493 | certificate the fingerprint can be changed without breaking the signature. | |
1494 | Although no details of the signed portion of the certificate can be changed | |
1495 | this can cause problems with some applications: e.g. those using the | |
1496 | certificate fingerprint for blacklists. | |
1497 | ||
1498 | 1. Reject signatures with non zero unused bits. | |
1499 | ||
1500 | If the BIT STRING containing the signature has non zero unused bits reject | |
1501 | the signature. All current signature algorithms require zero unused bits. | |
1502 | ||
1503 | 2. Check certificate algorithm consistency. | |
1504 | ||
1505 | Check the AlgorithmIdentifier inside TBS matches the one in the | |
1506 | certificate signature. NB: this will result in signature failure | |
1507 | errors for some broken certificates. | |
1508 | ||
1509 | Thanks to Konrad Kraszewski from Google for reporting this issue. | |
1510 | ||
1511 | 3. Check DSA/ECDSA signatures use DER. | |
1512 | ||
1513 | Reencode DSA/ECDSA signatures and compare with the original received | |
1514 | signature. Return an error if there is a mismatch. | |
1515 | ||
1516 | This will reject various cases including garbage after signature | |
1517 | (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS | |
1518 | program for discovering this case) and use of BER or invalid ASN.1 INTEGERs | |
1519 | (negative or with leading zeroes). | |
1520 | ||
1521 | Further analysis was conducted and fixes were developed by Stephen Henson | |
1522 | of the OpenSSL core team. | |
1523 | ||
1524 | (CVE-2014-8275) | |
1525 | [Steve Henson] | |
1526 | ||
a8b1e52f MC |
1527 | *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect |
1528 | results on some platforms, including x86_64. This bug occurs at random | |
1529 | with a very low probability, and is not known to be exploitable in any | |
1530 | way, though its exact impact is difficult to determine. Thanks to Pieter | |
1531 | Wuille (Blockstream) who reported this issue and also suggested an initial | |
1532 | fix. Further analysis was conducted by the OpenSSL development team and | |
1533 | Adam Langley of Google. The final fix was developed by Andy Polyakov of | |
1534 | the OpenSSL core team. | |
1535 | (CVE-2014-3570) | |
1536 | [Andy Polyakov] | |
1537 | ||
03d14f58 DB |
1538 | *) Do not resume sessions on the server if the negotiated protocol |
1539 | version does not match the session's version. Resuming with a different | |
1540 | version, while not strictly forbidden by the RFC, is of questionable | |
1541 | sanity and breaks all known clients. | |
9f0b86c6 | 1542 | [David Benjamin, Emilia Käsper] |
03d14f58 | 1543 | |
e5f261df EK |
1544 | *) Tighten handling of the ChangeCipherSpec (CCS) message: reject |
1545 | early CCS messages during renegotiation. (Note that because | |
1546 | renegotiation is encrypted, this early CCS was not exploitable.) | |
9f0b86c6 | 1547 | [Emilia Käsper] |
e5f261df | 1548 | |
4c75f4e5 EK |
1549 | *) Tighten client-side session ticket handling during renegotiation: |
1550 | ensure that the client only accepts a session ticket if the server sends | |
1551 | the extension anew in the ServerHello. Previously, a TLS client would | |
1552 | reuse the old extension state and thus accept a session ticket if one was | |
1553 | announced in the initial ServerHello. | |
9baee021 EK |
1554 | |
1555 | Similarly, ensure that the client requires a session ticket if one | |
1556 | was advertised in the ServerHello. Previously, a TLS client would | |
1557 | ignore a missing NewSessionTicket message. | |
9f0b86c6 | 1558 | [Emilia Käsper] |
4c75f4e5 | 1559 | |
13803174 EK |
1560 | Changes between 1.0.1i and 1.0.1j [15 Oct 2014] |
1561 | ||
1562 | *) SRTP Memory Leak. | |
1563 | ||
1564 | A flaw in the DTLS SRTP extension parsing code allows an attacker, who | |
1565 | sends a carefully crafted handshake message, to cause OpenSSL to fail | |
1566 | to free up to 64k of memory causing a memory leak. This could be | |
1567 | exploited in a Denial Of Service attack. This issue affects OpenSSL | |
1568 | 1.0.1 server implementations for both SSL/TLS and DTLS regardless of | |
1569 | whether SRTP is used or configured. Implementations of OpenSSL that | |
1570 | have been compiled with OPENSSL_NO_SRTP defined are not affected. | |
1571 | ||
1572 | The fix was developed by the OpenSSL team. | |
1573 | (CVE-2014-3513) | |
1574 | [OpenSSL team] | |
1575 | ||
1576 | *) Session Ticket Memory Leak. | |
1577 | ||
1578 | When an OpenSSL SSL/TLS/DTLS server receives a session ticket the | |
1579 | integrity of that ticket is first verified. In the event of a session | |
1580 | ticket integrity check failing, OpenSSL will fail to free memory | |
1581 | causing a memory leak. By sending a large number of invalid session | |
1582 | tickets an attacker could exploit this issue in a Denial Of Service | |
1583 | attack. | |
1584 | (CVE-2014-3567) | |
1585 | [Steve Henson] | |
1586 | ||
1587 | *) Build option no-ssl3 is incomplete. | |
1588 | ||
1589 | When OpenSSL is configured with "no-ssl3" as a build option, servers | |
1590 | could accept and complete a SSL 3.0 handshake, and clients could be | |
1591 | configured to send them. | |
1592 | (CVE-2014-3568) | |
1593 | [Akamai and the OpenSSL team] | |
1594 | ||
1595 | *) Add support for TLS_FALLBACK_SCSV. | |
1596 | Client applications doing fallback retries should call | |
1597 | SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV). | |
1598 | (CVE-2014-3566) | |
1599 | [Adam Langley, Bodo Moeller] | |
1600 | ||
1601 | *) Add additional DigestInfo checks. | |
1602 | ||
1603 | Reencode DigestInto in DER and check against the original when | |
1604 | verifying RSA signature: this will reject any improperly encoded | |
1605 | DigestInfo structures. | |
1606 | ||
1607 | Note: this is a precautionary measure and no attacks are currently known. | |
1608 | ||
1609 | [Steve Henson] | |
1610 | ||
5e60396f MC |
1611 | Changes between 1.0.1h and 1.0.1i [6 Aug 2014] |
1612 | ||
1613 | *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the | |
1614 | SRP code can be overrun an internal buffer. Add sanity check that | |
1615 | g, A, B < N to SRP code. | |
1616 | ||
1617 | Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC | |
1618 | Group for discovering this issue. | |
1619 | (CVE-2014-3512) | |
1620 | [Steve Henson] | |
1621 | ||
1622 | *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate | |
1623 | TLS 1.0 instead of higher protocol versions when the ClientHello message | |
1624 | is badly fragmented. This allows a man-in-the-middle attacker to force a | |
1625 | downgrade to TLS 1.0 even if both the server and the client support a | |
1626 | higher protocol version, by modifying the client's TLS records. | |
1627 | ||
1628 | Thanks to David Benjamin and Adam Langley (Google) for discovering and | |
1629 | researching this issue. | |
1630 | (CVE-2014-3511) | |
1631 | [David Benjamin] | |
1632 | ||
1633 | *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject | |
1634 | to a denial of service attack. A malicious server can crash the client | |
1635 | with a null pointer dereference (read) by specifying an anonymous (EC)DH | |
1636 | ciphersuite and sending carefully crafted handshake messages. | |
1637 | ||
9f0b86c6 | 1638 | Thanks to Felix Gröbert (Google) for discovering and researching this |
5e60396f MC |
1639 | issue. |
1640 | (CVE-2014-3510) | |
9f0b86c6 | 1641 | [Emilia Käsper] |
5e60396f MC |
1642 | |
1643 | *) By sending carefully crafted DTLS packets an attacker could cause openssl | |
1644 | to leak memory. This can be exploited through a Denial of Service attack. | |
1645 | Thanks to Adam Langley for discovering and researching this issue. | |
1646 | (CVE-2014-3507) | |
1647 | [Adam Langley] | |
1648 | ||
1649 | *) An attacker can force openssl to consume large amounts of memory whilst | |
1650 | processing DTLS handshake messages. This can be exploited through a | |
1651 | Denial of Service attack. | |
1652 | Thanks to Adam Langley for discovering and researching this issue. | |
1653 | (CVE-2014-3506) | |
1654 | [Adam Langley] | |
1655 | ||
1656 | *) An attacker can force an error condition which causes openssl to crash | |
1657 | whilst processing DTLS packets due to memory being freed twice. This | |
1658 | can be exploited through a Denial of Service attack. | |
1659 | Thanks to Adam Langley and Wan-Teh Chang for discovering and researching | |
1660 | this issue. | |
1661 | (CVE-2014-3505) | |
1662 | [Adam Langley] | |
1663 | ||
1664 | *) If a multithreaded client connects to a malicious server using a resumed | |
1665 | session and the server sends an ec point format extension it could write | |
1666 | up to 255 bytes to freed memory. | |
1667 | ||
1668 | Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this | |
1669 | issue. | |
1670 | (CVE-2014-3509) | |
1671 | [Gabor Tyukasz] | |
1672 | ||
1673 | *) A malicious server can crash an OpenSSL client with a null pointer | |
1674 | dereference (read) by specifying an SRP ciphersuite even though it was not | |
1675 | properly negotiated with the client. This can be exploited through a | |
1676 | Denial of Service attack. | |
1677 | ||
9f0b86c6 | 1678 | Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for |
5e60396f MC |
1679 | discovering and researching this issue. |
1680 | (CVE-2014-5139) | |
1681 | [Steve Henson] | |
1682 | ||
1683 | *) A flaw in OBJ_obj2txt may cause pretty printing functions such as | |
1684 | X509_name_oneline, X509_name_print_ex et al. to leak some information | |
1685 | from the stack. Applications may be affected if they echo pretty printing | |
1686 | output to the attacker. | |
1687 | ||
1688 | Thanks to Ivan Fratric (Google) for discovering this issue. | |
1689 | (CVE-2014-3508) | |
9f0b86c6 | 1690 | [Emilia Käsper, and Steve Henson] |
5e60396f MC |
1691 | |
1692 | *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.) | |
1693 | for corner cases. (Certain input points at infinity could lead to | |
1694 | bogus results, with non-infinity inputs mapped to infinity too.) | |
1695 | [Bodo Moeller] | |
1696 | ||
68a1e0bc RL |
1697 | Changes between 1.0.1g and 1.0.1h [5 Jun 2014] |
1698 | ||
1699 | *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted | |
1700 | handshake can force the use of weak keying material in OpenSSL | |
1701 | SSL/TLS clients and servers. | |
1702 | ||
1703 | Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and | |
1704 | researching this issue. (CVE-2014-0224) | |
1705 | [KIKUCHI Masashi, Steve Henson] | |
1706 | ||
1707 | *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an | |
1708 | OpenSSL DTLS client the code can be made to recurse eventually crashing | |
1709 | in a DoS attack. | |
1710 | ||
1711 | Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue. | |
1712 | (CVE-2014-0221) | |
1713 | [Imre Rad, Steve Henson] | |
1714 | ||
1715 | *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can | |
1716 | be triggered by sending invalid DTLS fragments to an OpenSSL DTLS | |
1717 | client or server. This is potentially exploitable to run arbitrary | |
1718 | code on a vulnerable client or server. | |
1719 | ||
9f0b86c6 RL |
1720 | Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195) |
1721 | [Jüri Aedla, Steve Henson] | |
68a1e0bc RL |
1722 | |
1723 | *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites | |
1724 | are subject to a denial of service attack. | |
1725 | ||
9f0b86c6 | 1726 | Thanks to Felix Gröbert and Ivan Fratric at Google for discovering |
68a1e0bc | 1727 | this issue. (CVE-2014-3470) |
9f0b86c6 | 1728 | [Felix Gröbert, Ivan Fratric, Steve Henson] |
68a1e0bc RL |
1729 | |
1730 | *) Harmonize version and its documentation. -f flag is used to display | |
1731 | compilation flags. | |
1732 | [mancha <mancha1@zoho.com>] | |
1733 | ||
1734 | *) Fix eckey_priv_encode so it immediately returns an error upon a failure | |
b68fa4d1 | 1735 | in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue. |
68a1e0bc RL |
1736 | [mancha <mancha1@zoho.com>] |
1737 | ||
1738 | *) Fix some double frees. These are not thought to be exploitable. | |
1739 | [mancha <mancha1@zoho.com>] | |
1740 | ||
13738d5f DSH |
1741 | Changes between 1.0.1f and 1.0.1g [7 Apr 2014] |
1742 | ||
1743 | *) A missing bounds check in the handling of the TLS heartbeat extension | |
1744 | can be used to reveal up to 64k of memory to a connected client or | |
1745 | server. | |
1746 | ||
1747 | Thanks for Neel Mehta of Google Security for discovering this bug and to | |
1748 | Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for | |
1749 | preparing the fix (CVE-2014-0160) | |
1750 | [Adam Langley, Bodo Moeller] | |
1751 | ||
1752 | *) Fix for the attack described in the paper "Recovering OpenSSL | |
1753 | ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack" | |
1754 | by Yuval Yarom and Naomi Benger. Details can be obtained from: | |
1755 | http://eprint.iacr.org/2014/140 | |
1756 | ||
1757 | Thanks to Yuval Yarom and Naomi Benger for discovering this | |
1758 | flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076) | |
1759 | [Yuval Yarom and Naomi Benger] | |
1760 | ||
1761 | *) TLS pad extension: draft-agl-tls-padding-03 | |
1762 | ||
1763 | Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the | |
1764 | TLS client Hello record length value would otherwise be > 255 and | |
1765 | less that 512 pad with a dummy extension containing zeroes so it | |
1766 | is at least 512 bytes long. | |
1767 | ||
1768 | [Adam Langley, Steve Henson] | |
1769 | ||
802db0fa DSH |
1770 | Changes between 1.0.1e and 1.0.1f [6 Jan 2014] |
1771 | ||
1772 | *) Fix for TLS record tampering bug. A carefully crafted invalid | |
1773 | handshake could crash OpenSSL with a NULL pointer exception. | |
1774 | Thanks to Anton Johansson for reporting this issues. | |
1775 | (CVE-2013-4353) | |
1776 | ||
1777 | *) Keep original DTLS digest and encryption contexts in retransmission | |
1778 | structures so we can use the previous session parameters if they need | |
1779 | to be resent. (CVE-2013-6450) | |
1780 | [Steve Henson] | |
1781 | ||
1782 | *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which | |
1783 | avoids preferring ECDHE-ECDSA ciphers when the client appears to be | |
1784 | Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for | |
1785 | several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug | |
1786 | is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing | |
1787 | 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer. | |
1788 | [Rob Stradling, Adam Langley] | |
1789 | ||
1b9a59c3 BM |
1790 | Changes between 1.0.1d and 1.0.1e [11 Feb 2013] |
1791 | ||
1792 | *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI | |
1793 | supporting platforms or when small records were transferred. | |
1794 | [Andy Polyakov, Steve Henson] | |
1795 | ||
1796 | Changes between 1.0.1c and 1.0.1d [5 Feb 2013] | |
1797 | ||
1798 | *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time. | |
1799 | ||
1800 | This addresses the flaw in CBC record processing discovered by | |
1801 | Nadhem Alfardan and Kenny Paterson. Details of this attack can be found | |
1802 | at: http://www.isg.rhul.ac.uk/tls/ | |
1803 | ||
1804 | Thanks go to Nadhem Alfardan and Kenny Paterson of the Information | |
1805 | Security Group at Royal Holloway, University of London | |
1806 | (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and | |
9f0b86c6 | 1807 | Emilia Käsper for the initial patch. |
1b9a59c3 | 1808 | (CVE-2013-0169) |
9f0b86c6 | 1809 | [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson] |
1b9a59c3 BM |
1810 | |
1811 | *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode | |
1812 | ciphersuites which can be exploited in a denial of service attack. | |
1813 | Thanks go to and to Adam Langley <agl@chromium.org> for discovering | |
1814 | and detecting this bug and to Wolfgang Ettlinger | |
1815 | <wolfgang.ettlinger@gmail.com> for independently discovering this issue. | |
1816 | (CVE-2012-2686) | |
1817 | [Adam Langley] | |
1818 | ||
1819 | *) Return an error when checking OCSP signatures when key is NULL. | |
1820 | This fixes a DoS attack. (CVE-2013-0166) | |
1821 | [Steve Henson] | |
5e145e54 | 1822 | |
5f4cf088 BL |
1823 | *) Make openssl verify return errors. |
1824 | [Chris Palmer <palmer@google.com> and Ben Laurie] | |
1825 | ||
1b9a59c3 BM |
1826 | *) Call OCSP Stapling callback after ciphersuite has been chosen, so |
1827 | the right response is stapled. Also change SSL_get_certificate() | |
1828 | so it returns the certificate actually sent. | |
1829 | See http://rt.openssl.org/Ticket/Display.html?id=2836. | |
1830 | [Rob Stradling <rob.stradling@comodo.com>] | |
1831 | ||
482f2380 DSH |
1832 | *) Fix possible deadlock when decoding public keys. |
1833 | [Steve Henson] | |
1834 | ||
5e145e54 DSH |
1835 | *) Don't use TLS 1.0 record version number in initial client hello |
1836 | if renegotiating. | |
1837 | [Steve Henson] | |
1838 | ||
1839 | Changes between 1.0.1b and 1.0.1c [10 May 2012] | |
a56f9a61 | 1840 | |
e7c84838 | 1841 | *) Sanity check record length before skipping explicit IV in TLS |
1b9a59c3 | 1842 | 1.2, 1.1 and DTLS to fix DoS attack. |
e7c84838 DSH |
1843 | |
1844 | Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic | |
1845 | fuzzing as a service testing platform. | |
1846 | (CVE-2012-2333) | |
1847 | [Steve Henson] | |
1848 | ||
24547c23 DSH |
1849 | *) Initialise tkeylen properly when encrypting CMS messages. |
1850 | Thanks to Solar Designer of Openwall for reporting this issue. | |
1851 | [Steve Henson] | |
1852 | ||
a56f9a61 DSH |
1853 | *) In FIPS mode don't try to use composite ciphers as they are not |
1854 | approved. | |
1855 | [Steve Henson] | |
1856 | ||
1857 | Changes between 1.0.1a and 1.0.1b [26 Apr 2012] | |
f69abd53 | 1858 | |
7e0c9630 | 1859 | *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and |
a6df6702 DSH |
1860 | 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately |
1861 | mean any application compiled against OpenSSL 1.0.0 headers setting | |
1862 | SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disablng | |
7e0c9630 | 1863 | TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to |
a6df6702 DSH |
1864 | 0x10000000L Any application which was previously compiled against |
1865 | OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1 | |
7e0c9630 AP |
1866 | will need to be recompiled as a result. Letting be results in |
1867 | inability to disable specifically TLS 1.1 and in client context, | |
1868 | in unlike event, limit maximum offered version to TLS 1.0 [see below]. | |
a6df6702 DSH |
1869 | [Steve Henson] |
1870 | ||
f69abd53 AP |
1871 | *) In order to ensure interoperabilty SSL_OP_NO_protocolX does not |
1872 | disable just protocol X, but all protocols above X *if* there are | |
1873 | protocols *below* X still enabled. In more practical terms it means | |
1874 | that if application wants to disable TLS1.0 in favor of TLS1.1 and | |
1875 | above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass | |
0ae89cf3 AP |
1876 | SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to |
1877 | client side. | |
f69abd53 AP |
1878 | [Andy Polyakov] |
1879 | ||
d6ef8165 | 1880 | Changes between 1.0.1 and 1.0.1a [19 Apr 2012] |
54543b95 | 1881 | |
564a503b DSH |
1882 | *) Check for potentially exploitable overflows in asn1_d2i_read_bio |
1883 | BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer | |
1884 | in CRYPTO_realloc_clean. | |
1885 | ||
1886 | Thanks to Tavis Ormandy, Google Security Team, for discovering this | |
1887 | issue and to Adam Langley <agl@chromium.org> for fixing it. | |
1888 | (CVE-2012-2110) | |
1889 | [Adam Langley (Google), Tavis Ormandy, Google Security Team] | |
1890 | ||
bb3add20 BM |
1891 | *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections. |
1892 | [Adam Langley] | |
1893 | ||
48e0f666 | 1894 | *) Workarounds for some broken servers that "hang" if a client hello |
1b9a59c3 BM |
1895 | record length exceeds 255 bytes. |
1896 | ||
48e0f666 DSH |
1897 | 1. Do not use record version number > TLS 1.0 in initial client |
1898 | hello: some (but not all) hanging servers will now work. | |
1899 | 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate | |
1b9a59c3 | 1900 | the number of ciphers sent in the client hello. This should be |
48e0f666 DSH |
1901 | set to an even number, such as 50, for example by passing: |
1902 | -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure. | |
1903 | Most broken servers should now work. | |
1904 | 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable | |
1b9a59c3 | 1905 | TLS 1.2 client support entirely. |
a6df6702 | 1906 | [Steve Henson] |
48e0f666 | 1907 | |
54543b95 AP |
1908 | *) Fix SEGV in Vector Permutation AES module observed in OpenSSH. |
1909 | [Andy Polyakov] | |
1910 | ||
f3dcae15 | 1911 | Changes between 1.0.0h and 1.0.1 [14 Mar 2012] |
9472baae | 1912 | |
0cd7a032 DSH |
1913 | *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET |
1914 | STRING form instead of a DigestInfo. | |
1915 | [Steve Henson] | |
1916 | ||
16b7c81d DSH |
1917 | *) The format used for MDC2 RSA signatures is inconsistent between EVP |
1918 | and the RSA_sign/RSA_verify functions. This was made more apparent when | |
1919 | OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular | |
1920 | those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect | |
1921 | the correct format in RSA_verify so both forms transparently work. | |
1922 | [Steve Henson] | |
1923 | ||
fc6800d1 DSH |
1924 | *) Some servers which support TLS 1.0 can choke if we initially indicate |
1925 | support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA | |
1926 | encrypted premaster secret. As a workaround use the maximum pemitted | |
1927 | client version in client hello, this should keep such servers happy | |
1928 | and still work with previous versions of OpenSSL. | |
1929 | [Steve Henson] | |
1930 | ||
bd6941cf DSH |
1931 | *) Add support for TLS/DTLS heartbeats. |
1932 | [Robin Seggelmann <seggelmann@fh-muenster.de>] | |
1933 | ||
6e750fcb DSH |
1934 | *) Add support for SCTP. |
1935 | [Robin Seggelmann <seggelmann@fh-muenster.de>] | |
1936 | ||
62308f3f DSH |
1937 | *) Improved PRNG seeding for VOS. |
1938 | [Paul Green <Paul.Green@stratus.com>] | |
1939 | ||
cecafcce AP |
1940 | *) Extensive assembler packs updates, most notably: |
1941 | ||
1942 | - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support; | |
1943 | - x86[_64]: SSSE3 support (SHA1, vector-permutation AES); | |
1944 | - x86_64: bit-sliced AES implementation; | |
1945 | - ARM: NEON support, contemporary platforms optimizations; | |
1946 | - s390x: z196 support; | |
1947 | - *: GHASH and GF(2^m) multiplication implementations; | |
1948 | ||
1949 | [Andy Polyakov] | |
1950 | ||
ca0efb75 DSH |
1951 | *) Make TLS-SRP code conformant with RFC 5054 API cleanup |
1952 | (removal of unnecessary code) | |
1953 | [Peter Sylvester <peter.sylvester@edelweb.fr>] | |
1954 | ||
b1d74291 BL |
1955 | *) Add TLS key material exporter from RFC 5705. |
1956 | [Eric Rescorla] | |
1957 | ||
060a38a2 BL |
1958 | *) Add DTLS-SRTP negotiation from RFC 5764. |
1959 | [Eric Rescorla] | |
1960 | ||
e2809bfb BL |
1961 | *) Add Next Protocol Negotiation, |
1962 | http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be | |
1963 | disabled with a no-npn flag to config or Configure. Code donated | |
1964 | by Google. | |
1965 | [Adam Langley <agl@google.com> and Ben Laurie] | |
1966 | ||
9c37519b BM |
1967 | *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224, |
1968 | NIST-P256, NIST-P521, with constant-time single point multiplication on | |
1969 | typical inputs. Compiler support for the nonstandard type __uint128_t is | |
3d520f7c BM |
1970 | required to use this (present in gcc 4.4 and later, for 64-bit builds). |
1971 | Code made available under Apache License version 2.0. | |
9c37519b | 1972 | |
3d520f7c BM |
1973 | Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command |
1974 | line to include this in your build of OpenSSL, and run "make depend" (or | |
1975 | "make update"). This enables the following EC_METHODs: | |
9c37519b BM |
1976 | |
1977 | EC_GFp_nistp224_method() | |
1978 | EC_GFp_nistp256_method() | |
1979 | EC_GFp_nistp521_method() | |
1980 | ||
1981 | EC_GROUP_new_by_curve_name() will automatically use these (while | |
1982 | EC_GROUP_new_curve_GFp() currently prefers the more flexible | |
1983 | implementations). | |
9f0b86c6 | 1984 | [Emilia Käsper, Adam Langley, Bodo Moeller (Google)] |
9c37519b | 1985 | |
cb70355d DSH |
1986 | *) Use type ossl_ssize_t instad of ssize_t which isn't available on |
1987 | all platforms. Move ssize_t definition from e_os.h to the public | |
1988 | header file e_os2.h as it now appears in public header file cms.h | |
1989 | [Steve Henson] | |
1990 | ||
9309ea66 DSH |
1991 | *) New -sigopt option to the ca, req and x509 utilities. Additional |
1992 | signature parameters can be passed using this option and in | |
1993 | particular PSS. | |
1994 | [Steve Henson] | |
1995 | ||
1996 | *) Add RSA PSS signing function. This will generate and set the | |
1997 | appropriate AlgorithmIdentifiers for PSS based on those in the | |
1998 | corresponding EVP_MD_CTX structure. No application support yet. | |
1999 | [Steve Henson] | |
2000 | ||
2001 | *) Support for companion algorithm specific ASN1 signing routines. | |
2002 | New function ASN1_item_sign_ctx() signs a pre-initialised | |
2003 | EVP_MD_CTX structure and sets AlgorithmIdentifiers based on | |
2004 | the appropriate parameters. | |
2005 | [Steve Henson] | |
2006 | ||
2007 | *) Add new algorithm specific ASN1 verification initialisation function | |
2008 | to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1 | |
2009 | handling will be the same no matter what EVP_PKEY_METHOD is used. | |
2010 | Add a PSS handler to support verification of PSS signatures: checked | |
2011 | against a number of sample certificates. | |
2012 | [Steve Henson] | |
2013 | ||
2014 | *) Add signature printing for PSS. Add PSS OIDs. | |
2015 | [Steve Henson, Martin Kaiser <lists@kaiser.cx>] | |
2016 | ||
2017 | *) Add algorithm specific signature printing. An individual ASN1 method | |
2018 | can now print out signatures instead of the standard hex dump. | |
2019 | ||
2020 | More complex signatures (e.g. PSS) can print out more meaningful | |
2021 | information. Include DSA version that prints out the signature | |
2022 | parameters r, s. | |
2023 | [Steve Henson] | |
2024 | ||
05c9e3ae DSH |
2025 | *) Password based recipient info support for CMS library: implementing |
2026 | RFC3211. | |
dc100d87 DSH |
2027 | [Steve Henson] |
2028 | ||
2029 | *) Split password based encryption into PBES2 and PBKDF2 functions. This | |
2030 | neatly separates the code into cipher and PBE sections and is required | |
2031 | for some algorithms that split PBES2 into separate pieces (such as | |
2032 | password based CMS). | |
2033 | [Steve Henson] | |
2034 | ||
3c3f0259 BM |
2035 | *) Session-handling fixes: |
2036 | - Fix handling of connections that are resuming with a session ID, | |
2037 | but also support Session Tickets. | |
2038 | - Fix a bug that suppressed issuing of a new ticket if the client | |
2039 | presented a ticket with an expired session. | |
2040 | - Try to set the ticket lifetime hint to something reasonable. | |
2041 | - Make tickets shorter by excluding irrelevant information. | |
2042 | - On the client side, don't ignore renewed tickets. | |
2043 | [Adam Langley, Bodo Moeller (Google)] | |
2044 | ||
5ff6e2df BM |
2045 | *) Fix PSK session representation. |
2046 | [Bodo Moeller] | |
2047 | ||
84e7485b AP |
2048 | *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations. |
2049 | ||
2050 | This work was sponsored by Intel. | |
2051 | [Andy Polyakov] | |
2052 | ||
aed53d6c DSH |
2053 | *) Add GCM support to TLS library. Some custom code is needed to split |
2054 | the IV between the fixed (from PRF) and explicit (from TLS record) | |
2055 | portions. This adds all GCM ciphersuites supported by RFC5288 and | |
2056 | RFC5289. Generalise some AES* cipherstrings to inlclude GCM and | |
2057 | add a special AESGCM string for GCM only. | |
2058 | [Steve Henson] | |
2059 | ||
2060 | *) Expand range of ctrls for AES GCM. Permit setting invocation | |
2061 | field on decrypt and retrieval of invocation field only on encrypt. | |
2062 | [Steve Henson] | |
2063 | ||
c8c6e9ec DSH |
2064 | *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support. |
2065 | As required by RFC5289 these ciphersuites cannot be used if for | |
2066 | versions of TLS earlier than 1.2. | |
2067 | [Steve Henson] | |
2068 | ||
3a5b97b7 DSH |
2069 | *) For FIPS capable OpenSSL interpret a NULL default public key method |
2070 | as unset and return the appopriate default but do *not* set the default. | |
2071 | This means we can return the appopriate method in applications that | |
2072 | swicth between FIPS and non-FIPS modes. | |
2073 | [Steve Henson] | |
2074 | ||
e8d23f78 DSH |
2075 | *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an |
2076 | ENGINE is used then we cannot handle that in the FIPS module so we | |
2077 | keep original code iff non-FIPS operations are allowed. | |
2078 | [Steve Henson] | |
2079 | ||
81739603 DSH |
2080 | *) Add -attime option to openssl utilities. |
2081 | [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson] | |
be23b71e | 2082 | |
752c1a0c DSH |
2083 | *) Redirect DSA and DH operations to FIPS module in FIPS mode. |
2084 | [Steve Henson] | |
2085 | ||
6342b6e3 DSH |
2086 | *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use |
2087 | FIPS EC methods unconditionally for now. | |
2088 | [Steve Henson] | |
2089 | ||
f610a516 DSH |
2090 | *) New build option no-ec2m to disable characteristic 2 code. |
2091 | [Steve Henson] | |
2092 | ||
5cacc82f | 2093 | *) Backport libcrypto audit of return value checking from 1.1.0-dev; not |
24d7159a DSH |
2094 | all cases can be covered as some introduce binary incompatibilities. |
2095 | [Steve Henson] | |
2096 | ||
53dd05d8 DSH |
2097 | *) Redirect RSA operations to FIPS module including keygen, |
2098 | encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods. | |
fbe70553 DSH |
2099 | [Steve Henson] |
2100 | ||
916bcab2 DSH |
2101 | *) Add similar low level API blocking to ciphers. |
2102 | [Steve Henson] | |
2103 | ||
65300dcf DSH |
2104 | *) Low level digest APIs are not approved in FIPS mode: any attempt |
2105 | to use these will cause a fatal error. Applications that *really* want | |
2106 | to use them can use the private_* version instead. | |
2107 | [Steve Henson] | |
2108 | ||
5792219d DSH |
2109 | *) Redirect cipher operations to FIPS module for FIPS builds. |
2110 | [Steve Henson] | |
2111 | ||
04dc5a9c DSH |
2112 | *) Redirect digest operations to FIPS module for FIPS builds. |
2113 | [Steve Henson] | |
2114 | ||
2115 | *) Update build system to add "fips" flag which will link in fipscanister.o | |
2116 | for static and shared library builds embedding a signature if needed. | |
2117 | [Steve Henson] | |
2118 | ||
55a47cd3 DSH |
2119 | *) Output TLS supported curves in preference order instead of numerical |
2120 | order. This is currently hardcoded for the highest order curves first. | |
2121 | This should be configurable so applications can judge speed vs strength. | |
2122 | [Steve Henson] | |
2123 | ||
b81fde02 DSH |
2124 | *) Add TLS v1.2 server support for client authentication. |
2125 | [Steve Henson] | |
2126 | ||
7043fa70 DSH |
2127 | *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers |
2128 | and enable MD5. | |
2129 | [Steve Henson] | |
2130 | ||
f98d2e5c DSH |
2131 | *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying |
2132 | FIPS modules versions. | |
2133 | [Steve Henson] | |
2134 | ||
4fe4c00e DSH |
2135 | *) Add TLS v1.2 client side support for client authentication. Keep cache |
2136 | of handshake records longer as we don't know the hash algorithm to use | |
2137 | until after the certificate request message is received. | |
2138 | [Steve Henson] | |
2139 | ||
9472baae DSH |
2140 | *) Initial TLS v1.2 client support. Add a default signature algorithms |
2141 | extension including all the algorithms we support. Parse new signature | |
2142 | format in client key exchange. Relax some ECC signing restrictions for | |
2143 | TLS v1.2 as indicated in RFC5246. | |
2144 | [Steve Henson] | |
2145 | ||
2146 | *) Add server support for TLS v1.2 signature algorithms extension. Switch | |
2147 | to new signature format when needed using client digest preference. | |
2148 | All server ciphersuites should now work correctly in TLS v1.2. No client | |
2149 | support yet and no support for client certificates. | |
2150 | [Steve Henson] | |
2151 | ||
2152 | *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch | |
2153 | to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based | |
2154 | ciphersuites. At present only RSA key exchange ciphersuites work with | |
2155 | TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete | |
2156 | SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods | |
2157 | and version checking. | |
2158 | [Steve Henson] | |
5cacc82f | 2159 | |
74096890 DSH |
2160 | *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled |
2161 | with this defined it will not be affected by any changes to ssl internal | |
2162 | structures. Add several utility functions to allow openssl application | |
2163 | to work with OPENSSL_NO_SSL_INTERN defined. | |
2164 | [Steve Henson] | |
c549810d | 2165 | |
726cd626 RS |
2166 | *) A long standing patch to add support for SRP from EdelWeb (Peter |
2167 | Sylvester and Christophe Renou) was integrated. | |
2168 | [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester | |
2169 | <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and | |
2170 | Ben Laurie] | |
a149b246 | 2171 | |
a618011c DSH |
2172 | *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id. |
2173 | [Steve Henson] | |
2174 | ||
48ae85b6 DSH |
2175 | *) Permit abbreviated handshakes when renegotiating using the function |
2176 | SSL_renegotiate_abbreviated(). | |
2177 | [Robin Seggelmann <seggelmann@fh-muenster.de>] | |
2178 | ||
160f9b5b DSH |
2179 | *) Add call to ENGINE_register_all_complete() to |
2180 | ENGINE_load_builtin_engines(), so some implementations get used | |
2181 | automatically instead of needing explicit application support. | |
2182 | [Steve Henson] | |
2183 | ||
53e7985c DSH |
2184 | *) Add support for TLS key exporter as described in RFC5705. |
2185 | [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson] | |
2186 | ||
1eb1cf45 DSH |
2187 | *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only |
2188 | a few changes are required: | |
2189 | ||
2190 | Add SSL_OP_NO_TLSv1_1 flag. | |
2191 | Add TLSv1_1 methods. | |
2192 | Update version checking logic to handle version 1.1. | |
2193 | Add explicit IV handling (ported from DTLS code). | |
2194 | Add command line options to s_client/s_server. | |
2195 | [Steve Henson] | |
c549810d | 2196 | |
54543b95 AP |
2197 | Changes between 1.0.0g and 1.0.0h [12 Mar 2012] |
2198 | ||
2199 | *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness | |
2200 | in CMS and PKCS7 code. When RSA decryption fails use a random key for | |
2201 | content decryption and always return the same error. Note: this attack | |
2202 | needs on average 2^20 messages so it only affects automated senders. The | |
2203 | old behaviour can be reenabled in the CMS code by setting the | |
2204 | CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where | |
2205 | an MMA defence is not necessary. | |
2206 | Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering | |
2207 | this issue. (CVE-2012-0884) | |
2208 | [Steve Henson] | |
a8314df9 DSH |
2209 | |
2210 | *) Fix CVE-2011-4619: make sure we really are receiving a | |
2211 | client hello before rejecting multiple SGC restarts. Thanks to | |
2212 | Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug. | |
2213 | [Steve Henson] | |
2214 | ||
2dc4b0db DSH |
2215 | Changes between 1.0.0f and 1.0.0g [18 Jan 2012] |
2216 | ||
2217 | *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109. | |
2218 | Thanks to Antonio Martin, Enterprise Secure Access Research and | |
2219 | Development, Cisco Systems, Inc. for discovering this bug and | |
2220 | preparing a fix. (CVE-2012-0050) | |
2221 | [Antonio Martin] | |
2222 | ||
801e5ef8 | 2223 | Changes between 1.0.0e and 1.0.0f [4 Jan 2012] |
f72c1a58 | 2224 | |
0044739a DSH |
2225 | *) Nadhem Alfardan and Kenny Paterson have discovered an extension |
2226 | of the Vaudenay padding oracle attack on CBC mode encryption | |
2227 | which enables an efficient plaintext recovery attack against | |
2228 | the OpenSSL implementation of DTLS. Their attack exploits timing | |
2229 | differences arising during decryption processing. A research | |
2230 | paper describing this attack can be found at: | |
2231 | http://www.isg.rhul.ac.uk/~kp/dtls.pdf | |
2232 | Thanks go to Nadhem Alfardan and Kenny Paterson of the Information | |
2233 | Security Group at Royal Holloway, University of London | |
2234 | (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann | |
2235 | <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de> | |
2236 | for preparing the fix. (CVE-2011-4108) | |
2237 | [Robin Seggelmann, Michael Tuexen] | |
2238 | ||
4e44bd36 DSH |
2239 | *) Clear bytes used for block padding of SSL 3.0 records. |
2240 | (CVE-2011-4576) | |
2241 | [Adam Langley (Google)] | |
2242 | ||
25e3d222 DSH |
2243 | *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George |
2244 | Kadianakis <desnacked@gmail.com> for discovering this issue and | |
2245 | Adam Langley for preparing the fix. (CVE-2011-4619) | |
0cffb0cd DSH |
2246 | [Adam Langley (Google)] |
2247 | ||
2248 | *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027) | |
2249 | [Andrey Kulikov <amdeich@gmail.com>] | |
2250 | ||
2251 | *) Prevent malformed RFC3779 data triggering an assertion failure. | |
2252 | Thanks to Andrew Chi, BBN Technologies, for discovering the flaw | |
2253 | and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577) | |
2254 | [Rob Austein <sra@hactrn.net>] | |
2255 | ||
767d3e00 BM |
2256 | *) Improved PRNG seeding for VOS. |
2257 | [Paul Green <Paul.Green@stratus.com>] | |
2258 | ||
9f2b4533 BM |
2259 | *) Fix ssl_ciph.c set-up race. |
2260 | [Adam Langley (Google)] | |
2261 | ||
a0dce9be | 2262 | *) Fix spurious failures in ecdsatest.c. |
9f0b86c6 | 2263 | [Emilia Käsper (Google)] |
a0dce9be | 2264 | |
cf2b9385 BM |
2265 | *) Fix the BIO_f_buffer() implementation (which was mixing different |
2266 | interpretations of the '..._len' fields). | |
2267 | [Adam Langley (Google)] | |
2268 | ||
2d95ceed BM |
2269 | *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than |
2270 | BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent | |
2271 | threads won't reuse the same blinding coefficients. | |
2272 | ||
2273 | This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING | |
2274 | lock to call BN_BLINDING_invert_ex, and avoids one use of | |
2275 | BN_BLINDING_update for each BN_BLINDING structure (previously, | |
2276 | the last update always remained unused). | |
9f0b86c6 | 2277 | [Emilia Käsper (Google)] |
2d95ceed | 2278 | |
f72c1a58 BM |
2279 | *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf. |
2280 | [Bob Buckholz (Google)] | |
2281 | ||
2282 | Changes between 1.0.0d and 1.0.0e [6 Sep 2011] | |
5cacc82f | 2283 | |
cd447875 DSH |
2284 | *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted |
2285 | by initialising X509_STORE_CTX properly. (CVE-2011-3207) | |
2286 | [Kaspar Brand <ossl@velox.ch>] | |
2287 | ||
61ac68f9 | 2288 | *) Fix SSL memory handling for (EC)DH ciphersuites, in particular |
cd447875 | 2289 | for multi-threaded use of ECDH. (CVE-2011-3210) |
61ac68f9 BM |
2290 | [Adam Langley (Google)] |
2291 | ||
7f1022a8 BM |
2292 | *) Fix x509_name_ex_d2i memory leak on bad inputs. |
2293 | [Bodo Moeller] | |
2294 | ||
cf199fec DSH |
2295 | *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check |
2296 | signature public key algorithm by using OID xref utilities instead. | |
2297 | Before this you could only use some ECC ciphersuites with SHA1 only. | |
2298 | [Steve Henson] | |
2299 | ||
5cacc82f BM |
2300 | *) Add protection against ECDSA timing attacks as mentioned in the paper |
2301 | by Billy Bob Brumley and Nicola Tuveri, see: | |
2302 | ||
2303 | http://eprint.iacr.org/2011/232.pdf | |
2304 | ||
2305 | [Billy Bob Brumley and Nicola Tuveri] | |
2306 | ||
cd77b3e8 BM |
2307 | Changes between 1.0.0c and 1.0.0d [8 Feb 2011] |
2308 | ||
2309 | *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014 | |
2310 | [Neel Mehta, Adam Langley, Bodo Moeller (Google)] | |
e501dbb6 DSH |
2311 | |
2312 | *) Fix bug in string printing code: if *any* escaping is enabled we must | |
2313 | escape the escape character (backslash) or the resulting string is | |
2314 | ambiguous. | |
2315 | [Steve Henson] | |
2316 | ||
2317 | Changes between 1.0.0b and 1.0.0c [2 Dec 2010] | |
2c5c4fca | 2318 | |
346601bc BM |
2319 | *) Disable code workaround for ancient and obsolete Netscape browsers |
2320 | and servers: an attacker can use it in a ciphersuite downgrade attack. | |
2321 | Thanks to Martin Rex for discovering this bug. CVE-2010-4180 | |
2322 | [Steve Henson] | |
2323 | ||
2c5c4fca DSH |
2324 | *) Fixed J-PAKE implementation error, originally discovered by |
2325 | Sebastien Martini, further info and confirmation from Stefan | |
2326 | Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252 | |
2327 | [Ben Laurie] | |
2328 | ||
2329 | Changes between 1.0.0a and 1.0.0b [16 Nov 2010] | |
1dba06e7 | 2330 | |
6e21ce59 DSH |
2331 | *) Fix extension code to avoid race conditions which can result in a buffer |
2332 | overrun vulnerability: resumed sessions must not be modified as they can | |
2333 | be shared by multiple threads. CVE-2010-3864 | |
346601bc | 2334 | [Steve Henson] |
6e21ce59 | 2335 | |
f6c29ba3 DSH |
2336 | *) Fix WIN32 build system to correctly link an ENGINE directory into |
2337 | a DLL. | |
2338 | [Steve Henson] | |
1dba06e7 | 2339 | |
9c7baca8 | 2340 | Changes between 1.0.0 and 1.0.0a [01 Jun 2010] |
6747de65 | 2341 | |
618265e6 DSH |
2342 | *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover |
2343 | (CVE-2010-1633) | |
2344 | [Steve Henson, Peter-Michael Hager <hager@dortmund.net>] | |
6747de65 | 2345 | |
91bad2b0 | 2346 | Changes between 0.9.8n and 1.0.0 [29 Mar 2010] |
93fac08e | 2347 | |
17004262 DSH |
2348 | *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher |
2349 | context. The operation can be customised via the ctrl mechanism in | |
2350 | case ENGINEs want to include additional functionality. | |
2351 | [Steve Henson] | |
2352 | ||
1699389a DSH |
2353 | *) Tolerate yet another broken PKCS#8 key format: private key value negative. |
2354 | [Steve Henson] | |
2355 | ||
93fac08e DSH |
2356 | *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to |
2357 | output hashes compatible with older versions of OpenSSL. | |
2358 | [Willy Weisz <weisz@vcpc.univie.ac.at>] | |
5b5464d5 | 2359 | |
e642fd7a DSH |
2360 | *) Fix compression algorithm handling: if resuming a session use the |
2361 | compression algorithm of the resumed session instead of determining | |
2362 | it from client hello again. Don't allow server to change algorithm. | |
2363 | [Steve Henson] | |
aaf35f11 | 2364 | |
96109228 DSH |
2365 | *) Add load_crls() function to apps tidying load_certs() too. Add option |
2366 | to verify utility to allow additional CRLs to be included. | |
2367 | [Steve Henson] | |
2368 | ||
0c690586 DSH |
2369 | *) Update OCSP request code to permit adding custom headers to the request: |
2370 | some responders need this. | |
2371 | [Steve Henson] | |
2372 | ||
80afb40a DSH |
2373 | *) The function EVP_PKEY_sign() returns <=0 on error: check return code |
2374 | correctly. | |
2375 | [Julia Lawall <julia@diku.dk>] | |
2376 | ||
c9add317 DSH |
2377 | *) Update verify callback code in apps/s_cb.c and apps/verify.c, it |
2378 | needlessly dereferenced structures, used obsolete functions and | |
2379 | didn't handle all updated verify codes correctly. | |
2380 | [Steve Henson] | |
2381 | ||
aefb9dc5 | 2382 | *) Disable MD2 in the default configuration. |
11ba084e DSH |
2383 | [Steve Henson] |
2384 | ||
0cb76e79 DSH |
2385 | *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to |
2386 | indicate the initial BIO being pushed or popped. This makes it possible | |
2387 | to determine whether the BIO is the one explicitly called or as a result | |
2388 | of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so | |
2389 | it handles reference counts correctly and doesn't zero out the I/O bio | |
2390 | when it is not being explicitly popped. WARNING: applications which | |
2391 | included workarounds for the old buggy behaviour will need to be modified | |
2392 | or they could free up already freed BIOs. | |
2393 | [Steve Henson] | |
2394 | ||
aefb9dc5 BM |
2395 | *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni |
2396 | renaming to all platforms (within the 0.9.8 branch, this was | |
2397 | done conditionally on Netware platforms to avoid a name clash). | |
6178da01 DSH |
2398 | [Guenter <lists@gknw.net>] |
2399 | ||
e1f09dfd DSH |
2400 | *) Add ECDHE and PSK support to DTLS. |
2401 | [Michael Tuexen <tuexen@fh-muenster.de>] | |
2402 | ||
376bbb58 DSH |
2403 | *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't |
2404 | be used on C++. | |
2405 | [Steve Henson] | |
2406 | ||
19ae0907 DSH |
2407 | *) Add "missing" function EVP_MD_flags() (without this the only way to |
2408 | retrieve a digest flags is by accessing the structure directly. Update | |
2409 | EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest | |
2410 | or cipher is registered as in the "from" argument. Print out all | |
2411 | registered digests in the dgst usage message instead of manually | |
2412 | attempting to work them out. | |
2413 | [Steve Henson] | |
2414 | ||
9ae57435 DSH |
2415 | *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello: |
2416 | this allows the use of compression and extensions. Change default cipher | |
2417 | string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2 | |
2418 | by default unless an application cipher string requests it. | |
2419 | [Steve Henson] | |
2420 | ||
5d487626 DSH |
2421 | *) Alter match criteria in PKCS12_parse(). It used to try to use local |
2422 | key ids to find matching certificates and keys but some PKCS#12 files | |
2423 | don't follow the (somewhat unwritten) rules and this strategy fails. | |
2424 | Now just gather all certificates together and the first private key | |
2425 | then look for the first certificate that matches the key. | |
2426 | [Steve Henson] | |
2427 | ||
aaf35f11 DSH |
2428 | *) Support use of registered digest and cipher names for dgst and cipher |
2429 | commands instead of having to add each one as a special case. So now | |
2430 | you can do: | |
2431 | ||
2432 | openssl sha256 foo | |
2433 | ||
2434 | as well as: | |
2435 | ||
2436 | openssl dgst -sha256 foo | |
2437 | ||
2438 | and this works for ENGINE based algorithms too. | |
2439 | ||
2440 | [Steve Henson] | |
3ff55e96 | 2441 | |
b6af2c7e DSH |
2442 | *) Update Gost ENGINE to support parameter files. |
2443 | [Victor B. Wagner <vitus@cryptocom.ru>] | |
2444 | ||
33ab2e31 DSH |
2445 | *) Support GeneralizedTime in ca utility. |
2446 | [Oliver Martin <oliver@volatilevoid.net>, Steve Henson] | |
2447 | ||
c2c99e28 DSH |
2448 | *) Enhance the hash format used for certificate directory links. The new |
2449 | form uses the canonical encoding (meaning equivalent names will work | |
2450 | even if they aren't identical) and uses SHA1 instead of MD5. This form | |
2451 | is incompatible with the older format and as a result c_rehash should | |
2452 | be used to rebuild symbolic links. | |
2453 | [Steve Henson] | |
2454 | ||
8125d9f9 DSH |
2455 | *) Make PKCS#8 the default write format for private keys, replacing the |
2456 | traditional format. This form is standardised, more secure and doesn't | |
2457 | include an implicit MD5 dependency. | |
2458 | [Steve Henson] | |
2459 | ||
363bd0b4 DSH |
2460 | *) Add a $gcc_devteam_warn option to Configure. The idea is that any code |
2461 | committed to OpenSSL should pass this lot as a minimum. | |
2462 | [Steve Henson] | |
2463 | ||
12bf56c0 DSH |
2464 | *) Add session ticket override functionality for use by EAP-FAST. |
2465 | [Jouni Malinen <j@w1.fi>] | |
2466 | ||
87d52468 DSH |
2467 | *) Modify HMAC functions to return a value. Since these can be implemented |
2468 | in an ENGINE errors can occur. | |
2469 | [Steve Henson] | |
2470 | ||
1ea6472e BL |
2471 | *) Type-checked OBJ_bsearch_ex. |
2472 | [Ben Laurie] | |
2473 | ||
babb3798 BL |
2474 | *) Type-checked OBJ_bsearch. Also some constification necessitated |
2475 | by type-checking. Still to come: TXT_DB, bsearch(?), | |
2476 | OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING, | |
1ea6472e BL |
2477 | CONF_VALUE. |
2478 | [Ben Laurie] | |
babb3798 | 2479 | |
87d3a0cd DSH |
2480 | *) New function OPENSSL_gmtime_adj() to add a specific number of days and |
2481 | seconds to a tm structure directly, instead of going through OS | |
2482 | specific date routines. This avoids any issues with OS routines such | |
2483 | as the year 2038 bug. New *_adj() functions for ASN1 time structures | |
2484 | and X509_time_adj_ex() to cover the extended range. The existing | |
2485 | X509_time_adj() is still usable and will no longer have any date issues. | |
2486 | [Steve Henson] | |
2487 | ||
d43c4497 DSH |
2488 | *) Delta CRL support. New use deltas option which will attempt to locate |
2489 | and search any appropriate delta CRLs available. | |
2490 | ||
2491 | This work was sponsored by Google. | |
2492 | [Steve Henson] | |
2493 | ||
4b96839f DSH |
2494 | *) Support for CRLs partitioned by reason code. Reorganise CRL processing |
2495 | code and add additional score elements. Validate alternate CRL paths | |
2496 | as part of the CRL checking and indicate a new error "CRL path validation | |
2497 | error" in this case. Applications wanting additional details can use | |
2498 | the verify callback and check the new "parent" field. If this is not | |
2499 | NULL CRL path validation is taking place. Existing applications wont | |
2500 | see this because it requires extended CRL support which is off by | |
2501 | default. | |
2502 | ||
2503 | This work was sponsored by Google. | |
2504 | [Steve Henson] | |
2505 | ||
249a77f5 DSH |
2506 | *) Support for freshest CRL extension. |
2507 | ||
2508 | This work was sponsored by Google. | |
2509 | [Steve Henson] | |
2510 | ||
d0fff69d DSH |
2511 | *) Initial indirect CRL support. Currently only supported in the CRLs |
2512 | passed directly and not via lookup. Process certificate issuer | |
2513 | CRL entry extension and lookup CRL entries by bother issuer name | |
4b96839f | 2514 | and serial number. Check and process CRL issuer entry in IDP extension. |
d0fff69d DSH |
2515 | |
2516 | This work was sponsored by Google. | |
2517 | [Steve Henson] | |
2518 | ||
9d84d4ed DSH |
2519 | *) Add support for distinct certificate and CRL paths. The CRL issuer |
2520 | certificate is validated separately in this case. Only enabled if | |
2521 | an extended CRL support flag is set: this flag will enable additional | |
2522 | CRL functionality in future. | |
2523 | ||
2524 | This work was sponsored by Google. | |
2525 | [Steve Henson] | |
9d84d4ed | 2526 | |
002e66c0 DSH |
2527 | *) Add support for policy mappings extension. |
2528 | ||
2529 | This work was sponsored by Google. | |
2530 | [Steve Henson] | |
2531 | ||
e9746e03 DSH |
2532 | *) Fixes to pathlength constraint, self issued certificate handling, |
2533 | policy processing to align with RFC3280 and PKITS tests. | |
2534 | ||
2535 | This work was sponsored by Google. | |
2536 | [Steve Henson] | |
2537 | ||
2538 | *) Support for name constraints certificate extension. DN, email, DNS | |
2539 | and URI types are currently supported. | |
2540 | ||
2541 | This work was sponsored by Google. | |
2542 | [Steve Henson] | |
2543 | ||
4c329696 GT |
2544 | *) To cater for systems that provide a pointer-based thread ID rather |
2545 | than numeric, deprecate the current numeric thread ID mechanism and | |
2546 | replace it with a structure and associated callback type. This | |
2547 | mechanism allows a numeric "hash" to be extracted from a thread ID in | |
2548 | either case, and on platforms where pointers are larger than 'long', | |
2549 | mixing is done to help ensure the numeric 'hash' is usable even if it | |
2550 | can't be guaranteed unique. The default mechanism is to use "&errno" | |
2551 | as a pointer-based thread ID to distinguish between threads. | |
2552 | ||
2553 | Applications that want to provide their own thread IDs should now use | |
2554 | CRYPTO_THREADID_set_callback() to register a callback that will call | |
2555 | either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer(). | |
2556 | ||
2ecd2ede BM |
2557 | Note that ERR_remove_state() is now deprecated, because it is tied |
2558 | to the assumption that thread IDs are numeric. ERR_remove_state(0) | |
2559 | to free the current thread's error state should be replaced by | |
2560 | ERR_remove_thread_state(NULL). | |
2561 | ||
4c329696 GT |
2562 | (This new approach replaces the functions CRYPTO_set_idptr_callback(), |
2563 | CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in | |
2564 | OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an | |
2565 | application was previously providing a numeric thread callback that | |
2566 | was inappropriate for distinguishing threads, then uniqueness might | |
2567 | have been obtained with &errno that happened immediately in the | |
2568 | intermediate development versions of OpenSSL; this is no longer the | |
2569 | case, the numeric thread callback will now override the automatic use | |
2570 | of &errno.) | |
2571 | [Geoff Thorpe, with help from Bodo Moeller] | |
2572 | ||
5cbd2033 DSH |
2573 | *) Initial support for different CRL issuing certificates. This covers a |
2574 | simple case where the self issued certificates in the chain exist and | |
2575 | the real CRL issuer is higher in the existing chain. | |
e9746e03 DSH |
2576 | |
2577 | This work was sponsored by Google. | |
5cbd2033 DSH |
2578 | [Steve Henson] |
2579 | ||
5ce278a7 BL |
2580 | *) Removed effectively defunct crypto/store from the build. |
2581 | [Ben Laurie] | |
2582 | ||
2583 | *) Revamp of STACK to provide stronger type-checking. Still to come: | |
2584 | TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE, | |
2585 | ASN1_STRING, CONF_VALUE. | |
2586 | [Ben Laurie] | |
2587 | ||
8671b898 BL |
2588 | *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer |
2589 | RAM on SSL connections. This option can save about 34k per idle SSL. | |
2590 | [Nick Mathewson] | |
2591 | ||
3c1d6bbc BL |
2592 | *) Revamp of LHASH to provide stronger type-checking. Still to come: |
2593 | STACK, TXT_DB, bsearch, qsort. | |
2594 | [Ben Laurie] | |
2595 | ||
8931b30d DSH |
2596 | *) Initial support for Cryptographic Message Syntax (aka CMS) based |
2597 | on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility, | |
fd47c361 | 2598 | support for data, signedData, compressedData, digestedData and |
eb9d8d8c DSH |
2599 | encryptedData, envelopedData types included. Scripts to check against |
2600 | RFC4134 examples draft and interop and consistency checks of many | |
2601 | content types and variants. | |
8931b30d DSH |
2602 | [Steve Henson] |
2603 | ||
3df93571 | 2604 | *) Add options to enc utility to support use of zlib compression BIO. |
8931b30d DSH |
2605 | [Steve Henson] |
2606 | ||
73980531 DSH |
2607 | *) Extend mk1mf to support importing of options and assembly language |
2608 | files from Configure script, currently only included in VC-WIN32. | |
2609 | The assembly language rules can now optionally generate the source | |
2610 | files from the associated perl scripts. | |
2611 | [Steve Henson] | |
2612 | ||
0e1dba93 DSH |
2613 | *) Implement remaining functionality needed to support GOST ciphersuites. |
2614 | Interop testing has been performed using CryptoPro implementations. | |
2615 | [Victor B. Wagner <vitus@cryptocom.ru>] | |
2616 | ||
0023adb4 AP |
2617 | *) s390x assembler pack. |
2618 | [Andy Polyakov] | |
2619 | ||
4c7c5ff6 AP |
2620 | *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU |
2621 | "family." | |
2622 | [Andy Polyakov] | |
2623 | ||
761772d7 BM |
2624 | *) Implement Opaque PRF Input TLS extension as specified in |
2625 | draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an | |
2626 | official specification yet and no extension type assignment by | |
2627 | IANA exists, this extension (for now) will have to be explicitly | |
2628 | enabled when building OpenSSL by providing the extension number | |
2629 | to use. For example, specify an option | |
2630 | ||
2631 | -DTLSEXT_TYPE_opaque_prf_input=0x9527 | |
2632 | ||
2633 | to the "config" or "Configure" script to enable the extension, | |
2634 | assuming extension number 0x9527 (which is a completely arbitrary | |
2635 | and unofficial assignment based on the MD5 hash of the Internet | |
2636 | Draft). Note that by doing so, you potentially lose | |
2637 | interoperability with other TLS implementations since these might | |
2638 | be using the same extension number for other purposes. | |
2639 | ||
2640 | SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the | |
2641 | opaque PRF input value to use in the handshake. This will create | |
2642 | an interal copy of the length-'len' string at 'src', and will | |
2643 | return non-zero for success. | |
2644 | ||
2645 | To get more control and flexibility, provide a callback function | |
2646 | by using | |
2647 | ||
2648 | SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb) | |
2649 | SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg) | |
2650 | ||
2651 | where | |
2652 | ||
2653 | int (*cb)(SSL *, void *peerinput, size_t len, void *arg); | |
2654 | void *arg; | |
2655 | ||
2656 | Callback function 'cb' will be called in handshakes, and is | |
2657 | expected to use SSL_set_tlsext_opaque_prf_input() as appropriate. | |
2658 | Argument 'arg' is for application purposes (the value as given to | |
2659 | SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly | |
2660 | be provided to the callback function). The callback function | |
2661 | has to return non-zero to report success: usually 1 to use opaque | |
2662 | PRF input just if possible, or 2 to enforce use of the opaque PRF | |
2663 | input. In the latter case, the library will abort the handshake | |
2664 | if opaque PRF input is not successfully negotiated. | |
2665 | ||
2666 | Arguments 'peerinput' and 'len' given to the callback function | |
2667 | will always be NULL and 0 in the case of a client. A server will | |
2668 | see the client's opaque PRF input through these variables if | |
2669 | available (NULL and 0 otherwise). Note that if the server | |
2670 | provides an opaque PRF input, the length must be the same as the | |
2671 | length of the client's opaque PRF input. | |
2672 | ||
2673 | Note that the callback function will only be called when creating | |
2674 | a new session (session resumption can resume whatever was | |
2675 | previously negotiated), and will not be called in SSL 2.0 | |
2676 | handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or | |
2677 | SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended | |
2678 | for applications that need to enforce opaque PRF input. | |
2679 | ||
2680 | [Bodo Moeller] | |
2681 | ||
81025661 DSH |
2682 | *) Update ssl code to support digests other than SHA1+MD5 for handshake |
2683 | MAC. | |
2684 | ||
2685 | [Victor B. Wagner <vitus@cryptocom.ru>] | |
2686 | ||
6434abbf DSH |
2687 | *) Add RFC4507 support to OpenSSL. This includes the corrections in |
2688 | RFC4507bis. The encrypted ticket format is an encrypted encoded | |
2689 | SSL_SESSION structure, that way new session features are automatically | |
2690 | supported. | |
2691 | ||
ba0e826d DSH |
2692 | If a client application caches session in an SSL_SESSION structure |
2693 | support is transparent because tickets are now stored in the encoded | |
2694 | SSL_SESSION. | |
2695 | ||
2696 | The SSL_CTX structure automatically generates keys for ticket | |
2697 | protection in servers so again support should be possible | |
6434abbf DSH |
2698 | with no application modification. |
2699 | ||
2700 | If a client or server wishes to disable RFC4507 support then the option | |
2701 | SSL_OP_NO_TICKET can be set. | |
2702 | ||
2703 | Add a TLS extension debugging callback to allow the contents of any client | |
2704 | or server extensions to be examined. | |
ec5d7473 DSH |
2705 | |
2706 | This work was sponsored by Google. | |
6434abbf DSH |
2707 | [Steve Henson] |
2708 | ||
3c07d3a3 DSH |
2709 | *) Final changes to avoid use of pointer pointer casts in OpenSSL. |
2710 | OpenSSL should now compile cleanly on gcc 4.2 | |
2711 | [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson] | |
2712 | ||
b948e2c5 DSH |
2713 | *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC |
2714 | support including streaming MAC support: this is required for GOST | |
2715 | ciphersuite support. | |
2716 | [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson] | |
2717 | ||
9cfc8a9d DSH |
2718 | *) Add option -stream to use PKCS#7 streaming in smime utility. New |
2719 | function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream() | |
2720 | to output in BER and PEM format. | |
2721 | [Steve Henson] | |
2722 | ||
47b71e6e DSH |
2723 | *) Experimental support for use of HMAC via EVP_PKEY interface. This |
2724 | allows HMAC to be handled via the EVP_DigestSign*() interface. The | |
2725 | EVP_PKEY "key" in this case is the HMAC key, potentially allowing | |
2022cfe0 DSH |
2726 | ENGINE support for HMAC keys which are unextractable. New -mac and |
2727 | -macopt options to dgst utility. | |
47b71e6e DSH |
2728 | [Steve Henson] |
2729 | ||
d952c79a DSH |
2730 | *) New option -sigopt to dgst utility. Update dgst to use |
2731 | EVP_Digest{Sign,Verify}*. These two changes make it possible to use | |
2732 | alternative signing paramaters such as X9.31 or PSS in the dgst | |
2733 | utility. | |
2734 | [Steve Henson] | |
2735 | ||
fd5bc65c BM |
2736 | *) Change ssl_cipher_apply_rule(), the internal function that does |
2737 | the work each time a ciphersuite string requests enabling | |
2738 | ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or | |
2739 | removing ("!foo+bar") a class of ciphersuites: Now it maintains | |
2740 | the order of disabled ciphersuites such that those ciphersuites | |
2741 | that most recently went from enabled to disabled not only stay | |
2742 | in order with respect to each other, but also have higher priority | |
2743 | than other disabled ciphersuites the next time ciphersuites are | |
2744 | enabled again. | |
2745 | ||
2746 | This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable | |
2747 | the same ciphersuites as with "HIGH" alone, but in a specific | |
2748 | order where the PSK ciphersuites come first (since they are the | |
2749 | most recently disabled ciphersuites when "HIGH" is parsed). | |
2750 | ||
2751 | Also, change ssl_create_cipher_list() (using this new | |
2752 | funcionality) such that between otherwise identical | |
2753 | cihpersuites, ephemeral ECDH is preferred over ephemeral DH in | |
2754 | the default order. | |
2755 | [Bodo Moeller] | |
2756 | ||
0a05123a BM |
2757 | *) Change ssl_create_cipher_list() so that it automatically |
2758 | arranges the ciphersuites in reasonable order before starting | |
2759 | to process the rule string. Thus, the definition for "DEFAULT" | |
2760 | (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but | |
2761 | remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH". | |
2762 | This makes it much easier to arrive at a reasonable default order | |
2763 | in applications for which anonymous ciphers are OK (meaning | |
2764 | that you can't actually use DEFAULT). | |
2765 | [Bodo Moeller; suggested by Victor Duchovni] | |
2766 | ||
52b8dad8 BM |
2767 | *) Split the SSL/TLS algorithm mask (as used for ciphersuite string |
2768 | processing) into multiple integers instead of setting | |
2769 | "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK", | |
2770 | "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer. | |
2771 | (These masks as well as the individual bit definitions are hidden | |
2772 | away into the non-exported interface ssl/ssl_locl.h, so this | |
2773 | change to the definition of the SSL_CIPHER structure shouldn't | |
2774 | affect applications.) This give us more bits for each of these | |
2775 | categories, so there is no longer a need to coagulate AES128 and | |
2776 | AES256 into a single algorithm bit, and to coagulate Camellia128 | |
2777 | and Camellia256 into a single algorithm bit, which has led to all | |
2778 | kinds of kludges. | |
2779 | ||
2780 | Thus, among other things, the kludge introduced in 0.9.7m and | |
2781 | 0.9.8e for masking out AES256 independently of AES128 or masking | |
2782 | out Camellia256 independently of AES256 is not needed here in 0.9.9. | |
2783 | ||
2784 | With the change, we also introduce new ciphersuite aliases that | |
2785 | so far were missing: "AES128", "AES256", "CAMELLIA128", and | |
2786 | "CAMELLIA256". | |
2787 | [Bodo Moeller] | |
2788 | ||
357d5de5 NL |
2789 | *) Add support for dsa-with-SHA224 and dsa-with-SHA256. |
2790 | Use the leftmost N bytes of the signature input if the input is | |
2791 | larger than the prime q (with N being the size in bytes of q). | |
2792 | [Nils Larsch] | |
2793 | ||
11d8cdc6 DSH |
2794 | *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses |
2795 | it yet and it is largely untested. | |
2796 | [Steve Henson] | |
2797 | ||
06e2dd03 NL |
2798 | *) Add support for the ecdsa-with-SHA224/256/384/512 signature types. |
2799 | [Nils Larsch] | |
2800 | ||
de121164 | 2801 | *) Initial incomplete changes to avoid need for function casts in OpenSSL |
297e6f19 | 2802 | some compilers (gcc 4.2 and later) reject their use. Safestack is |
a6fbcb42 | 2803 | reimplemented. Update ASN1 to avoid use of legacy functions. |
de121164 DSH |
2804 | [Steve Henson] |
2805 | ||
3189772e AP |
2806 | *) Win32/64 targets are linked with Winsock2. |
2807 | [Andy Polyakov] | |
2808 | ||
010fa0b3 DSH |
2809 | *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected |
2810 | to external functions. This can be used to increase CRL handling | |
2811 | efficiency especially when CRLs are very large by (for example) storing | |
2812 | the CRL revoked certificates in a database. | |
2813 | [Steve Henson] | |
2814 | ||
5d20c4fb DSH |
2815 | *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so |
2816 | new CRLs added to a directory can be used. New command line option | |
2817 | -verify_return_error to s_client and s_server. This causes real errors | |
2818 | to be returned by the verify callback instead of carrying on no matter | |
2819 | what. This reflects the way a "real world" verify callback would behave. | |
2820 | [Steve Henson] | |
2821 | ||
2822 | *) GOST engine, supporting several GOST algorithms and public key formats. | |
2823 | Kindly donated by Cryptocom. | |
2824 | [Cryptocom] | |
2825 | ||
bc7535bc DSH |
2826 | *) Partial support for Issuing Distribution Point CRL extension. CRLs |
2827 | partitioned by DP are handled but no indirect CRL or reason partitioning | |
2828 | (yet). Complete overhaul of CRL handling: now the most suitable CRL is | |
2829 | selected via a scoring technique which handles IDP and AKID in CRLs. | |
2830 | [Steve Henson] | |
2831 | ||
2832 | *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which | |
2833 | will ultimately be used for all verify operations: this will remove the | |
2834 | X509_STORE dependency on certificate verification and allow alternative | |
2835 | lookup methods. X509_STORE based implementations of these two callbacks. | |
2836 | [Steve Henson] | |
2837 | ||
f6e7d014 DSH |
2838 | *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names. |
2839 | Modify get_crl() to find a valid (unexpired) CRL if possible. | |
2840 | [Steve Henson] | |
2841 | ||
edc54021 DSH |
2842 | *) New function X509_CRL_match() to check if two CRLs are identical. Normally |
2843 | this would be called X509_CRL_cmp() but that name is already used by | |
2844 | a function that just compares CRL issuer names. Cache several CRL | |
2845 | extensions in X509_CRL structure and cache CRLDP in X509. | |
2846 | [Steve Henson] | |
2847 | ||
450ea834 DSH |
2848 | *) Store a "canonical" representation of X509_NAME structure (ASN1 Name) |
2849 | this maps equivalent X509_NAME structures into a consistent structure. | |
2850 | Name comparison can then be performed rapidly using memcmp(). | |
2851 | [Steve Henson] | |
2852 | ||
454dbbc5 DSH |
2853 | *) Non-blocking OCSP request processing. Add -timeout option to ocsp |
2854 | utility. | |
c1c6c0bf DSH |
2855 | [Steve Henson] |
2856 | ||
b7683e3a DSH |
2857 | *) Allow digests to supply their own micalg string for S/MIME type using |
2858 | the ctrl EVP_MD_CTRL_MICALG. | |
2859 | [Steve Henson] | |
2860 | ||
2861 | *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the | |
2862 | EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN | |
2863 | ctrl. It can then customise the structure before and/or after signing | |
2864 | if necessary. | |
2865 | [Steve Henson] | |
2866 | ||
0ee2166c DSH |
2867 | *) New function OBJ_add_sigid() to allow application defined signature OIDs |
2868 | to be added to OpenSSLs internal tables. New function OBJ_sigid_free() | |
2869 | to free up any added signature OIDs. | |
2870 | [Steve Henson] | |
2871 | ||
5ba4bf35 DSH |
2872 | *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(), |
2873 | EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal | |
2874 | digest and cipher tables. New options added to openssl utility: | |
2875 | list-message-digest-algorithms and list-cipher-algorithms. | |
2876 | [Steve Henson] | |
2877 | ||
c4e7870a BM |
2878 | *) Change the array representation of binary polynomials: the list |
2879 | of degrees of non-zero coefficients is now terminated with -1. | |
2880 | Previously it was terminated with 0, which was also part of the | |
2881 | value; thus, the array representation was not applicable to | |
2882 | polynomials where t^0 has coefficient zero. This change makes | |
2883 | the array representation useful in a more general context. | |
2884 | [Douglas Stebila] | |
2885 | ||
89bbe14c BM |
2886 | *) Various modifications and fixes to SSL/TLS cipher string |
2887 | handling. For ECC, the code now distinguishes between fixed ECDH | |
2888 | with RSA certificates on the one hand and with ECDSA certificates | |
2889 | on the other hand, since these are separate ciphersuites. The | |
2890 | unused code for Fortezza ciphersuites has been removed. | |
2891 | ||
2892 | For consistency with EDH, ephemeral ECDH is now called "EECDH" | |
2893 | (not "ECDHE"). For consistency with the code for DH | |
2894 | certificates, use of ECDH certificates is now considered ECDH | |
2895 | authentication, not RSA or ECDSA authentication (the latter is | |
2896 | merely the CA's signing algorithm and not actively used in the | |
2897 | protocol). | |
2898 | ||
2899 | The temporary ciphersuite alias "ECCdraft" is no longer | |
2900 | available, and ECC ciphersuites are no longer excluded from "ALL" | |
2901 | and "DEFAULT". The following aliases now exist for RFC 4492 | |
2902 | ciphersuites, most of these by analogy with the DH case: | |
2903 | ||
2904 | kECDHr - ECDH cert, signed with RSA | |
2905 | kECDHe - ECDH cert, signed with ECDSA | |
2906 | kECDH - ECDH cert (signed with either RSA or ECDSA) | |
2907 | kEECDH - ephemeral ECDH | |
2908 | ECDH - ECDH cert or ephemeral ECDH | |
2909 | ||
2910 | aECDH - ECDH cert | |
2911 | aECDSA - ECDSA cert | |
2912 | ECDSA - ECDSA cert | |
2913 | ||
2914 | AECDH - anonymous ECDH | |
2915 | EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH") | |
2916 | ||
2917 | [Bodo Moeller] | |
2918 | ||
fb7b3932 DSH |
2919 | *) Add additional S/MIME capabilities for AES and GOST ciphers if supported. |
2920 | Use correct micalg parameters depending on digest(s) in signed message. | |
2921 | [Steve Henson] | |
2922 | ||
01b8b3c7 DSH |
2923 | *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process |
2924 | an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code. | |
2925 | [Steve Henson] | |
de9fcfe3 | 2926 | |
58aa573a | 2927 | *) Initial engine support for EVP_PKEY_METHOD. New functions to permit |
c9777d26 DSH |
2928 | an engine to register a method. Add ENGINE lookups for methods and |
2929 | functional reference processing. | |
58aa573a DSH |
2930 | [Steve Henson] |
2931 | ||
91c9e621 DSH |
2932 | *) New functions EVP_Digest{Sign,Verify)*. These are enchance versions of |
2933 | EVP_{Sign,Verify}* which allow an application to customise the signature | |
2934 | process. | |
2935 | [Steve Henson] | |
2936 | ||
55311921 DSH |
2937 | *) New -resign option to smime utility. This adds one or more signers |
2938 | to an existing PKCS#7 signedData structure. Also -md option to use an | |
2939 | alternative message digest algorithm for signing. | |
2940 | [Steve Henson] | |
2941 | ||
a6e7fcd1 DSH |
2942 | *) Tidy up PKCS#7 routines and add new functions to make it easier to |
2943 | create PKCS7 structures containing multiple signers. Update smime | |
2944 | application to support multiple signers. | |
2945 | [Steve Henson] | |
2946 | ||
121dd39f DSH |
2947 | *) New -macalg option to pkcs12 utility to allow setting of an alternative |
2948 | digest MAC. | |
2949 | [Steve Henson] | |
2950 | ||
856640b5 | 2951 | *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC. |
b8f702a0 | 2952 | Reorganize PBE internals to lookup from a static table using NIDs, |
6d3a1eac DSH |
2953 | add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl: |
2954 | EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative | |
2955 | PRF which will be automatically used with PBES2. | |
856640b5 DSH |
2956 | [Steve Henson] |
2957 | ||
34b3c72e | 2958 | *) Replace the algorithm specific calls to generate keys in "req" with the |
959e8dfe DSH |
2959 | new API. |
2960 | [Steve Henson] | |
2961 | ||
399a6f0b DSH |
2962 | *) Update PKCS#7 enveloped data routines to use new API. This is now |
2963 | supported by any public key method supporting the encrypt operation. A | |
2964 | ctrl is added to allow the public key algorithm to examine or modify | |
2965 | the PKCS#7 RecipientInfo structure if it needs to: for RSA this is | |
2966 | a no op. | |
2967 | [Steve Henson] | |
28e4fe34 | 2968 | |
03919683 DSH |
2969 | *) Add a ctrl to asn1 method to allow a public key algorithm to express |
2970 | a default digest type to use. In most cases this will be SHA1 but some | |
2971 | algorithms (such as GOST) need to specify an alternative digest. The | |
2972 | return value indicates how strong the prefernce is 1 means optional and | |
2973 | 2 is mandatory (that is it is the only supported type). Modify | |
2974 | ASN1_item_sign() to accept a NULL digest argument to indicate it should | |
2975 | use the default md. Update openssl utilities to use the default digest | |
2976 | type for signing if it is not explicitly indicated. | |
2977 | [Steve Henson] | |
2978 | ||
ee1d9ec0 DSH |
2979 | *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New |
2980 | EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant | |
2981 | signing method from the key type. This effectively removes the link | |
2982 | between digests and public key types. | |
2983 | [Steve Henson] | |
2984 | ||
d2027098 DSH |
2985 | *) Add an OID cross reference table and utility functions. Its purpose is to |
2986 | translate between signature OIDs such as SHA1WithrsaEncryption and SHA1, | |
2987 | rsaEncryption. This will allow some of the algorithm specific hackery | |
2988 | needed to use the correct OID to be removed. | |
2989 | [Steve Henson] | |
2990 | ||
492a9e24 DSH |
2991 | *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO |
2992 | structures for PKCS7_sign(). They are now set up by the relevant public | |
2993 | key ASN1 method. | |
2994 | [Steve Henson] | |
2995 | ||
9ca7047d DSH |
2996 | *) Add provisional EC pkey method with support for ECDSA and ECDH. |
2997 | [Steve Henson] | |
2998 | ||
ffb1ac67 DSH |
2999 | *) Add support for key derivation (agreement) in the API, DH method and |
3000 | pkeyutl. | |
3001 | [Steve Henson] | |
3002 | ||
3ba0885a DSH |
3003 | *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support |
3004 | public and private key formats. As a side effect these add additional | |
3005 | command line functionality not previously available: DSA signatures can be | |
3006 | generated and verified using pkeyutl and DH key support and generation in | |
3007 | pkey, genpkey. | |
3008 | [Steve Henson] | |
3009 | ||
4700aea9 UM |
3010 | *) BeOS support. |
3011 | [Oliver Tappe <zooey@hirschkaefer.de>] | |
3012 | ||
3013 | *) New make target "install_html_docs" installs HTML renditions of the | |
3014 | manual pages. | |
3015 | [Oliver Tappe <zooey@hirschkaefer.de>] | |
3016 | ||
f5cda4cb DSH |
3017 | *) New utility "genpkey" this is analagous to "genrsa" etc except it can |
3018 | generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to | |
3019 | support key and parameter generation and add initial key generation | |
3020 | functionality for RSA. | |
3021 | [Steve Henson] | |
3022 | ||
f733a5ef DSH |
3023 | *) Add functions for main EVP_PKEY_method operations. The undocumented |
3024 | functions EVP_PKEY_{encrypt,decrypt} have been renamed to | |
3025 | EVP_PKEY_{encrypt,decrypt}_old. | |
3026 | [Steve Henson] | |
3027 | ||
0b6f3c66 DSH |
3028 | *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public |
3029 | key API, doesn't do much yet. | |
3030 | [Steve Henson] | |
3031 | ||
0b33dac3 DSH |
3032 | *) New function EVP_PKEY_asn1_get0_info() to retrieve information about |
3033 | public key algorithms. New option to openssl utility: | |
3034 | "list-public-key-algorithms" to print out info. | |
3035 | [Steve Henson] | |
3036 | ||
33273721 BM |
3037 | *) Implement the Supported Elliptic Curves Extension for |
3038 | ECC ciphersuites from draft-ietf-tls-ecc-12.txt. | |
3039 | [Douglas Stebila] | |
3040 | ||
246e0931 DSH |
3041 | *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or |
3042 | EVP_CIPHER structures to avoid later problems in EVP_cleanup(). | |
3043 | [Steve Henson] | |
3044 | ||
3e4585c8 | 3045 | *) New utilities pkey and pkeyparam. These are similar to algorithm specific |
f5cda4cb | 3046 | utilities such as rsa, dsa, dsaparam etc except they process any key |
3e4585c8 | 3047 | type. |
3e84b6e1 DSH |
3048 | [Steve Henson] |
3049 | ||
35208f36 DSH |
3050 | *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New |
3051 | functions EVP_PKEY_print_public(), EVP_PKEY_print_private(), | |
3052 | EVP_PKEY_print_param() to print public key data from an EVP_PKEY | |
3053 | structure. | |
3054 | [Steve Henson] | |
3055 | ||
448be743 DSH |
3056 | *) Initial support for pluggable public key ASN1. |
3057 | De-spaghettify the public key ASN1 handling. Move public and private | |
3058 | key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate | |
3059 | algorithm specific handling to a single module within the relevant | |
3060 | algorithm directory. Add functions to allow (near) opaque processing | |
3061 | of public and private key structures. | |
3062 | [Steve Henson] | |
3063 | ||
36ca4ba6 BM |
3064 | *) Implement the Supported Point Formats Extension for |
3065 | ECC ciphersuites from draft-ietf-tls-ecc-12.txt. | |
3066 | [Douglas Stebila] | |
3067 | ||
ddac1974 NL |
3068 | *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members |
3069 | for the psk identity [hint] and the psk callback functions to the | |
3070 | SSL_SESSION, SSL and SSL_CTX structure. | |
3071 | ||
3072 | New ciphersuites: | |
3073 | PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA, | |
3074 | PSK-AES256-CBC-SHA | |
3075 | ||
3076 | New functions: | |
3077 | SSL_CTX_use_psk_identity_hint | |
3078 | SSL_get_psk_identity_hint | |
3079 | SSL_get_psk_identity | |
3080 | SSL_use_psk_identity_hint | |
3081 | ||
3082 | [Mika Kousa and Pasi Eronen of Nokia Corporation] | |
3083 | ||
c7235be6 UM |
3084 | *) Add RFC 3161 compliant time stamp request creation, response generation |
3085 | and response verification functionality. | |
9f0b86c6 | 3086 | [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project] |
c7235be6 | 3087 | |
1aeb3da8 BM |
3088 | *) Add initial support for TLS extensions, specifically for the server_name |
3089 | extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now | |
3090 | have new members for a host name. The SSL data structure has an | |
3091 | additional member SSL_CTX *initial_ctx so that new sessions can be | |
3092 | stored in that context to allow for session resumption, even after the | |
3093 | SSL has been switched to a new SSL_CTX in reaction to a client's | |
3094 | server_name extension. | |
f1fd4544 BM |
3095 | |
3096 | New functions (subject to change): | |
3097 | ||
3098 | SSL_get_servername() | |
3099 | SSL_get_servername_type() | |
3100 | SSL_set_SSL_CTX() | |
3101 | ||
3102 | New CTRL codes and macros (subject to change): | |
3103 | ||
3104 | SSL_CTRL_SET_TLSEXT_SERVERNAME_CB | |
3105 | - SSL_CTX_set_tlsext_servername_callback() | |
3106 | SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG | |
3107 | - SSL_CTX_set_tlsext_servername_arg() | |
fec38ca4 | 3108 | SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() |
b1277b99 | 3109 | |
241520e6 BM |
3110 | openssl s_client has a new '-servername ...' option. |
3111 | ||
3112 | openssl s_server has new options '-servername_host ...', '-cert2 ...', | |
3113 | '-key2 ...', '-servername_fatal' (subject to change). This allows | |
3114 | testing the HostName extension for a specific single host name ('-cert' | |
3115 | and '-key' remain fallbacks for handshakes without HostName | |
3116 | negotiation). If the unrecogninzed_name alert has to be sent, this by | |
3117 | default is a warning; it becomes fatal with the '-servername_fatal' | |
3118 | option. | |
b1277b99 | 3119 | |
e8e5b46e | 3120 | [Peter Sylvester, Remy Allais, Christophe Renou] |
b1277b99 | 3121 | |
ed26604a AP |
3122 | *) Whirlpool hash implementation is added. |
3123 | [Andy Polyakov] | |
3124 | ||
0cb9d93d AP |
3125 | *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to |
3126 | bn(64,32). Because of instruction set limitations it doesn't have | |
3127 | any negative impact on performance. This was done mostly in order | |
3128 | to make it possible to share assembler modules, such as bn_mul_mont | |
3129 | implementations, between 32- and 64-bit builds without hassle. | |
3130 | [Andy Polyakov] | |
3131 | ||
8dee9f84 BM |
3132 | *) Move code previously exiled into file crypto/ec/ec2_smpt.c |
3133 | to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP | |
3134 | macro. | |
3135 | [Bodo Moeller] | |
3136 | ||
4d524040 AP |
3137 | *) New candidate for BIGNUM assembler implementation, bn_mul_mont, |
3138 | dedicated Montgomery multiplication procedure, is introduced. | |
3139 | BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher | |
3140 | "64-bit" performance on certain 32-bit targets. | |
3141 | [Andy Polyakov] | |
3142 | ||
566dda07 DSH |
3143 | *) New option SSL_OP_NO_COMP to disable use of compression selectively |
3144 | in SSL structures. New SSL ctrl to set maximum send fragment size. | |
3145 | Save memory by seeting the I/O buffer sizes dynamically instead of | |
3146 | using the maximum available value. | |
3147 | [Steve Henson] | |
3148 | ||
13e4670c BM |
3149 | *) New option -V for 'openssl ciphers'. This prints the ciphersuite code |
3150 | in addition to the text details. | |
3151 | [Bodo Moeller] | |
3152 | ||
1ef7acfe DSH |
3153 | *) Very, very preliminary EXPERIMENTAL support for printing of general |
3154 | ASN1 structures. This currently produces rather ugly output and doesn't | |
3155 | handle several customised structures at all. | |
3156 | [Steve Henson] | |
3157 | ||
a0156a92 DSH |
3158 | *) Integrated support for PVK file format and some related formats such |
3159 | as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support | |
3160 | these in the 'rsa' and 'dsa' utilities. | |
3161 | [Steve Henson] | |
3162 | ||
eea374fd DSH |
3163 | *) Support for PKCS#1 RSAPublicKey format on rsa utility command line. |
3164 | [Steve Henson] | |
3165 | ||
45e27385 DSH |
3166 | *) Remove the ancient ASN1_METHOD code. This was only ever used in one |
3167 | place for the (very old) "NETSCAPE" format certificates which are now | |
3168 | handled using new ASN1 code equivalents. | |
eea374fd | 3169 | [Steve Henson] |
45e27385 | 3170 | |
4ebb342f NL |
3171 | *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD |
3172 | pointer and make the SSL_METHOD parameter in SSL_CTX_new, | |
3173 | SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'. | |
3174 | [Nils Larsch] | |
3175 | ||
9aa9d70d | 3176 | *) Modify CRL distribution points extension code to print out previously |
0745d089 DSH |
3177 | unsupported fields. Enhance extension setting code to allow setting of |
3178 | all fields. | |
9aa9d70d DSH |
3179 | [Steve Henson] |
3180 | ||
0537f968 | 3181 | *) Add print and set support for Issuing Distribution Point CRL extension. |
231493c9 | 3182 | [Steve Henson] |
28e4fe34 | 3183 | |
f3dea9a5 BM |
3184 | *) Change 'Configure' script to enable Camellia by default. |
3185 | [NTT] | |
2dc4b0db | 3186 | |
5b5464d5 BM |
3187 | Changes between 0.9.8m and 0.9.8n [24 Mar 2010] |
3188 | ||
3189 | *) When rejecting SSL/TLS records due to an incorrect version number, never | |
3190 | update s->server with a new major version number. As of | |
3191 | - OpenSSL 0.9.8m if 'short' is a 16-bit type, | |
3192 | - OpenSSL 0.9.8f if 'short' is longer than 16 bits, | |
3193 | the previous behavior could result in a read attempt at NULL when | |
3194 | receiving specific incorrect SSL/TLS records once record payload | |
4ecd2baf BM |
3195 | protection is active. (CVE-2010-0740) |
3196 | [Bodo Moeller, Adam Langley <agl@chromium.org>] | |
5b5464d5 | 3197 | |
47333a34 DSH |
3198 | *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL |
3199 | could be crashed if the relevant tables were not present (e.g. chrooted). | |
3200 | [Tomas Hoger <thoger@redhat.com>] | |
f3dea9a5 | 3201 | |
5b5464d5 BM |
3202 | Changes between 0.9.8l and 0.9.8m [25 Feb 2010] |
3203 | ||
3204 | *) Always check bn_wexpend() return values for failure. (CVE-2009-3245) | |
3205 | [Martin Olsson, Neel Mehta] | |
32567c9f BM |
3206 | |
3207 | *) Fix X509_STORE locking: Every 'objs' access requires a lock (to | |
3208 | accommodate for stack sorting, always a write lock!). | |
3209 | [Bodo Moeller] | |
2c627637 | 3210 | |
9051fc53 DSH |
3211 | *) On some versions of WIN32 Heap32Next is very slow. This can cause |
3212 | excessive delays in the RAND_poll(): over a minute. As a workaround | |
3213 | include a time check in the inner Heap32Next loop too. | |
3214 | [Steve Henson] | |
3215 | ||
57cffe90 | 3216 | *) The code that handled flushing of data in SSL/TLS originally used the |
d793c292 DSH |
3217 | BIO_CTRL_INFO ctrl to see if any data was pending first. This caused |
3218 | the problem outlined in PR#1949. The fix suggested there however can | |
3219 | trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions | |
3220 | of Apache). So instead simplify the code to flush unconditionally. | |
3221 | This should be fine since flushing with no data to flush is a no op. | |
3222 | [Steve Henson] | |
3223 | ||
41c0f686 DSH |
3224 | *) Handle TLS versions 2.0 and later properly and correctly use the |
3225 | highest version of TLS/SSL supported. Although TLS >= 2.0 is some way | |
3226 | off ancient servers have a habit of sticking around for a while... | |
3227 | [Steve Henson] | |
3228 | ||
2c627637 DSH |
3229 | *) Modify compression code so it frees up structures without using the |
3230 | ex_data callbacks. This works around a problem where some applications | |
d8f07f16 | 3231 | call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when |
2c627637 DSH |
3232 | restarting) then use compression (e.g. SSL with compression) later. |
3233 | This results in significant per-connection memory leaks and | |
3234 | has caused some security issues including CVE-2008-1678 and | |
3235 | CVE-2009-4355. | |
3236 | [Steve Henson] | |
3237 | ||
a0b72777 BM |
3238 | *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't |
3239 | change when encrypting or decrypting. | |
3240 | [Bodo Moeller] | |
3241 | ||
67556483 | 3242 | *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to |
98923880 | 3243 | connect and renegotiate with servers which do not support RI. |
67556483 DSH |
3244 | Until RI is more widely deployed this option is enabled by default. |
3245 | [Steve Henson] | |
ddcfc25a | 3246 | |
52a08e90 DSH |
3247 | *) Add "missing" ssl ctrls to clear options and mode. |
3248 | [Steve Henson] | |
3249 | ||
6b5f0458 | 3250 | *) If client attempts to renegotiate and doesn't support RI respond with |
81d87a2a DSH |
3251 | a no_renegotiation alert as required by RFC5746. Some renegotiating |
3252 | TLS clients will continue a connection gracefully when they receive | |
3253 | the alert. Unfortunately OpenSSL mishandled this alert and would hang | |
3254 | waiting for a server hello which it will never receive. Now we treat a | |
3255 | received no_renegotiation alert as a fatal error. This is because | |
3256 | applications requesting a renegotiation might well expect it to succeed | |
3257 | and would have no code in place to handle the server denying it so the | |
3258 | only safe thing to do is to terminate the connection. | |
6b5f0458 DSH |
3259 | [Steve Henson] |
3260 | ||
b52a2738 DSH |
3261 | *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if |
3262 | peer supports secure renegotiation and 0 otherwise. Print out peer | |
3263 | renegotiation support in s_client/s_server. | |
3264 | [Steve Henson] | |
3265 | ||
7b1856e5 DSH |
3266 | *) Replace the highly broken and deprecated SPKAC certification method with |
3267 | the updated NID creation version. This should correctly handle UTF8. | |
3268 | [Steve Henson] | |
3269 | ||
81d87a2a DSH |
3270 | *) Implement RFC5746. Re-enable renegotiation but require the extension |
3271 | as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION | |
3272 | turns out to be a bad idea. It has been replaced by | |
bc9058d0 DSH |
3273 | SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with |
3274 | SSL_CTX_set_options(). This is really not recommended unless you | |
3275 | know what you are doing. | |
10f99d7b | 3276 | [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson] |
bc9058d0 | 3277 | |
9ac5c355 DSH |
3278 | *) Fixes to stateless session resumption handling. Use initial_ctx when |
3279 | issuing and attempting to decrypt tickets in case it has changed during | |
3280 | servername handling. Use a non-zero length session ID when attempting | |
3281 | stateless session resumption: this makes it possible to determine if | |
aefb9dc5 | 3282 | a resumption has occurred immediately after receiving server hello |
9ac5c355 DSH |
3283 | (several places in OpenSSL subtly assume this) instead of later in |
3284 | the handshake. | |
3285 | [Steve Henson] | |
3286 | ||
80afb40a DSH |
3287 | *) The functions ENGINE_ctrl(), OPENSSL_isservice(), |
3288 | CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error | |
3289 | fixes for a few places where the return code is not checked | |
3290 | correctly. | |
3291 | [Julia Lawall <julia@diku.dk>] | |
3292 | ||
b5b65403 DSH |
3293 | *) Add --strict-warnings option to Configure script to include devteam |
3294 | warnings in other configurations. | |
3295 | [Steve Henson] | |
3296 | ||
d5ec7d66 | 3297 | *) Add support for --libdir option and LIBDIR variable in makefiles. This |
aefb9dc5 | 3298 | makes it possible to install openssl libraries in locations which |
d5ec7d66 DSH |
3299 | have names other than "lib", for example "/usr/lib64" which some |
3300 | systems need. | |
3301 | [Steve Henson, based on patch from Jeremy Utley] | |
3302 | ||
52828ca2 DSH |
3303 | *) Don't allow the use of leading 0x80 in OIDs. This is a violation of |
3304 | X690 8.9.12 and can produce some misleading textual output of OIDs. | |
3305 | [Steve Henson, reported by Dan Kaminsky] | |
3306 | ||
aefb9dc5 BM |
3307 | *) Delete MD2 from algorithm tables. This follows the recommendation in |
3308 | several standards that it is not used in new applications due to | |
3309 | several cryptographic weaknesses. For binary compatibility reasons | |
3310 | the MD2 API is still compiled in by default. | |
3311 | [Steve Henson] | |
3312 | ||
76ec9151 DSH |
3313 | *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved |
3314 | and restored. | |
3315 | [Steve Henson] | |
3316 | ||
aefb9dc5 BM |
3317 | *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and |
3318 | OPENSSL_asc2uni conditionally on Netware platforms to avoid a name | |
3319 | clash. | |
3320 | [Guenter <lists@gknw.net>] | |
3321 | ||
dbb834ff DSH |
3322 | *) Fix the server certificate chain building code to use X509_verify_cert(), |
3323 | it used to have an ad-hoc builder which was unable to cope with anything | |
3324 | other than a simple chain. | |
3325 | [David Woodhouse <dwmw2@infradead.org>, Steve Henson] | |
3326 | ||
710c1c34 DSH |
3327 | *) Don't check self signed certificate signatures in X509_verify_cert() |
3328 | by default (a flag can override this): it just wastes time without | |
3329 | adding any security. As a useful side effect self signed root CAs | |
3330 | with non-FIPS digests are now usable in FIPS mode. | |
f1ed5fa8 DSH |
3331 | [Steve Henson] |
3332 | ||
32fbeacd DSH |
3333 | *) In dtls1_process_out_of_seq_message() the check if the current message |
3334 | is already buffered was missing. For every new message was memory | |
3335 | allocated, allowing an attacker to perform an denial of service attack | |
3336 | with sending out of seq handshake messages until there is no memory | |
3337 | left. Additionally every future messege was buffered, even if the | |
3338 | sequence number made no sense and would be part of another handshake. | |
3339 | So only messages with sequence numbers less than 10 in advance will be | |
aefb9dc5 | 3340 | buffered. (CVE-2009-1378) |
32fbeacd DSH |
3341 | [Robin Seggelmann, discovered by Daniel Mentz] |
3342 | ||
3343 | *) Records are buffered if they arrive with a future epoch to be | |
3344 | processed after finishing the corresponding handshake. There is | |
3345 | currently no limitation to this buffer allowing an attacker to perform | |
3346 | a DOS attack with sending records with future epochs until there is no | |
3347 | memory left. This patch adds the pqueue_size() function to detemine | |
3348 | the size of a buffer and limits the record buffer to 100 entries. | |
aefb9dc5 | 3349 | (CVE-2009-1377) |
32fbeacd DSH |
3350 | [Robin Seggelmann, discovered by Daniel Mentz] |
3351 | ||
3352 | *) Keep a copy of frag->msg_header.frag_len so it can be used after the | |
aefb9dc5 | 3353 | parent structure is freed. (CVE-2009-1379) |
32fbeacd DSH |
3354 | [Daniel Mentz] |
3355 | ||
c184b140 DSH |
3356 | *) Handle non-blocking I/O properly in SSL_shutdown() call. |
3357 | [Darryl Miles <darryl-mailinglists@netbauds.net>] | |
3358 | ||
ddcfc25a DSH |
3359 | *) Add 2.5.4.* OIDs |
3360 | [Ilya O. <vrghost@gmail.com>] | |
3361 | ||
aefb9dc5 BM |
3362 | Changes between 0.9.8k and 0.9.8l [5 Nov 2009] |
3363 | ||
3364 | *) Disable renegotiation completely - this fixes a severe security | |
3365 | problem (CVE-2009-3555) at the cost of breaking all | |
3366 | renegotiation. Renegotiation can be re-enabled by setting | |
3367 | SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at | |
3368 | run-time. This is really not recommended unless you know what | |
3369 | you're doing. | |
3370 | [Ben Laurie] | |
3371 | ||
4d7b7c62 | 3372 | Changes between 0.9.8j and 0.9.8k [25 Mar 2009] |
64895732 | 3373 | |
73ba116e DSH |
3374 | *) Don't set val to NULL when freeing up structures, it is freed up by |
3375 | underlying code. If sizeof(void *) > sizeof(long) this can result in | |
3376 | zeroing past the valid field. (CVE-2009-0789) | |
3377 | [Paolo Ganci <Paolo.Ganci@AdNovum.CH>] | |
3378 | ||
80b2ff97 DSH |
3379 | *) Fix bug where return value of CMS_SignerInfo_verify_content() was not |
3380 | checked correctly. This would allow some invalid signed attributes to | |
3381 | appear to verify correctly. (CVE-2009-0591) | |
3382 | [Ivan Nestlerode <inestlerode@us.ibm.com>] | |
3383 | ||
7ce8c95d DSH |
3384 | *) Reject UniversalString and BMPString types with invalid lengths. This |
3385 | prevents a crash in ASN1_STRING_print_ex() which assumes the strings have | |
3386 | a legal length. (CVE-2009-0590) | |
3387 | [Steve Henson] | |
3388 | ||
237d7b6c DSH |
3389 | *) Set S/MIME signing as the default purpose rather than setting it |
3390 | unconditionally. This allows applications to override it at the store | |
3391 | level. | |
3392 | [Steve Henson] | |
3393 | ||
854a225a DSH |
3394 | *) Permit restricted recursion of ASN1 strings. This is needed in practice |
3395 | to handle some structures. | |
3396 | [Steve Henson] | |
3397 | ||
77202a85 DSH |
3398 | *) Improve efficiency of mem_gets: don't search whole buffer each time |
3399 | for a '\n' | |
3400 | [Jeremy Shapiro <jnshapir@us.ibm.com>] | |
3401 | ||
7ca1cfba BM |
3402 | *) New -hex option for openssl rand. |
3403 | [Matthieu Herrb] | |
3404 | ||
57f39cc8 DSH |
3405 | *) Print out UTF8String and NumericString when parsing ASN1. |
3406 | [Steve Henson] | |
3407 | ||
64895732 DSH |
3408 | *) Support NumericString type for name components. |
3409 | [Steve Henson] | |
aefb9dc5 | 3410 | |
7f625320 BL |
3411 | *) Allow CC in the environment to override the automatically chosen |
3412 | compiler. Note that nothing is done to ensure flags work with the | |
3413 | chosen compiler. | |
3414 | [Ben Laurie] | |
aefb9dc5 | 3415 | |
bab53405 DSH |
3416 | Changes between 0.9.8i and 0.9.8j [07 Jan 2009] |
3417 | ||
3418 | *) Properly check EVP_VerifyFinal() and similar return values | |
3419 | (CVE-2008-5077). | |
3420 | [Ben Laurie, Bodo Moeller, Google Security Team] | |
31636a3e | 3421 | |
60aee6ce BL |
3422 | *) Enable TLS extensions by default. |
3423 | [Ben Laurie] | |
3424 | ||
31636a3e | 3425 | *) Allow the CHIL engine to be loaded, whether the application is |
7a762197 BM |
3426 | multithreaded or not. (This does not release the developer from the |
3427 | obligation to set up the dynamic locking callbacks.) | |
3428 | [Sander Temme <sander@temme.net>] | |
31636a3e | 3429 | |
31636a3e GT |
3430 | *) Use correct exit code if there is an error in dgst command. |
3431 | [Steve Henson; problem pointed out by Roland Dirlewanger] | |
3432 | ||
7a762197 BM |
3433 | *) Tweak Configure so that you need to say "experimental-jpake" to enable |
3434 | JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications. | |
3435 | [Bodo Moeller] | |
3436 | ||
3437 | *) Add experimental JPAKE support, including demo authentication in | |
3438 | s_client and s_server. | |
6caa4edd BL |
3439 | [Ben Laurie] |
3440 | ||
28b6d502 BL |
3441 | *) Set the comparison function in v3_addr_canonize(). |
3442 | [Rob Austein <sra@hactrn.net>] | |
3443 | ||
d5bbead4 BL |
3444 | *) Add support for XMPP STARTTLS in s_client. |
3445 | [Philip Paeps <philip@freebsd.org>] | |
3446 | ||
837f2fc7 BM |
3447 | *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior |
3448 | to ensure that even with this option, only ciphersuites in the | |
3449 | server's preference list will be accepted. (Note that the option | |
3450 | applies only when resuming a session, so the earlier behavior was | |
3451 | just about the algorithm choice for symmetric cryptography.) | |
3452 | [Bodo Moeller] | |
3453 | ||
1a489c9a | 3454 | Changes between 0.9.8h and 0.9.8i [15 Sep 2008] |
6bf79e30 | 3455 | |
aefb9dc5 BM |
3456 | *) Fix NULL pointer dereference if a DTLS server received |
3457 | ChangeCipherSpec as first record (CVE-2009-1386). | |
3458 | [PR #1679] | |
3459 | ||
e65bcbce BM |
3460 | *) Fix a state transitition in s3_srvr.c and d1_srvr.c |
3461 | (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...). | |
3462 | [Nagendra Modadugu] | |
3463 | ||
db99c525 BM |
3464 | *) The fix in 0.9.8c that supposedly got rid of unsafe |
3465 | double-checked locking was incomplete for RSA blinding, | |
3466 | addressing just one layer of what turns out to have been | |
3467 | doubly unsafe triple-checked locking. | |
3468 | ||
3469 | So now fix this for real by retiring the MONT_HELPER macro | |
3470 | in crypto/rsa/rsa_eay.c. | |
3471 | ||
3472 | [Bodo Moeller; problem pointed out by Marius Schilder] | |
3473 | ||
f8d6be3f BM |
3474 | *) Various precautionary measures: |
3475 | ||
3476 | - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h). | |
3477 | ||
3478 | - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c). | |
3479 | (NB: This would require knowledge of the secret session ticket key | |
3480 | to exploit, in which case you'd be SOL either way.) | |
3481 | ||
3482 | - Change bn_nist.c so that it will properly handle input BIGNUMs | |
3483 | outside the expected range. | |
3484 | ||
3485 | - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG | |
3486 | builds. | |
3487 | ||
3488 | [Neel Mehta, Bodo Moeller] | |
3489 | ||
1a489c9a BM |
3490 | *) Allow engines to be "soft loaded" - i.e. optionally don't die if |
3491 | the load fails. Useful for distros. | |
3492 | [Ben Laurie and the FreeBSD team] | |
3493 | ||
8528128b DSH |
3494 | *) Add support for Local Machine Keyset attribute in PKCS#12 files. |
3495 | [Steve Henson] | |
3496 | ||
8228fd89 BM |
3497 | *) Fix BN_GF2m_mod_arr() top-bit cleanup code. |
3498 | [Huang Ying] | |
3499 | ||
6bf79e30 | 3500 | *) Expand ENGINE to support engine supplied SSL client certificate functions. |
adb92d56 DSH |
3501 | |
3502 | This work was sponsored by Logica. | |
6bf79e30 DSH |
3503 | [Steve Henson] |
3504 | ||
8228fd89 BM |
3505 | *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows |
3506 | keystores. Support for SSL/TLS client authentication too. | |
6bf79e30 | 3507 | Not compiled unless enable-capieng specified to Configure. |
adb92d56 DSH |
3508 | |
3509 | This work was sponsored by Logica. | |
6bf79e30 DSH |
3510 | [Steve Henson] |
3511 | ||
1a489c9a BM |
3512 | *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using |
3513 | ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain | |
3514 | attribute creation routines such as certifcate requests and PKCS#12 | |
3515 | files. | |
3516 | [Steve Henson] | |
db99c525 | 3517 | |
2cd81830 | 3518 | Changes between 0.9.8g and 0.9.8h [28 May 2008] |
94fd382f | 3519 | |
e194fe8f BM |
3520 | *) Fix flaw if 'Server Key exchange message' is omitted from a TLS |
3521 | handshake which could lead to a cilent crash as found using the | |
3522 | Codenomicon TLS test suite (CVE-2008-1672) | |
3523 | [Steve Henson, Mark Cox] | |
3524 | ||
40a70628 BM |
3525 | *) Fix double free in TLS server name extensions which could lead to |
3526 | a remote crash found by Codenomicon TLS test suite (CVE-2008-0891) | |
3527 | [Joe Orton] | |
3528 | ||
c2c2e7a4 LJ |
3529 | *) Clear error queue in SSL_CTX_use_certificate_chain_file() |
3530 | ||
3531 | Clear the error queue to ensure that error entries left from | |
3532 | older function calls do not interfere with the correct operation. | |
3533 | [Lutz Jaenicke, Erik de Castro Lopo] | |
3534 | ||
d18ef847 LJ |
3535 | *) Remove root CA certificates of commercial CAs: |
3536 | ||
3537 | The OpenSSL project does not recommend any specific CA and does not | |
3538 | have any policy with respect to including or excluding any CA. | |
3539 | Therefore it does not make any sense to ship an arbitrary selection | |
3540 | of root CA certificates with the OpenSSL software. | |
3541 | [Lutz Jaenicke] | |
3542 | ||
94fd382f DSH |
3543 | *) RSA OAEP patches to fix two separate invalid memory reads. |
3544 | The first one involves inputs when 'lzero' is greater than | |
3545 | 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes | |
3546 | before the beginning of from). The second one involves inputs where | |
3547 | the 'db' section contains nothing but zeroes (there is a one-byte | |
3548 | invalid read after the end of 'db'). | |
5c0d90a6 | 3549 | [Ivan Nestlerode <inestlerode@us.ibm.com>] |
db99c525 BM |
3550 | |
3551 | *) Partial backport from 0.9.9-dev: | |
3552 | ||
3553 | Introduce bn_mul_mont (dedicated Montgomery multiplication | |
3554 | procedure) as a candidate for BIGNUM assembler implementation. | |
3555 | While 0.9.9-dev uses assembler for various architectures, only | |
3556 | x86_64 is available by default here in the 0.9.8 branch, and | |
3557 | 32-bit x86 is available through a compile-time setting. | |
3558 | ||
3559 | To try the 32-bit x86 assembler implementation, use Configure | |
3560 | option "enable-montasm" (which exists only for this backport). | |
3561 | ||
3562 | As "enable-montasm" for 32-bit x86 disclaims code stability | |
3563 | anyway, in this constellation we activate additional code | |
3564 | backported from 0.9.9-dev for further performance improvements, | |
3565 | namely BN_from_montgomery_word. (To enable this otherwise, | |
3566 | e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".) | |
3567 | ||
3568 | [Andy Polyakov (backport partially by Bodo Moeller)] | |
3569 | ||
8a2062fe DSH |
3570 | *) Add TLS session ticket callback. This allows an application to set |
3571 | TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed | |
3572 | values. This is useful for key rollover for example where several key | |
3573 | sets may exist with different names. | |
3574 | [Steve Henson] | |
a6db6a00 | 3575 | |
e7b097f5 GT |
3576 | *) Reverse ENGINE-internal logic for caching default ENGINE handles. |
3577 | This was broken until now in 0.9.8 releases, such that the only way | |
3578 | a registered ENGINE could be used (assuming it initialises | |
3579 | successfully on the host) was to explicitly set it as the default | |
3580 | for the relevant algorithms. This is in contradiction with 0.9.7 | |
3581 | behaviour and the documentation. With this fix, when an ENGINE is | |
3582 | registered into a given algorithm's table of implementations, the | |
3583 | 'uptodate' flag is reset so that auto-discovery will be used next | |
3584 | time a new context for that algorithm attempts to select an | |
3585 | implementation. | |
3586 | [Ian Lister (tweaked by Geoff Thorpe)] | |
3587 | ||
db99c525 BM |
3588 | *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9 |
3589 | implemention in the following ways: | |
3590 | ||
3591 | Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be | |
3592 | hard coded. | |
3593 | ||
3594 | Lack of BER streaming support means one pass streaming processing is | |
3595 | only supported if data is detached: setting the streaming flag is | |
3596 | ignored for embedded content. | |
3597 | ||
3598 | CMS support is disabled by default and must be explicitly enabled | |
3599 | with the enable-cms configuration option. | |
3600 | [Steve Henson] | |
3601 | ||
5ee6f96c GT |
3602 | *) Update the GMP engine glue to do direct copies between BIGNUM and |
3603 | mpz_t when openssl and GMP use the same limb size. Otherwise the | |
3604 | existing "conversion via a text string export" trick is still used. | |
db99c525 | 3605 | [Paul Sheer <paulsheer@gmail.com>] |
5ee6f96c | 3606 | |
3df93571 DSH |
3607 | *) Zlib compression BIO. This is a filter BIO which compressed and |
3608 | uncompresses any data passed through it. | |
3609 | [Steve Henson] | |
3610 | ||
992e92a4 DSH |
3611 | *) Add AES_wrap_key() and AES_unwrap_key() functions to implement |
3612 | RFC3394 compatible AES key wrapping. | |
3613 | [Steve Henson] | |
3614 | ||
3615 | *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0(): | |
3616 | sets string data without copying. X509_ALGOR_set0() and | |
3617 | X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier) | |
3618 | data. Attribute function X509at_get0_data_by_OBJ(): retrieves data | |
3619 | from an X509_ATTRIBUTE structure optionally checking it occurs only | |
3620 | once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied | |
3621 | data. | |
3622 | [Steve Henson] | |
3623 | ||
7c9882eb BM |
3624 | *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set() |
3625 | to get the expected BN_FLG_CONSTTIME behavior. | |
3626 | [Bodo Moeller (Google)] | |
3627 | ||
76d761cc DSH |
3628 | *) Netware support: |
3629 | ||
3630 | - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets | |
3631 | - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT) | |
3632 | - added some more tests to do_tests.pl | |
3633 | - fixed RunningProcess usage so that it works with newer LIBC NDKs too | |
3634 | - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency | |
3635 | - added new Configure targets netware-clib-bsdsock, netware-clib-gcc, | |
3636 | netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc | |
3637 | - various changes to netware.pl to enable gcc-cross builds on Win32 | |
3638 | platform | |
3639 | - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD) | |
3640 | - various changes to fix missing prototype warnings | |
3641 | - fixed x86nasm.pl to create correct asm files for NASM COFF output | |
3642 | - added AES, WHIRLPOOL and CPUID assembler code to build files | |
3643 | - added missing AES assembler make rules to mk1mf.pl | |
3644 | - fixed order of includes in apps/ocsp.c so that e_os.h settings apply | |
3645 | [Guenter Knauf <eflash@gmx.net>] | |
3646 | ||
a6db6a00 DSH |
3647 | *) Implement certificate status request TLS extension defined in RFC3546. |
3648 | A client can set the appropriate parameters and receive the encoded | |
3649 | OCSP response via a callback. A server can query the supplied parameters | |
3650 | and set the encoded OCSP response in the callback. Add simplified examples | |
3651 | to s_client and s_server. | |
3652 | [Steve Henson] | |
3653 | ||
11d01d37 LJ |
3654 | Changes between 0.9.8f and 0.9.8g [19 Oct 2007] |
3655 | ||
3656 | *) Fix various bugs: | |
3657 | + Binary incompatibility of ssl_ctx_st structure | |
3658 | + DTLS interoperation with non-compliant servers | |
3659 | + Don't call get_session_cb() without proposed session | |
3660 | + Fix ia64 assembler code | |
3661 | [Andy Polyakov, Steve Henson] | |
3662 | ||
a6db6a00 | 3663 | Changes between 0.9.8e and 0.9.8f [11 Oct 2007] |
0f32c841 | 3664 | |
0d89e456 AP |
3665 | *) DTLS Handshake overhaul. There were longstanding issues with |
3666 | OpenSSL DTLS implementation, which were making it impossible for | |
3667 | RFC 4347 compliant client to communicate with OpenSSL server. | |
3668 | Unfortunately just fixing these incompatibilities would "cut off" | |
3669 | pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e | |
3670 | server keeps tolerating non RFC compliant syntax. The opposite is | |
3671 | not true, 0.9.8f client can not communicate with earlier server. | |
3672 | This update even addresses CVE-2007-4995. | |
3673 | [Andy Polyakov] | |
3674 | ||
3675 | *) Changes to avoid need for function casts in OpenSSL: some compilers | |
3676 | (gcc 4.2 and later) reject their use. | |
3677 | [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>, | |
3678 | Steve Henson] | |
3679 | ||
3680 | *) Add RFC4507 support to OpenSSL. This includes the corrections in | |
3681 | RFC4507bis. The encrypted ticket format is an encrypted encoded | |
3682 | SSL_SESSION structure, that way new session features are automatically | |
3683 | supported. | |
3684 | ||
3685 | If a client application caches session in an SSL_SESSION structure | |
3686 | support is transparent because tickets are now stored in the encoded | |
3687 | SSL_SESSION. | |
3688 | ||
3689 | The SSL_CTX structure automatically generates keys for ticket | |
3690 | protection in servers so again support should be possible | |
3691 | with no application modification. | |
3692 | ||
3693 | If a client or server wishes to disable RFC4507 support then the option | |
3694 | SSL_OP_NO_TICKET can be set. | |
3695 | ||
3696 | Add a TLS extension debugging callback to allow the contents of any client | |
3697 | or server extensions to be examined. | |
3698 | ||
3699 | This work was sponsored by Google. | |
3700 | [Steve Henson] | |
3701 | ||
3702 | *) Add initial support for TLS extensions, specifically for the server_name | |
3703 | extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now | |
3704 | have new members for a host name. The SSL data structure has an | |
3705 | additional member SSL_CTX *initial_ctx so that new sessions can be | |
3706 | stored in that context to allow for session resumption, even after the | |
3707 | SSL has been switched to a new SSL_CTX in reaction to a client's | |
3708 | server_name extension. | |
3709 | ||
3710 | New functions (subject to change): | |
3711 | ||
3712 | SSL_get_servername() | |
3713 | SSL_get_servername_type() | |
3714 | SSL_set_SSL_CTX() | |
3715 | ||
3716 | New CTRL codes and macros (subject to change): | |
3717 | ||
3718 | SSL_CTRL_SET_TLSEXT_SERVERNAME_CB | |
3719 | - SSL_CTX_set_tlsext_servername_callback() | |
3720 | SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG | |
3721 | - SSL_CTX_set_tlsext_servername_arg() | |
3722 | SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name() | |
3723 | ||
3724 | openssl s_client has a new '-servername ...' option. | |
3725 | ||
3726 | openssl s_server has new options '-servername_host ...', '-cert2 ...', | |
3727 | '-key2 ...', '-servername_fatal' (subject to change). This allows | |
3728 | testing the HostName extension for a specific single host name ('-cert' | |
3729 | and '-key' remain fallbacks for handshakes without HostName | |
3730 | negotiation). If the unrecogninzed_name alert has to be sent, this by | |
3731 | default is a warning; it becomes fatal with the '-servername_fatal' | |
3732 | option. | |
3733 | ||
3734 | [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson] | |
3735 | ||
3736 | *) Add AES and SSE2 assembly language support to VC++ build. | |
3737 | [Steve Henson] | |
3738 | ||
85a5668d AP |
3739 | *) Mitigate attack on final subtraction in Montgomery reduction. |
3740 | [Andy Polyakov] | |
3741 | ||
19f6c524 BM |
3742 | *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0 |
3743 | (which previously caused an internal error). | |
3744 | [Bodo Moeller] | |
3745 | ||
69ab0852 BL |
3746 | *) Squeeze another 10% out of IGE mode when in != out. |
3747 | [Ben Laurie] | |
3748 | ||
5f09d0ec BL |
3749 | *) AES IGE mode speedup. |
3750 | [Dean Gaudet (Google)] | |
3751 | ||
96afc1cf BM |
3752 | *) Add the Korean symmetric 128-bit cipher SEED (see |
3753 | http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and | |
3754 | add SEED ciphersuites from RFC 4162: | |
3755 | ||
3756 | TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA" | |
3757 | TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA" | |
3758 | TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA" | |
3759 | TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA" | |
3760 | ||
3761 | To minimize changes between patchlevels in the OpenSSL 0.9.8 | |
3762 | series, SEED remains excluded from compilation unless OpenSSL | |
3763 | is configured with 'enable-seed'. | |
3764 | [KISA, Bodo Moeller] | |
3765 | ||
bd31fb21 BM |
3766 | *) Mitigate branch prediction attacks, which can be practical if a |
3767 | single processor is shared, allowing a spy process to extract | |
3768 | information. For detailed background information, see | |
3769 | http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron, | |
3770 | J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL | |
3771 | and Necessary Software Countermeasures"). The core of the change | |
3772 | are new versions BN_div_no_branch() and | |
3773 | BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(), | |
3774 | respectively, which are slower, but avoid the security-relevant | |
3775 | conditional branches. These are automatically called by BN_div() | |
b002265e BM |
3776 | and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one |
3777 | of the input BIGNUMs. Also, BN_is_bit_set() has been changed to | |
3778 | remove a conditional branch. | |
bd31fb21 BM |
3779 | |
3780 | BN_FLG_CONSTTIME is the new name for the previous | |
3781 | BN_FLG_EXP_CONSTTIME flag, since it now affects more than just | |
3782 | modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag | |
3783 | in the exponent causes BN_mod_exp_mont() to use the alternative | |
3784 | implementation in BN_mod_exp_mont_consttime().) The old name | |
3785 | remains as a deprecated alias. | |
3786 | ||
3787 | Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general | |
3788 | RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses | |
3789 | constant-time implementations for more than just exponentiation. | |
3790 | Here too the old name is kept as a deprecated alias. | |
3791 | ||
3792 | BN_BLINDING_new() will now use BN_dup() for the modulus so that | |
3793 | the BN_BLINDING structure gets an independent copy of the | |
3794 | modulus. This means that the previous "BIGNUM *m" argument to | |
3795 | BN_BLINDING_new() and to BN_BLINDING_create_param() now | |
3796 | essentially becomes "const BIGNUM *m", although we can't actually | |
3797 | change this in the header file before 0.9.9. It allows | |
3798 | RSA_setup_blinding() to use BN_with_flags() on the modulus to | |
3799 | enable BN_FLG_CONSTTIME. | |
3800 | ||
3801 | [Matthew D Wood (Intel Corp)] | |
3802 | ||
0f32c841 BM |
3803 | *) In the SSL/TLS server implementation, be strict about session ID |
3804 | context matching (which matters if an application uses a single | |
3805 | external cache for different purposes). Previously, | |
3806 | out-of-context reuse was forbidden only if SSL_VERIFY_PEER was | |
3807 | set. This did ensure strict client verification, but meant that, | |
3808 | with applications using a single external cache for quite | |
3809 | different requirements, clients could circumvent ciphersuite | |
3810 | restrictions for a given session ID context by starting a session | |
3811 | in a different context. | |
3812 | [Bodo Moeller] | |
61118caa | 3813 | |
0a05123a BM |
3814 | *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that |
3815 | a ciphersuite string such as "DEFAULT:RSA" cannot enable | |
3816 | authentication-only ciphersuites. | |
3817 | [Bodo Moeller] | |
3818 | ||
db99c525 BM |
3819 | *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was |
3820 | not complete and could lead to a possible single byte overflow | |
3821 | (CVE-2007-5135) [Ben Laurie] | |
3822 | ||
0f32c841 BM |
3823 | Changes between 0.9.8d and 0.9.8e [23 Feb 2007] |
3824 | ||
52b8dad8 BM |
3825 | *) Since AES128 and AES256 (and similarly Camellia128 and |
3826 | Camellia256) share a single mask bit in the logic of | |
3827 | ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a | |
3828 | kludge to work properly if AES128 is available and AES256 isn't | |
3829 | (or if Camellia128 is available and Camellia256 isn't). | |
3830 | [Victor Duchovni] | |
3831 | ||
772e3c07 BM |
3832 | *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c |
3833 | (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters): | |
3834 | When a point or a seed is encoded in a BIT STRING, we need to | |
3835 | prevent the removal of trailing zero bits to get the proper DER | |
3836 | encoding. (By default, crypto/asn1/a_bitstr.c assumes the case | |
3837 | of a NamedBitList, for which trailing 0 bits need to be removed.) | |
3838 | [Bodo Moeller] | |
3839 | ||
1e24b3a0 BM |
3840 | *) Have SSL/TLS server implementation tolerate "mismatched" record |
3841 | protocol version while receiving ClientHello even if the | |
3842 | ClientHello is fragmented. (The server can't insist on the | |
3843 | particular protocol version it has chosen before the ServerHello | |
3844 | message has informed the client about his choice.) | |
3845 | [Bodo Moeller] | |
3846 | ||
96ea4ae9 BL |
3847 | *) Add RFC 3779 support. |
3848 | [Rob Austein for ARIN, Ben Laurie] | |
3849 | ||
1e24b3a0 BM |
3850 | *) Load error codes if they are not already present instead of using a |
3851 | static variable. This allows them to be cleanly unloaded and reloaded. | |
3852 | Improve header file function name parsing. | |
3853 | [Steve Henson] | |
3854 | ||
8d72476e LJ |
3855 | *) extend SMTP and IMAP protocol emulation in s_client to use EHLO |
3856 | or CAPABILITY handshake as required by RFCs. | |
3857 | [Goetz Babin-Ebell] | |
3858 | ||
61118caa | 3859 | Changes between 0.9.8c and 0.9.8d [28 Sep 2006] |
b79aa05e | 3860 | |
3ff55e96 MC |
3861 | *) Introduce limits to prevent malicious keys being able to |
3862 | cause a denial of service. (CVE-2006-2940) | |
3863 | [Steve Henson, Bodo Moeller] | |
3864 | ||
3865 | *) Fix ASN.1 parsing of certain invalid structures that can result | |
3866 | in a denial of service. (CVE-2006-2937) [Steve Henson] | |
3867 | ||
3868 | *) Fix buffer overflow in SSL_get_shared_ciphers() function. | |
3869 | (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] | |
3870 | ||
3871 | *) Fix SSL client code which could crash if connecting to a | |
3872 | malicious SSLv2 server. (CVE-2006-4343) | |
3873 | [Tavis Ormandy and Will Drewry, Google Security Team] | |
3874 | ||
ed65f7dc BM |
3875 | *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites |
3876 | match only those. Before that, "AES256-SHA" would be interpreted | |
b6699c3f BM |
3877 | as a pattern and match "AES128-SHA" too (since AES128-SHA got |
3878 | the same strength classification in 0.9.7h) as we currently only | |
ed65f7dc BM |
3879 | have a single AES bit in the ciphersuite description bitmap. |
3880 | That change, however, also applied to ciphersuite strings such as | |
3881 | "RC4-MD5" that intentionally matched multiple ciphersuites -- | |
3882 | namely, SSL 2.0 ciphersuites in addition to the more common ones | |
3883 | from SSL 3.0/TLS 1.0. | |
3884 | ||
3885 | So we change the selection algorithm again: Naming an explicit | |
3886 | ciphersuite selects this one ciphersuite, and any other similar | |
3887 | ciphersuite (same bitmap) from *other* protocol versions. | |
3888 | Thus, "RC4-MD5" again will properly select both the SSL 2.0 | |
3889 | ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite. | |
3890 | ||
3891 | Since SSL 2.0 does not have any ciphersuites for which the | |
3892 | 128/256 bit distinction would be relevant, this works for now. | |
3893 | The proper fix will be to use different bits for AES128 and | |
3894 | AES256, which would have avoided the problems from the beginning; | |
3895 | however, bits are scarce, so we can only do this in a new release | |
3896 | (not just a patchlevel) when we can change the SSL_CIPHER | |
3897 | definition to split the single 'unsigned long mask' bitmap into | |
3898 | multiple values to extend the available space. | |
3899 | ||
3900 | [Bodo Moeller] | |
3901 | ||
b79aa05e MC |
3902 | Changes between 0.9.8b and 0.9.8c [05 Sep 2006] |
3903 | ||
3904 | *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher | |
3905 | (CVE-2006-4339) [Ben Laurie and Google Security Team] | |
5cda6c45 | 3906 | |
aa6d1a0c BL |
3907 | *) Add AES IGE and biIGE modes. |
3908 | [Ben Laurie] | |
3909 | ||
e34aa5a3 BM |
3910 | *) Change the Unix randomness entropy gathering to use poll() when |
3911 | possible instead of select(), since the latter has some | |
3912 | undesirable limitations. | |
3913 | [Darryl Miles via Richard Levitte and Bodo Moeller] | |
3914 | ||
81de1028 BM |
3915 | *) Disable "ECCdraft" ciphersuites more thoroughly. Now special |
3916 | treatment in ssl/ssl_ciph.s makes sure that these ciphersuites | |
3917 | cannot be implicitly activated as part of, e.g., the "AES" alias. | |
3918 | However, please upgrade to OpenSSL 0.9.9[-dev] for | |
3919 | non-experimental use of the ECC ciphersuites to get TLS extension | |
3920 | support, which is required for curve and point format negotiation | |
3921 | to avoid potential handshake problems. | |
850815cb BM |
3922 | [Bodo Moeller] |
3923 | ||
5b57fe0a BM |
3924 | *) Disable rogue ciphersuites: |
3925 | ||
3926 | - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") | |
3927 | - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") | |
3928 | - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") | |
3929 | ||
3930 | The latter two were purportedly from | |
3931 | draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really | |
3932 | appear there. | |
3933 | ||
fec38ca4 | 3934 | Also deactivate the remaining ciphersuites from |
5b57fe0a BM |
3935 | draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as |
3936 | unofficial, and the ID has long expired. | |
3937 | [Bodo Moeller] | |
3938 | ||
675f605d BM |
3939 | *) Fix RSA blinding Heisenbug (problems sometimes occured on |
3940 | dual-core machines) and other potential thread-safety issues. | |
3941 | [Bodo Moeller] | |
3942 | ||
f3dea9a5 BM |
3943 | *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key |
3944 | versions), which is now available for royalty-free use | |
3945 | (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html). | |
3946 | Also, add Camellia TLS ciphersuites from RFC 4132. | |
3947 | ||
3948 | To minimize changes between patchlevels in the OpenSSL 0.9.8 | |
3949 | series, Camellia remains excluded from compilation unless OpenSSL | |
3950 | is configured with 'enable-camellia'. | |
3951 | [NTT] | |
3952 | ||
5cda6c45 DSH |
3953 | *) Disable the padding bug check when compression is in use. The padding |
3954 | bug check assumes the first packet is of even length, this is not | |
3955 | necessarily true if compresssion is enabled and can result in false | |
3956 | positives causing handshake failure. The actual bug test is ancient | |
3957 | code so it is hoped that implementations will either have fixed it by | |
3958 | now or any which still have the bug do not support compression. | |
3959 | [Steve Henson] | |
3960 | ||
3961 | Changes between 0.9.8a and 0.9.8b [04 May 2006] | |
998ac55e | 3962 | |
ba1ba5f0 DSH |
3963 | *) When applying a cipher rule check to see if string match is an explicit |
3964 | cipher suite and only match that one cipher suite if it is. | |
3965 | [Steve Henson] | |
3966 | ||
31676a35 DSH |
3967 | *) Link in manifests for VC++ if needed. |
3968 | [Austin Ziegler <halostatue@gmail.com>] | |
3969 | ||
d56349a2 | 3970 | *) Update support for ECC-based TLS ciphersuites according to |
ed4a1d12 BM |
3971 | draft-ietf-tls-ecc-12.txt with proposed changes (but without |
3972 | TLS extensions, which are supported starting with the 0.9.9 | |
3973 | branch, not in the OpenSSL 0.9.8 branch). | |
d56349a2 BM |
3974 | [Douglas Stebila] |
3975 | ||
b40228a6 DSH |
3976 | *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support |
3977 | opaque EVP_CIPHER_CTX handling. | |
3978 | [Steve Henson] | |
3979 | ||
ad2695b1 DSH |
3980 | *) Fixes and enhancements to zlib compression code. We now only use |
3981 | "zlib1.dll" and use the default __cdecl calling convention on Win32 | |
3982 | to conform with the standards mentioned here: | |
3983 | http://www.zlib.net/DLL_FAQ.txt | |
3984 | Static zlib linking now works on Windows and the new --with-zlib-include | |
3985 | --with-zlib-lib options to Configure can be used to supply the location | |
3986 | of the headers and library. Gracefully handle case where zlib library | |
3987 | can't be loaded. | |
3988 | [Steve Henson] | |
3989 | ||
452ae49d DSH |
3990 | *) Several fixes and enhancements to the OID generation code. The old code |
3991 | sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't | |
3992 | handle numbers larger than ULONG_MAX, truncated printing and had a | |
3993 | non standard OBJ_obj2txt() behaviour. | |
3994 | [Steve Henson] | |
3995 | ||
fbf002bb DSH |
3996 | *) Add support for building of engines under engine/ as shared libraries |
3997 | under VC++ build system. | |
3998 | [Steve Henson] | |
3999 | ||
998ac55e RL |
4000 | *) Corrected the numerous bugs in the Win32 path splitter in DSO. |
4001 | Hopefully, we will not see any false combination of paths any more. | |
4002 | [Richard Levitte] | |
4003 | ||
d357be38 MC |
4004 | Changes between 0.9.8 and 0.9.8a [11 Oct 2005] |
4005 | ||
4006 | *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING | |
4007 | (part of SSL_OP_ALL). This option used to disable the | |
4008 | countermeasure against man-in-the-middle protocol-version | |
4009 | rollback in the SSL 2.0 server implementation, which is a bad | |
04fac373 | 4010 | idea. (CVE-2005-2969) |
d357be38 MC |
4011 | |
4012 | [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center | |
4013 | for Information Security, National Institute of Advanced Industrial | |
4014 | Science and Technology [AIST], Japan)] | |
2bd2cd9b | 4015 | |
f022c177 DSH |
4016 | *) Add two function to clear and return the verify parameter flags. |
4017 | [Steve Henson] | |
4018 | ||
6e119bb0 NL |
4019 | *) Keep cipherlists sorted in the source instead of sorting them at |
4020 | runtime, thus removing the need for a lock. | |
4021 | [Nils Larsch] | |
4022 | ||
770bc596 | 4023 | *) Avoid some small subgroup attacks in Diffie-Hellman. |
bf3d6c0c BL |
4024 | [Nick Mathewson and Ben Laurie] |
4025 | ||
4026 | *) Add functions for well-known primes. | |
4027 | [Nick Mathewson] | |
4028 | ||
0491e058 AP |
4029 | *) Extended Windows CE support. |
4030 | [Satoshi Nakamura and Andy Polyakov] | |
a1006c37 | 4031 | |
f3b656b2 DSH |
4032 | *) Initialize SSL_METHOD structures at compile time instead of during |
4033 | runtime, thus removing the need for a lock. | |
4034 | [Steve Henson] | |
4035 | ||
8f2e4fdf DSH |
4036 | *) Make PKCS7_decrypt() work even if no certificate is supplied by |
4037 | attempting to decrypt each encrypted key in turn. Add support to | |
4038 | smime utility. | |
4039 | [Steve Henson] | |
2bd2cd9b RL |
4040 | |
4041 | Changes between 0.9.7h and 0.9.8 [05 Jul 2005] | |
12bdb643 | 4042 | |
675f605d BM |
4043 | [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after |
4044 | OpenSSL 0.9.8.] | |
4045 | ||
c8310124 RL |
4046 | *) Add libcrypto.pc and libssl.pc for those who feel they need them. |
4047 | [Richard Levitte] | |
4048 | ||
4049 | *) Change CA.sh and CA.pl so they don't bundle the CSR and the private | |
4050 | key into the same file any more. | |
4051 | [Richard Levitte] | |
4052 | ||
8d3509b9 AP |
4053 | *) Add initial support for Win64, both IA64 and AMD64/x64 flavors. |
4054 | [Andy Polyakov] | |
4055 | ||
cbdac46d DSH |
4056 | *) Add -utf8 command line and config file option to 'ca'. |
4057 | [Stefan <stf@udoma.org] | |
4058 | ||
c8310124 RL |
4059 | *) Removed the macro des_crypt(), as it seems to conflict with some |
4060 | libraries. Use DES_crypt(). | |
4061 | [Richard Levitte] | |
4062 | ||
a2c32e2d GT |
4063 | *) Correct naming of the 'chil' and '4758cca' ENGINEs. This |
4064 | involves renaming the source and generated shared-libs for | |
4065 | both. The engines will accept the corrected or legacy ids | |
4066 | ('ncipher' and '4758_cca' respectively) when binding. NB, | |
4067 | this only applies when building 'shared'. | |
4068 | [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe] | |
4069 | ||
b6995add DSH |
4070 | *) Add attribute functions to EVP_PKEY structure. Modify |
4071 | PKCS12_create() to recognize a CSP name attribute and | |
4072 | use it. Make -CSP option work again in pkcs12 utility. | |
4073 | [Steve Henson] | |
4074 | ||
800e400d NL |
4075 | *) Add new functionality to the bn blinding code: |
4076 | - automatic re-creation of the BN_BLINDING parameters after | |
4077 | a fixed number of uses (currently 32) | |
4078 | - add new function for parameter creation | |
4079 | - introduce flags to control the update behaviour of the | |
4080 | BN_BLINDING parameters | |
4081 | - hide BN_BLINDING structure | |
4082 | Add a second BN_BLINDING slot to the RSA structure to improve | |
4083 | performance when a single RSA object is shared among several | |
4084 | threads. | |
4085 | [Nils Larsch] | |
4086 | ||
36d16f8e BL |
4087 | *) Add support for DTLS. |
4088 | [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie] | |
4089 | ||
dc0ed30c NL |
4090 | *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1) |
4091 | to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file() | |
4092 | [Walter Goulet] | |
4093 | ||
6049399b NL |
4094 | *) Remove buggy and incompletet DH cert support from |
4095 | ssl/ssl_rsa.c and ssl/s3_both.c | |
4096 | [Nils Larsch] | |
4097 | ||
12bdb643 NL |
4098 | *) Use SHA-1 instead of MD5 as the default digest algorithm for |
4099 | the apps/openssl applications. | |
4100 | [Nils Larsch] | |
4d94ae00 | 4101 | |
41a15c4f BL |
4102 | *) Compile clean with "-Wall -Wmissing-prototypes |
4103 | -Wstrict-prototypes -Wmissing-declarations -Werror". Currently | |
4104 | DEBUG_SAFESTACK must also be set. | |
4105 | [Ben Laurie] | |
4106 | ||
c9a112f5 | 4107 | *) Change ./Configure so that certain algorithms can be disabled by default. |
ecc5ef87 BM |
4108 | The new counterpiece to "no-xxx" is "enable-xxx". |
4109 | ||
4110 | The patented RC5 and MDC2 algorithms will now be disabled unless | |
4111 | "enable-rc5" and "enable-mdc2", respectively, are specified. | |
4112 | ||
4113 | (IDEA remains enabled despite being patented. This is because IDEA | |
4114 | is frequently required for interoperability, and there is no license | |
4115 | fee for non-commercial use. As before, "no-idea" can be used to | |
4116 | avoid this algorithm.) | |
4117 | ||
c9a112f5 BM |
4118 | [Bodo Moeller] |
4119 | ||
6951c23a RL |
4120 | *) Add processing of proxy certificates (see RFC 3820). This work was |
4121 | sponsored by KTH (The Royal Institute of Technology in Stockholm) and | |
4122 | EGEE (Enabling Grids for E-science in Europe). | |
4123 | [Richard Levitte] | |
4124 | ||
ea681ba8 AP |
4125 | *) RC4 performance overhaul on modern architectures/implementations, such |
4126 | as Intel P4, IA-64 and AMD64. | |
4127 | [Andy Polyakov] | |
4128 | ||
401ee37a DSH |
4129 | *) New utility extract-section.pl. This can be used specify an alternative |
4130 | section number in a pod file instead of having to treat each file as | |
4131 | a separate case in Makefile. This can be done by adding two lines to the | |
4132 | pod file: | |
4133 | ||
4134 | =for comment openssl_section:XXX | |
4135 | ||
4136 | The blank line is mandatory. | |
4137 | ||
4138 | [Steve Henson] | |
4139 | ||
826a42a0 DSH |
4140 | *) New arguments -certform, -keyform and -pass for s_client and s_server |
4141 | to allow alternative format key and certificate files and passphrase | |
4142 | sources. | |
4143 | [Steve Henson] | |
4144 | ||
5d7c222d DSH |
4145 | *) New structure X509_VERIFY_PARAM which combines current verify parameters, |
4146 | update associated structures and add various utility functions. | |
4147 | ||
4148 | Add new policy related verify parameters, include policy checking in | |
4149 | standard verify code. Enhance 'smime' application with extra parameters | |
4150 | to support policy checking and print out. | |
4151 | [Steve Henson] | |
4152 | ||
30fe028f GT |
4153 | *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3 |
4154 | Nehemiah processors. These extensions support AES encryption in hardware | |
4155 | as well as RNG (though RNG support is currently disabled). | |
4156 | [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov] | |
4157 | ||
df11e1e9 GT |
4158 | *) Deprecate BN_[get|set]_params() functions (they were ignored internally). |
4159 | [Geoff Thorpe] | |
4160 | ||
ad500340 AP |
4161 | *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented. |
4162 | [Andy Polyakov and a number of other people] | |
4163 | ||
e14f4aab AP |
4164 | *) Improved PowerPC platform support. Most notably BIGNUM assembler |
4165 | implementation contributed by IBM. | |
4166 | [Suresh Chari, Peter Waltenberg, Andy Polyakov] | |
4167 | ||
bcfea9fb GT |
4168 | *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public |
4169 | exponent rather than 'unsigned long'. There is a corresponding change to | |
4170 | the new 'rsa_keygen' element of the RSA_METHOD structure. | |
4171 | [Jelte Jansen, Geoff Thorpe] | |
4172 | ||
d5f686d8 BM |
4173 | *) Functionality for creating the initial serial number file is now |
4174 | moved from CA.pl to the 'ca' utility with a new option -create_serial. | |
4175 | ||
4176 | (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial | |
4177 | number file to 1, which is bound to cause problems. To avoid | |
4178 | the problems while respecting compatibility between different 0.9.7 | |
4179 | patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in | |
4180 | CA.pl for serial number initialization. With the new release 0.9.8, | |
4181 | we can fix the problem directly in the 'ca' utility.) | |
64674bcc DSH |
4182 | [Steve Henson] |
4183 | ||
3a87a9b9 GT |
4184 | *) Reduced header interdepencies by declaring more opaque objects in |
4185 | ossl_typ.h. As a consequence, including some headers (eg. engine.h) will | |
4186 | give fewer recursive includes, which could break lazy source code - so | |
4187 | this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always, | |
4188 | developers should define this symbol when building and using openssl to | |
4189 | ensure they track the recommended behaviour, interfaces, [etc], but | |
4190 | backwards-compatible behaviour prevails when this isn't defined. | |
4191 | [Geoff Thorpe] | |
4192 | ||
bf5773fa DSH |
4193 | *) New function X509_POLICY_NODE_print() which prints out policy nodes. |
4194 | [Steve Henson] | |
4195 | ||
216659eb DSH |
4196 | *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality. |
4197 | This will generate a random key of the appropriate length based on the | |
4198 | cipher context. The EVP_CIPHER can provide its own random key generation | |
4199 | routine to support keys of a specific form. This is used in the des and | |
4200 | 3des routines to generate a key of the correct parity. Update S/MIME | |
4201 | code to use new functions and hence generate correct parity DES keys. | |
4202 | Add EVP_CHECK_DES_KEY #define to return an error if the key is not | |
4203 | valid (weak or incorrect parity). | |
4204 | [Steve Henson] | |
4205 | ||
e1a27eb3 DSH |
4206 | *) Add a local set of CRLs that can be used by X509_verify_cert() as well |
4207 | as looking them up. This is useful when the verified structure may contain | |
4208 | CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs | |
4209 | present unless the new PKCS7_NO_CRL flag is asserted. | |
4210 | [Steve Henson] | |
4211 | ||
6446e0c3 DSH |
4212 | *) Extend ASN1 oid configuration module. It now additionally accepts the |
4213 | syntax: | |
4214 | ||
4215 | shortName = some long name, 1.2.3.4 | |
4216 | [Steve Henson] | |
4217 | ||
5c98b2ca GT |
4218 | *) Reimplemented the BN_CTX implementation. There is now no more static |
4219 | limitation on the number of variables it can handle nor the depth of the | |
4220 | "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack | |
4221 | information can now expand as required, and rather than having a single | |
4222 | static array of bignums, BN_CTX now uses a linked-list of such arrays | |
4223 | allowing it to expand on demand whilst maintaining the usefulness of | |
4224 | BN_CTX's "bundling". | |
4225 | [Geoff Thorpe] | |
4226 | ||
46ef873f GT |
4227 | *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD |
4228 | to allow all RSA operations to function using a single BN_CTX. | |
4229 | [Geoff Thorpe] | |
4230 | ||
4acc3e90 DSH |
4231 | *) Preliminary support for certificate policy evaluation and checking. This |
4232 | is initially intended to pass the tests outlined in "Conformance Testing | |
4233 | of Relying Party Client Certificate Path Processing Logic" v1.07. | |
4234 | [Steve Henson] | |
4235 | ||
7f663ce4 GT |
4236 | *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and |
4237 | remained unused and not that useful. A variety of other little bignum | |
4238 | tweaks and fixes have also been made continuing on from the audit (see | |
4239 | below). | |
4240 | [Geoff Thorpe] | |
4241 | ||
875a644a RL |
4242 | *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with |
4243 | associated ASN1, EVP and SSL functions and old ASN1 macros. | |
7f663ce4 | 4244 | [Richard Levitte] |
875a644a | 4245 | |
b6358c89 GT |
4246 | *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results, |
4247 | and this should never fail. So the return value from the use of | |
4248 | BN_set_word() (which can fail due to needless expansion) is now deprecated; | |
4249 | if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro. | |
4250 | [Geoff Thorpe] | |
4251 | ||
9e051bac GT |
4252 | *) BN_CTX_get() should return zero-valued bignums, providing the same |
4253 | initialised value as BN_new(). | |
9f0b86c6 | 4254 | [Geoff Thorpe, suggested by Ulf Möller] |
9e051bac | 4255 | |
edec614e DSH |
4256 | *) Support for inhibitAnyPolicy certificate extension. |
4257 | [Steve Henson] | |
4258 | ||
d870740c GT |
4259 | *) An audit of the BIGNUM code is underway, for which debugging code is |
4260 | enabled when BN_DEBUG is defined. This makes stricter enforcements on what | |
4261 | is considered valid when processing BIGNUMs, and causes execution to | |
4262 | assert() when a problem is discovered. If BN_DEBUG_RAND is defined, | |
4263 | further steps are taken to deliberately pollute unused data in BIGNUM | |
4264 | structures to try and expose faulty code further on. For now, openssl will | |
4265 | (in its default mode of operation) continue to tolerate the inconsistent | |
4266 | forms that it has tolerated in the past, but authors and packagers should | |
4267 | consider trying openssl and their own applications when compiled with | |
4268 | these debugging symbols defined. It will help highlight potential bugs in | |
4269 | their own code, and will improve the test coverage for OpenSSL itself. At | |
4270 | some point, these tighter rules will become openssl's default to improve | |
4271 | maintainability, though the assert()s and other overheads will remain only | |
4272 | in debugging configurations. See bn.h for more details. | |
9f0b86c6 | 4273 | [Geoff Thorpe, Nils Larsch, Ulf Möller] |
d870740c | 4274 | |
2ce90b9b GT |
4275 | *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure |
4276 | that can only be obtained through BN_CTX_new() (which implicitly | |
4277 | initialises it). The presence of this function only made it possible | |
4278 | to overwrite an existing structure (and cause memory leaks). | |
4279 | [Geoff Thorpe] | |
4280 | ||
8dc344cc GT |
4281 | *) Because of the callback-based approach for implementing LHASH as a |
4282 | template type, lh_insert() adds opaque objects to hash-tables and | |
4283 | lh_doall() or lh_doall_arg() are typically used with a destructor callback | |
4284 | to clean up those corresponding objects before destroying the hash table | |
4285 | (and losing the object pointers). So some over-zealous constifications in | |
4286 | LHASH have been relaxed so that lh_insert() does not take (nor store) the | |
4287 | objects as "const" and the lh_doall[_arg] callback wrappers are not | |
4288 | prototyped to have "const" restrictions on the object pointers they are | |
4289 | given (and so aren't required to cast them away any more). | |
4290 | [Geoff Thorpe] | |
4291 | ||
0991f070 GT |
4292 | *) The tmdiff.h API was so ugly and minimal that our own timing utility |
4293 | (speed) prefers to use its own implementation. The two implementations | |
4294 | haven't been consolidated as yet (volunteers?) but the tmdiff API has had | |
4295 | its object type properly exposed (MS_TM) instead of casting to/from "char | |
4296 | *". This may still change yet if someone realises MS_TM and "ms_time_***" | |
4297 | aren't necessarily the greatest nomenclatures - but this is what was used | |
4298 | internally to the implementation so I've used that for now. | |
4299 | [Geoff Thorpe] | |
4300 | ||
9d473aa2 | 4301 | *) Ensure that deprecated functions do not get compiled when |
2aaec9cc GT |
4302 | OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of |
4303 | the self-tests were still using deprecated key-generation functions so | |
4304 | these have been updated also. | |
9d473aa2 GT |
4305 | [Geoff Thorpe] |
4306 | ||
c5a55463 DSH |
4307 | *) Reorganise PKCS#7 code to separate the digest location functionality |
4308 | into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest(). | |
4309 | New function PKCS7_set_digest() to set the digest type for PKCS#7 | |
4310 | digestedData type. Add additional code to correctly generate the | |
4311 | digestedData type and add support for this type in PKCS7 initialization | |
4312 | functions. | |
8d9086df DSH |
4313 | [Steve Henson] |
4314 | ||
c5a55463 DSH |
4315 | *) New function PKCS7_set0_type_other() this initializes a PKCS7 |
4316 | structure of type "other". | |
8d9086df DSH |
4317 | [Steve Henson] |
4318 | ||
6bd27f86 RE |
4319 | *) Fix prime generation loop in crypto/bn/bn_prime.pl by making |
4320 | sure the loop does correctly stop and breaking ("division by zero") | |
4321 | modulus operations are not performed. The (pre-generated) prime | |
4322 | table crypto/bn/bn_prime.h was already correct, but it could not be | |
4323 | re-generated on some platforms because of the "division by zero" | |
4324 | situation in the script. | |
4325 | [Ralf S. Engelschall] | |
4326 | ||
968766ca BM |
4327 | *) Update support for ECC-based TLS ciphersuites according to |
4328 | draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with | |
4329 | SHA-1 now is only used for "small" curves (where the | |
4330 | representation of a field element takes up to 24 bytes); for | |
4331 | larger curves, the field element resulting from ECDH is directly | |
4332 | used as premaster secret. | |
4333 | [Douglas Stebila (Sun Microsystems Laboratories)] | |
4334 | ||
652ae06b BM |
4335 | *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2 |
4336 | curve secp160r1 to the tests. | |
4337 | [Douglas Stebila (Sun Microsystems Laboratories)] | |
4338 | ||
e666c459 | 4339 | *) Add the possibility to load symbols globally with DSO. |
9f0b86c6 | 4340 | [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte] |
e666c459 | 4341 | |
54f64516 RL |
4342 | *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better |
4343 | control of the error stack. | |
4344 | [Richard Levitte] | |
4345 | ||
3bbb0212 RL |
4346 | *) Add support for STORE in ENGINE. |
4347 | [Richard Levitte] | |
4348 | ||
a5db6fa5 RL |
4349 | *) Add the STORE type. The intention is to provide a common interface |
4350 | to certificate and key stores, be they simple file-based stores, or | |
4351 | HSM-type store, or LDAP stores, or... | |
4352 | NOTE: The code is currently UNTESTED and isn't really used anywhere. | |
4353 | [Richard Levitte] | |
4354 | ||
535fba49 RL |
4355 | *) Add a generic structure called OPENSSL_ITEM. This can be used to |
4356 | pass a list of arguments to any function as well as provide a way | |
4357 | for a function to pass data back to the caller. | |
4358 | [Richard Levitte] | |
4359 | ||
1ae0a83b RL |
4360 | *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup() |
4361 | works like BUF_strdup() but can be used to duplicate a portion of | |
4362 | a string. The copy gets NUL-terminated. BUF_memdup() duplicates | |
4363 | a memory area. | |
4364 | [Richard Levitte] | |
4365 | ||
9d6c32d6 RL |
4366 | *) Add the function sk_find_ex() which works like sk_find(), but will |
4367 | return an index to an element even if an exact match couldn't be | |
4368 | found. The index is guaranteed to point at the element where the | |
4369 | searched-for key would be inserted to preserve sorting order. | |
4370 | [Richard Levitte] | |
4371 | ||
ea5240a5 RL |
4372 | *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but |
4373 | takes an extra flags argument for optional functionality. Currently, | |
4374 | the following flags are defined: | |
4375 | ||
4376 | OBJ_BSEARCH_VALUE_ON_NOMATCH | |
4377 | This one gets OBJ_bsearch_ex() to return a pointer to the first | |
4378 | element where the comparing function returns a negative or zero | |
4379 | number. | |
4380 | ||
4381 | OBJ_BSEARCH_FIRST_VALUE_ON_MATCH | |
4382 | This one gets OBJ_bsearch_ex() to return a pointer to the first | |
4383 | element where the comparing function returns zero. This is useful | |
4384 | if there are more than one element where the comparing function | |
4385 | returns zero. | |
9d6c32d6 | 4386 | [Richard Levitte] |
ea5240a5 | 4387 | |
16b1b035 RL |
4388 | *) Make it possible to create self-signed certificates with 'openssl ca' |
4389 | in such a way that the self-signed certificate becomes part of the | |
4390 | CA database and uses the same mechanisms for serial number generation | |
4391 | as all other certificate signing. The new flag '-selfsign' enables | |
4392 | this functionality. Adapt CA.sh and CA.pl.in. | |
4393 | [Richard Levitte] | |
4394 | ||
e6526fbf RL |
4395 | *) Add functionality to check the public key of a certificate request |
4396 | against a given private. This is useful to check that a certificate | |
4397 | request can be signed by that key (self-signing). | |
4398 | [Richard Levitte] | |
4399 | ||
f85b68cd RL |
4400 | *) Make it possible to have multiple active certificates with the same |
4401 | subject in the CA index file. This is done only if the keyword | |
4402 | 'unique_subject' is set to 'no' in the main CA section (default | |
4403 | if 'CA_default') of the configuration file. The value is saved | |
4404 | with the database itself in a separate index attribute file, | |
4405 | named like the index file with '.attr' appended to the name. | |
4406 | [Richard Levitte] | |
4407 | ||
1a15c899 DSH |
4408 | *) Generate muti valued AVAs using '+' notation in config files for |
4409 | req and dirName. | |
4410 | [Steve Henson] | |
4411 | ||
520b76ff DSH |
4412 | *) Support for nameConstraints certificate extension. |
4413 | [Steve Henson] | |
4414 | ||
f80153e2 DSH |
4415 | *) Support for policyConstraints certificate extension. |
4416 | [Steve Henson] | |
4417 | ||
a1d12dae DSH |
4418 | *) Support for policyMappings certificate extension. |
4419 | [Steve Henson] | |
4420 | ||
879650b8 GT |
4421 | *) Make sure the default DSA_METHOD implementation only uses its |
4422 | dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL, | |
4423 | and change its own handlers to be NULL so as to remove unnecessary | |
4424 | indirection. This lets alternative implementations fallback to the | |
4425 | default implementation more easily. | |
4426 | [Geoff Thorpe] | |
4427 | ||
f0dc08e6 DSH |
4428 | *) Support for directoryName in GeneralName related extensions |
4429 | in config files. | |
4430 | [Steve Henson] | |
4431 | ||
132eaa59 RL |
4432 | *) Make it possible to link applications using Makefile.shared. |
4433 | Make that possible even when linking against static libraries! | |
4434 | [Richard Levitte] | |
4435 | ||
27068df7 DSH |
4436 | *) Support for single pass processing for S/MIME signing. This now |
4437 | means that S/MIME signing can be done from a pipe, in addition | |
4438 | cleartext signing (multipart/signed type) is effectively streaming | |
4439 | and the signed data does not need to be all held in memory. | |
4440 | ||
e9ec6396 | 4441 | This is done with a new flag PKCS7_STREAM. When this flag is set |
27068df7 DSH |
4442 | PKCS7_sign() only initializes the PKCS7 structure and the actual signing |
4443 | is done after the data is output (and digests calculated) in | |
4444 | SMIME_write_PKCS7(). | |
4445 | [Steve Henson] | |
4446 | ||
2d3de726 RL |
4447 | *) Add full support for -rpath/-R, both in shared libraries and |
4448 | applications, at least on the platforms where it's known how | |
4449 | to do it. | |
4450 | [Richard Levitte] | |
4451 | ||
37c660ff | 4452 | *) In crypto/ec/ec_mult.c, implement fast point multiplication with |
24893ca9 | 4453 | precomputation, based on wNAF splitting: EC_GROUP_precompute_mult() |
37c660ff | 4454 | will now compute a table of multiples of the generator that |
24893ca9 | 4455 | makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul() |
37c660ff BM |
4456 | faster (notably in the case of a single point multiplication, |
4457 | scalar * generator). | |
4458 | [Nils Larsch, Bodo Moeller] | |
4459 | ||
4e5d3a7f DSH |
4460 | *) IPv6 support for certificate extensions. The various extensions |
4461 | which use the IP:a.b.c.d can now take IPv6 addresses using the | |
4462 | formats of RFC1884 2.2 . IPv6 addresses are now also displayed | |
4463 | correctly. | |
4464 | [Steve Henson] | |
4465 | ||
96f7065f GT |
4466 | *) Added an ENGINE that implements RSA by performing private key |
4467 | exponentiations with the GMP library. The conversions to and from | |
4468 | GMP's mpz_t format aren't optimised nor are any montgomery forms | |
4469 | cached, and on x86 it appears OpenSSL's own performance has caught up. | |
4470 | However there are likely to be other architectures where GMP could | |
4471 | provide a boost. This ENGINE is not built in by default, but it can be | |
4472 | specified at Configure time and should be accompanied by the necessary | |
4473 | linker additions, eg; | |
4474 | ./config -DOPENSSL_USE_GMP -lgmp | |
4475 | [Geoff Thorpe] | |
4476 | ||
4477 | *) "openssl engine" will not display ENGINE/DSO load failure errors when | |
4478 | testing availability of engines with "-t" - the old behaviour is | |
4479 | produced by increasing the feature's verbosity with "-tt". | |
4480 | [Geoff Thorpe] | |
4481 | ||
a74333f9 LJ |
4482 | *) ECDSA routines: under certain error conditions uninitialized BN objects |
4483 | could be freed. Solution: make sure initialization is performed early | |
4484 | enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de> | |
4485 | via PR#459) | |
4486 | [Lutz Jaenicke] | |
4487 | ||
0e4aa0d2 GT |
4488 | *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD |
4489 | and DH_METHOD (eg. by ENGINE implementations) to override the normal | |
4490 | software implementations. For DSA and DH, parameter generation can | |
4491 | also be overriden by providing the appropriate method callbacks. | |
4492 | [Geoff Thorpe] | |
4493 | ||
e9224c71 GT |
4494 | *) Change the "progress" mechanism used in key-generation and |
4495 | primality testing to functions that take a new BN_GENCB pointer in | |
4496 | place of callback/argument pairs. The new API functions have "_ex" | |
4497 | postfixes and the older functions are reimplemented as wrappers for | |
4498 | the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide | |
4499 | declarations of the old functions to help (graceful) attempts to | |
4500 | migrate to the new functions. Also, the new key-generation API | |
4501 | functions operate on a caller-supplied key-structure and return | |
4502 | success/failure rather than returning a key or NULL - this is to | |
4503 | help make "keygen" another member function of RSA_METHOD etc. | |
9d5390a0 BM |
4504 | |
4505 | Example for using the new callback interface: | |
4506 | ||
4507 | int (*my_callback)(int a, int b, BN_GENCB *cb) = ...; | |
4508 | void *my_arg = ...; | |
4509 | BN_GENCB my_cb; | |
4510 | ||
4511 | BN_GENCB_set(&my_cb, my_callback, my_arg); | |
4512 | ||
4513 | return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb); | |
4514 | /* For the meaning of a, b in calls to my_callback(), see the | |
4515 | * documentation of the function that calls the callback. | |
4516 | * cb will point to my_cb; my_arg can be retrieved as cb->arg. | |
4517 | * my_callback should return 1 if it wants BN_is_prime_ex() | |
4518 | * to continue, or 0 to stop. | |
4519 | */ | |
4520 | ||
e9224c71 GT |
4521 | [Geoff Thorpe] |
4522 | ||
fdaea9ed RL |
4523 | *) Change the ZLIB compression method to be stateful, and make it |
4524 | available to TLS with the number defined in | |
4525 | draft-ietf-tls-compression-04.txt. | |
4526 | [Richard Levitte] | |
4527 | ||
20199ca8 RL |
4528 | *) Add the ASN.1 structures and functions for CertificatePair, which |
4529 | is defined as follows (according to X.509_4thEditionDraftV6.pdf): | |
4530 | ||
4531 | CertificatePair ::= SEQUENCE { | |
9d5390a0 BM |
4532 | forward [0] Certificate OPTIONAL, |
4533 | reverse [1] Certificate OPTIONAL, | |
4534 | -- at least one of the pair shall be present -- } | |
20199ca8 RL |
4535 | |
4536 | Also implement the PEM functions to read and write certificate | |
4537 | pairs, and defined the PEM tag as "CERTIFICATE PAIR". | |
4538 | ||
4539 | This needed to be defined, mostly for the sake of the LDAP | |
4540 | attribute crossCertificatePair, but may prove useful elsewhere as | |
4541 | well. | |
4542 | [Richard Levitte] | |
4543 | ||
6f17f16f RL |
4544 | *) Make it possible to inhibit symlinking of shared libraries in |
4545 | Makefile.shared, for Cygwin's sake. | |
4546 | [Richard Levitte] | |
4547 | ||
ff22e913 NL |
4548 | *) Extend the BIGNUM API by creating a function |
4549 | void BN_set_negative(BIGNUM *a, int neg); | |
4550 | and a macro that behave like | |
4551 | int BN_is_negative(const BIGNUM *a); | |
b53e44e5 | 4552 | |
ff22e913 NL |
4553 | to avoid the need to access 'a->neg' directly in applications. |
4554 | [Nils Larsch] | |
b53e44e5 | 4555 | |
5c6bf031 BM |
4556 | *) Implement fast modular reduction for pseudo-Mersenne primes |
4557 | used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c). | |
4558 | EC_GROUP_new_curve_GFp() will now automatically use this | |
4559 | if applicable. | |
4560 | [Nils Larsch <nla@trustcenter.de>] | |
4561 | ||
19b8d06a BM |
4562 | *) Add new lock type (CRYPTO_LOCK_BN). |
4563 | [Bodo Moeller] | |
4564 | ||
6f7c2cb3 RL |
4565 | *) Change the ENGINE framework to automatically load engines |
4566 | dynamically from specific directories unless they could be | |
4567 | found to already be built in or loaded. Move all the | |
4568 | current engines except for the cryptodev one to a new | |
4569 | directory engines/. | |
4570 | The engines in engines/ are built as shared libraries if | |
4571 | the "shared" options was given to ./Configure or ./config. | |
4572 | Otherwise, they are inserted in libcrypto.a. | |
4573 | /usr/local/ssl/engines is the default directory for dynamic | |
874fee47 RL |
4574 | engines, but that can be overriden at configure time through |
4575 | the usual use of --prefix and/or --openssldir, and at run | |
4576 | time with the environment variable OPENSSL_ENGINES. | |
6f7c2cb3 RL |
4577 | [Geoff Thorpe and Richard Levitte] |
4578 | ||
30afcc07 RL |
4579 | *) Add Makefile.shared, a helper makefile to build shared |
4580 | libraries. Addapt Makefile.org. | |
4581 | [Richard Levitte] | |
4582 | ||
fc6a6a10 DSH |
4583 | *) Add version info to Win32 DLLs. |
4584 | [Peter 'Luna' Runestig" <peter@runestig.com>] | |
4585 | ||
9a48b07e DSH |
4586 | *) Add new 'medium level' PKCS#12 API. Certificates and keys |
4587 | can be added using this API to created arbitrary PKCS#12 | |
4588 | files while avoiding the low level API. | |
4589 | ||
4590 | New options to PKCS12_create(), key or cert can be NULL and | |
4591 | will then be omitted from the output file. The encryption | |
4592 | algorithm NIDs can be set to -1 for no encryption, the mac | |
4593 | iteration count can be set to 0 to omit the mac. | |
4594 | ||
4595 | Enhance pkcs12 utility by making the -nokeys and -nocerts | |
4596 | options work when creating a PKCS#12 file. New option -nomac | |
4597 | to omit the mac, NONE can be set for an encryption algorithm. | |
4598 | New code is modified to use the enhanced PKCS12_create() | |
4599 | instead of the low level API. | |
4600 | [Steve Henson] | |
4601 | ||
230fd6b7 DSH |
4602 | *) Extend ASN1 encoder to support indefinite length constructed |
4603 | encoding. This can output sequences tags and octet strings in | |
4604 | this form. Modify pk7_asn1.c to support indefinite length | |
4605 | encoding. This is experimental and needs additional code to | |
4606 | be useful, such as an ASN1 bio and some enhanced streaming | |
4607 | PKCS#7 code. | |
4608 | ||
4609 | Extend template encode functionality so that tagging is passed | |
4610 | down to the template encoder. | |
4611 | [Steve Henson] | |
4612 | ||
9226e218 BM |
4613 | *) Let 'openssl req' fail if an argument to '-newkey' is not |
4614 | recognized instead of using RSA as a default. | |
4615 | [Bodo Moeller] | |
4616 | ||
ea262260 BM |
4617 | *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt. |
4618 | As these are not official, they are not included in "ALL"; | |
4619 | the "ECCdraft" ciphersuite group alias can be used to select them. | |
4620 | [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)] | |
4621 | ||
e172d60d BM |
4622 | *) Add ECDH engine support. |
4623 | [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)] | |
4624 | ||
4625 | *) Add ECDH in new directory crypto/ecdh/. | |
4626 | [Douglas Stebila (Sun Microsystems Laboratories)] | |
4627 | ||
95ecacf8 BM |
4628 | *) Let BN_rand_range() abort with an error after 100 iterations |
4629 | without success (which indicates a broken PRNG). | |
4630 | [Bodo Moeller] | |
4631 | ||
6fb60a84 BM |
4632 | *) Change BN_mod_sqrt() so that it verifies that the input value |
4633 | is really the square of the return value. (Previously, | |
4634 | BN_mod_sqrt would show GIGO behaviour.) | |
4635 | [Bodo Moeller] | |
4636 | ||
7793f30e BM |
4637 | *) Add named elliptic curves over binary fields from X9.62, SECG, |
4638 | and WAP/WTLS; add OIDs that were still missing. | |
4639 | ||
4640 | [Sheueling Chang Shantz and Douglas Stebila | |
4641 | (Sun Microsystems Laboratories)] | |
4642 | ||
4643 | *) Extend the EC library for elliptic curves over binary fields | |
4644 | (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/). | |
4645 | New EC_METHOD: | |
4646 | ||
4647 | EC_GF2m_simple_method | |
4648 | ||
4649 | New API functions: | |
4650 | ||
4651 | EC_GROUP_new_curve_GF2m | |
4652 | EC_GROUP_set_curve_GF2m | |
4653 | EC_GROUP_get_curve_GF2m | |
7793f30e BM |
4654 | EC_POINT_set_affine_coordinates_GF2m |
4655 | EC_POINT_get_affine_coordinates_GF2m | |
4656 | EC_POINT_set_compressed_coordinates_GF2m | |
4657 | ||
4658 | Point compression for binary fields is disabled by default for | |
4659 | patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to | |
4660 | enable it). | |
4661 | ||
4662 | As binary polynomials are represented as BIGNUMs, various members | |
4663 | of the EC_GROUP and EC_POINT data structures can be shared | |
4664 | between the implementations for prime fields and binary fields; | |
4665 | the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m) | |
4666 | are essentially identical to their ..._GFp counterparts. | |
9e4f9b36 BM |
4667 | (For simplicity, the '..._GFp' prefix has been dropped from |
4668 | various internal method names.) | |
7793f30e BM |
4669 | |
4670 | An internal 'field_div' method (similar to 'field_mul' and | |
4671 | 'field_sqr') has been added; this is used only for binary fields. | |
4672 | ||
4673 | [Sheueling Chang Shantz and Douglas Stebila | |
4674 | (Sun Microsystems Laboratories)] | |
4675 | ||
9e4f9b36 | 4676 | *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult() |
7793f30e BM |
4677 | through methods ('mul', 'precompute_mult'). |
4678 | ||
4679 | The generic implementations (now internally called 'ec_wNAF_mul' | |
4680 | and 'ec_wNAF_precomputed_mult') remain the default if these | |
4681 | methods are undefined. | |
4682 | ||
4683 | [Sheueling Chang Shantz and Douglas Stebila | |
4684 | (Sun Microsystems Laboratories)] | |
4685 | ||
4686 | *) New function EC_GROUP_get_degree, which is defined through | |
4687 | EC_METHOD. For curves over prime fields, this returns the bit | |
4688 | length of the modulus. | |
4689 | ||
4690 | [Sheueling Chang Shantz and Douglas Stebila | |
4691 | (Sun Microsystems Laboratories)] | |
4692 | ||
4693 | *) New functions EC_GROUP_dup, EC_POINT_dup. | |
4694 | (These simply call ..._new and ..._copy). | |
4695 | ||
4696 | [Sheueling Chang Shantz and Douglas Stebila | |
4697 | (Sun Microsystems Laboratories)] | |
4698 | ||
1dc920c8 BM |
4699 | *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c. |
4700 | Polynomials are represented as BIGNUMs (where the sign bit is not | |
4701 | used) in the following functions [macros]: | |
4702 | ||
4703 | BN_GF2m_add | |
4704 | BN_GF2m_sub [= BN_GF2m_add] | |
4705 | BN_GF2m_mod [wrapper for BN_GF2m_mod_arr] | |
4706 | BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr] | |
4707 | BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr] | |
4708 | BN_GF2m_mod_inv | |
4709 | BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr] | |
4710 | BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr] | |
4711 | BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr] | |
4712 | BN_GF2m_cmp [= BN_ucmp] | |
4713 | ||
4714 | (Note that only the 'mod' functions are actually for fields GF(2^m). | |
4715 | BN_GF2m_add() is misnomer, but this is for the sake of consistency.) | |
4716 | ||
4717 | For some functions, an the irreducible polynomial defining a | |
4718 | field can be given as an 'unsigned int[]' with strictly | |
4719 | decreasing elements giving the indices of those bits that are set; | |
4720 | i.e., p[] represents the polynomial | |
4721 | f(t) = t^p[0] + t^p[1] + ... + t^p[k] | |
4722 | where | |
4723 | p[0] > p[1] > ... > p[k] = 0. | |
4724 | This applies to the following functions: | |
4725 | ||
4726 | BN_GF2m_mod_arr | |
4727 | BN_GF2m_mod_mul_arr | |
4728 | BN_GF2m_mod_sqr_arr | |
4729 | BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv] | |
4730 | BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div] | |
4731 | BN_GF2m_mod_exp_arr | |
4732 | BN_GF2m_mod_sqrt_arr | |
4733 | BN_GF2m_mod_solve_quad_arr | |
4734 | BN_GF2m_poly2arr | |
4735 | BN_GF2m_arr2poly | |
4736 | ||
4737 | Conversion can be performed by the following functions: | |
4738 | ||
4739 | BN_GF2m_poly2arr | |
4740 | BN_GF2m_arr2poly | |
4741 | ||
4742 | bntest.c has additional tests for binary polynomial arithmetic. | |
4743 | ||
909abce8 BM |
4744 | Two implementations for BN_GF2m_mod_div() are available. |
4745 | The default algorithm simply uses BN_GF2m_mod_inv() and | |
4746 | BN_GF2m_mod_mul(). The alternative algorithm is compiled in only | |
4747 | if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the | |
4748 | copyright notice in crypto/bn/bn_gf2m.c before enabling it). | |
1dc920c8 BM |
4749 | |
4750 | [Sheueling Chang Shantz and Douglas Stebila | |
4751 | (Sun Microsystems Laboratories)] | |
4752 | ||
16dc1cfb BM |
4753 | *) Add new error code 'ERR_R_DISABLED' that can be used when some |
4754 | functionality is disabled at compile-time. | |
4755 | [Douglas Stebila <douglas.stebila@sun.com>] | |
4756 | ||
ea4f109c BM |
4757 | *) Change default behaviour of 'openssl asn1parse' so that more |
4758 | information is visible when viewing, e.g., a certificate: | |
4759 | ||
4760 | Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump' | |
4761 | mode the content of non-printable OCTET STRINGs is output in a | |
4762 | style similar to INTEGERs, but with '[HEX DUMP]' prepended to | |
4763 | avoid the appearance of a printable string. | |
4764 | [Nils Larsch <nla@trustcenter.de>] | |
4765 | ||
254ef80d BM |
4766 | *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access |
4767 | functions | |
4768 | EC_GROUP_set_asn1_flag() | |
4769 | EC_GROUP_get_asn1_flag() | |
4770 | EC_GROUP_set_point_conversion_form() | |
4771 | EC_GROUP_get_point_conversion_form() | |
4772 | These control ASN1 encoding details: | |
b8e0e123 BM |
4773 | - Curves (i.e., groups) are encoded explicitly unless asn1_flag |
4774 | has been set to OPENSSL_EC_NAMED_CURVE. | |
5f3d6f70 | 4775 | - Points are encoded in uncompressed form by default; options for |
254ef80d BM |
4776 | asn1_for are as for point2oct, namely |
4777 | POINT_CONVERSION_COMPRESSED | |
4778 | POINT_CONVERSION_UNCOMPRESSED | |
4779 | POINT_CONVERSION_HYBRID | |
5f3d6f70 BM |
4780 | |
4781 | Also add 'seed' and 'seed_len' members to EC_GROUP with access | |
4782 | functions | |
4783 | EC_GROUP_set_seed() | |
4784 | EC_GROUP_get0_seed() | |
4785 | EC_GROUP_get_seed_len() | |
4786 | This is used only for ASN1 purposes (so far). | |
458c2917 BM |
4787 | [Nils Larsch <nla@trustcenter.de>] |
4788 | ||
4789 | *) Add 'field_type' member to EC_METHOD, which holds the NID | |
4790 | of the appropriate field type OID. The new function | |
4791 | EC_METHOD_get_field_type() returns this value. | |
4792 | [Nils Larsch <nla@trustcenter.de>] | |
4793 | ||
6cbe6382 BM |
4794 | *) Add functions |
4795 | EC_POINT_point2bn() | |
4796 | EC_POINT_bn2point() | |
4797 | EC_POINT_point2hex() | |
4798 | EC_POINT_hex2point() | |
4799 | providing useful interfaces to EC_POINT_point2oct() and | |
4800 | EC_POINT_oct2point(). | |
4801 | [Nils Larsch <nla@trustcenter.de>] | |
4802 | ||
b6db386f BM |
4803 | *) Change internals of the EC library so that the functions |
4804 | EC_GROUP_set_generator() | |
4805 | EC_GROUP_get_generator() | |
4806 | EC_GROUP_get_order() | |
4807 | EC_GROUP_get_cofactor() | |
4808 | are implemented directly in crypto/ec/ec_lib.c and not dispatched | |
4809 | to methods, which would lead to unnecessary code duplication when | |
4810 | adding different types of curves. | |
6cbe6382 | 4811 | [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller] |
b6db386f | 4812 | |
47234cd3 BM |
4813 | *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM |
4814 | arithmetic, and such that modified wNAFs are generated | |
4815 | (which avoid length expansion in many cases). | |
4816 | [Bodo Moeller] | |
4817 | ||
82652aaf BM |
4818 | *) Add a function EC_GROUP_check_discriminant() (defined via |
4819 | EC_METHOD) that verifies that the curve discriminant is non-zero. | |
4820 | ||
4821 | Add a function EC_GROUP_check() that makes some sanity tests | |
4822 | on a EC_GROUP, its generator and order. This includes | |
4823 | EC_GROUP_check_discriminant(). | |
4824 | [Nils Larsch <nla@trustcenter.de>] | |
4825 | ||
4d94ae00 BM |
4826 | *) Add ECDSA in new directory crypto/ecdsa/. |
4827 | ||
5dbd3efc BM |
4828 | Add applications 'openssl ecparam' and 'openssl ecdsa' |
4829 | (these are based on 'openssl dsaparam' and 'openssl dsa'). | |
4d94ae00 BM |
4830 | |
4831 | ECDSA support is also included in various other files across the | |
4832 | library. Most notably, | |
4833 | - 'openssl req' now has a '-newkey ecdsa:file' option; | |
4834 | - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA; | |
4835 | - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and | |
4836 | d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make | |
4837 | them suitable for ECDSA where domain parameters must be | |
e172d60d BM |
4838 | extracted before the specific public key; |
4839 | - ECDSA engine support has been added. | |
f8e21776 | 4840 | [Nils Larsch <nla@trustcenter.de>] |
4d94ae00 | 4841 | |
af28dd6c | 4842 | *) Include some named elliptic curves, and add OIDs from X9.62, |
ed5e37c3 | 4843 | SECG, and WAP/WTLS. Each curve can be obtained from the new |
7eb18f12 | 4844 | function |
8b15c740 | 4845 | EC_GROUP_new_by_curve_name(), |
ed5e37c3 BM |
4846 | and the list of available named curves can be obtained with |
4847 | EC_get_builtin_curves(). | |
254ef80d BM |
4848 | Also add a 'curve_name' member to EC_GROUP objects, which can be |
4849 | accessed via | |
0f449936 BM |
4850 | EC_GROUP_set_curve_name() |
4851 | EC_GROUP_get_curve_name() | |
8b15c740 | 4852 | [Nils Larsch <larsch@trustcenter.de, Bodo Moeller] |
4d94ae00 | 4853 | |
c1862f91 BM |
4854 | *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there |
4855 | was actually never needed) and in BN_mul(). The removal in BN_mul() | |
4856 | required a small change in bn_mul_part_recursive() and the addition | |
4857 | of the functions bn_cmp_part_words(), bn_sub_part_words() and | |
4858 | bn_add_part_words(), which do the same thing as bn_cmp_words(), | |
4859 | bn_sub_words() and bn_add_words() except they take arrays with | |
4860 | differing sizes. | |
4861 | [Richard Levitte] | |
4862 | ||
dd2b6750 | 4863 | Changes between 0.9.7l and 0.9.7m [23 Feb 2007] |
1e24b3a0 | 4864 | |
a2e623c0 DSH |
4865 | *) Cleanse PEM buffers before freeing them since they may contain |
4866 | sensitive data. | |
4867 | [Benjamin Bennett <ben@psc.edu>] | |
4868 | ||
0a05123a BM |
4869 | *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that |
4870 | a ciphersuite string such as "DEFAULT:RSA" cannot enable | |
4871 | authentication-only ciphersuites. | |
4872 | [Bodo Moeller] | |
4873 | ||
52b8dad8 BM |
4874 | *) Since AES128 and AES256 share a single mask bit in the logic of |
4875 | ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a | |
4876 | kludge to work properly if AES128 is available and AES256 isn't. | |
4877 | [Victor Duchovni] | |
4878 | ||
dd2b6750 BM |
4879 | *) Expand security boundary to match 1.1.1 module. |
4880 | [Steve Henson] | |
4881 | ||
4882 | *) Remove redundant features: hash file source, editing of test vectors | |
4883 | modify fipsld to use external fips_premain.c signature. | |
4884 | [Steve Henson] | |
4885 | ||
4886 | *) New perl script mkfipsscr.pl to create shell scripts or batch files to | |
4887 | run algorithm test programs. | |
4888 | [Steve Henson] | |
4889 | ||
4890 | *) Make algorithm test programs more tolerant of whitespace. | |
4891 | [Steve Henson] | |
4892 | ||
1e24b3a0 BM |
4893 | *) Have SSL/TLS server implementation tolerate "mismatched" record |
4894 | protocol version while receiving ClientHello even if the | |
4895 | ClientHello is fragmented. (The server can't insist on the | |
4896 | particular protocol version it has chosen before the ServerHello | |
4897 | message has informed the client about his choice.) | |
4898 | [Bodo Moeller] | |
4899 | ||
4900 | *) Load error codes if they are not already present instead of using a | |
4901 | static variable. This allows them to be cleanly unloaded and reloaded. | |
4902 | [Steve Henson] | |
4903 | ||
61118caa BM |
4904 | Changes between 0.9.7k and 0.9.7l [28 Sep 2006] |
4905 | ||
4906 | *) Introduce limits to prevent malicious keys being able to | |
4907 | cause a denial of service. (CVE-2006-2940) | |
4908 | [Steve Henson, Bodo Moeller] | |
4909 | ||
4910 | *) Fix ASN.1 parsing of certain invalid structures that can result | |
4911 | in a denial of service. (CVE-2006-2937) [Steve Henson] | |
4912 | ||
4913 | *) Fix buffer overflow in SSL_get_shared_ciphers() function. | |
4914 | (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team] | |
4915 | ||
4916 | *) Fix SSL client code which could crash if connecting to a | |
4917 | malicious SSLv2 server. (CVE-2006-4343) | |
4918 | [Tavis Ormandy and Will Drewry, Google Security Team] | |
b6699c3f BM |
4919 | |
4920 | *) Change ciphersuite string processing so that an explicit | |
4921 | ciphersuite selects this one ciphersuite (so that "AES256-SHA" | |
4922 | will no longer include "AES128-SHA"), and any other similar | |
4923 | ciphersuite (same bitmap) from *other* protocol versions (so that | |
4924 | "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the | |
4925 | SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining | |
4926 | changes from 0.9.8b and 0.9.8d. | |
4927 | [Bodo Moeller] | |
4928 | ||
b79aa05e MC |
4929 | Changes between 0.9.7j and 0.9.7k [05 Sep 2006] |
4930 | ||
4931 | *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher | |
4932 | (CVE-2006-4339) [Ben Laurie and Google Security Team] | |
675f605d | 4933 | |
27a3d9f9 RL |
4934 | *) Change the Unix randomness entropy gathering to use poll() when |
4935 | possible instead of select(), since the latter has some | |
4936 | undesirable limitations. | |
e34aa5a3 | 4937 | [Darryl Miles via Richard Levitte and Bodo Moeller] |
27a3d9f9 | 4938 | |
5b57fe0a BM |
4939 | *) Disable rogue ciphersuites: |
4940 | ||
4941 | - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5") | |
4942 | - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5") | |
4943 | - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5") | |
4944 | ||
4945 | The latter two were purportedly from | |
4946 | draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really | |
4947 | appear there. | |
4948 | ||
4949 | Also deactive the remaining ciphersuites from | |
4950 | draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as | |
4951 | unofficial, and the ID has long expired. | |
4952 | [Bodo Moeller] | |
4953 | ||
675f605d BM |
4954 | *) Fix RSA blinding Heisenbug (problems sometimes occured on |
4955 | dual-core machines) and other potential thread-safety issues. | |
4956 | [Bodo Moeller] | |
4957 | ||
4958 | Changes between 0.9.7i and 0.9.7j [04 May 2006] | |
4959 | ||
4960 | *) Adapt fipsld and the build system to link against the validated FIPS | |
4961 | module in FIPS mode. | |
4962 | [Steve Henson] | |
4963 | ||
4964 | *) Fixes for VC++ 2005 build under Windows. | |
4965 | [Steve Henson] | |
4966 | ||
4967 | *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make | |
4968 | from a Windows bash shell such as MSYS. It is autodetected from the | |
4969 | "config" script when run from a VC++ environment. Modify standard VC++ | |
4970 | build to use fipscanister.o from the GNU make build. | |
4971 | [Steve Henson] | |
4972 | ||
89ec4332 RL |
4973 | Changes between 0.9.7h and 0.9.7i [14 Oct 2005] |
4974 | ||
4975 | *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS. | |
4976 | The value now differs depending on if you build for FIPS or not. | |
4977 | BEWARE! A program linked with a shared FIPSed libcrypto can't be | |
4978 | safely run with a non-FIPSed libcrypto, as it may crash because of | |
4979 | the difference induced by this change. | |
4980 | [Andy Polyakov] | |
4981 | ||
d357be38 MC |
4982 | Changes between 0.9.7g and 0.9.7h [11 Oct 2005] |
4983 | ||
4984 | *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING | |
4985 | (part of SSL_OP_ALL). This option used to disable the | |
4986 | countermeasure against man-in-the-middle protocol-version | |
4987 | rollback in the SSL 2.0 server implementation, which is a bad | |
04fac373 | 4988 | idea. (CVE-2005-2969) |
d357be38 MC |
4989 | |
4990 | [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center | |
4991 | for Information Security, National Institute of Advanced Industrial | |
4992 | Science and Technology [AIST], Japan)] | |
bc3cae7e | 4993 | |
b615ad90 | 4994 | *) Minimal support for X9.31 signatures and PSS padding modes. This is |
5d6c4985 | 4995 | mainly for FIPS compliance and not fully integrated at this stage. |
b615ad90 DSH |
4996 | [Steve Henson] |
4997 | ||
0ebfcc8f BM |
4998 | *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform |
4999 | the exponentiation using a fixed-length exponent. (Otherwise, | |
5000 | the information leaked through timing could expose the secret key | |
5001 | after many signatures; cf. Bleichenbacher's attack on DSA with | |
5002 | biased k.) | |
5003 | [Bodo Moeller] | |
5004 | ||
46a64376 | 5005 | *) Make a new fixed-window mod_exp implementation the default for |
91b17fba BM |
5006 | RSA, DSA, and DH private-key operations so that the sequence of |
5007 | squares and multiplies and the memory access pattern are | |
5008 | independent of the particular secret key. This will mitigate | |
5009 | cache-timing and potential related attacks. | |
46a64376 BM |
5010 | |
5011 | BN_mod_exp_mont_consttime() is the new exponentiation implementation, | |
5012 | and this is automatically used by BN_mod_exp_mont() if the new flag | |
db99c525 | 5013 | BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH |
46a64376 BM |
5014 | will use this BN flag for private exponents unless the flag |
5015 | RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or | |
5016 | DH_FLAG_NO_EXP_CONSTTIME, respectively, is set. | |
5017 | ||
5018 | [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller] | |
5019 | ||
c6c2e313 BM |
5020 | *) Change the client implementation for SSLv23_method() and |
5021 | SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0 | |
5022 | Client Hello message format if the SSL_OP_NO_SSLv2 option is set. | |
5023 | (Previously, the SSL 2.0 backwards compatible Client Hello | |
5024 | message format would be used even with SSL_OP_NO_SSLv2.) | |
5025 | [Bodo Moeller] | |
5026 | ||
05338b58 DSH |
5027 | *) Add support for smime-type MIME parameter in S/MIME messages which some |
5028 | clients need. | |
5029 | [Steve Henson] | |
5030 | ||
6ec8e63a DSH |
5031 | *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in |
5032 | a threadsafe manner. Modify rsa code to use new function and add calls | |
5033 | to dsa and dh code (which had race conditions before). | |
5034 | [Steve Henson] | |
5035 | ||
bc3cae7e DSH |
5036 | *) Include the fixed error library code in the C error file definitions |
5037 | instead of fixing them up at runtime. This keeps the error code | |
5038 | structures constant. | |
5039 | [Steve Henson] | |
5040 | ||
5041 | Changes between 0.9.7f and 0.9.7g [11 Apr 2005] | |
7a8c7288 | 5042 | |
a1006c37 BM |
5043 | [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after |
5044 | OpenSSL 0.9.8.] | |
5045 | ||
0858b71b DSH |
5046 | *) Fixes for newer kerberos headers. NB: the casts are needed because |
5047 | the 'length' field is signed on one version and unsigned on another | |
5048 | with no (?) obvious way to tell the difference, without these VC++ | |
5049 | complains. Also the "definition" of FAR (blank) is no longer included | |
5050 | nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up | |
5051 | some needed definitions. | |
5052 | [Steve Henson] | |
5053 | ||
7a8c7288 | 5054 | *) Undo Cygwin change. |
9f0b86c6 | 5055 | [Ulf Möller] |
7a8c7288 | 5056 | |
d9bfe4f9 RL |
5057 | *) Added support for proxy certificates according to RFC 3820. |
5058 | Because they may be a security thread to unaware applications, | |
5059 | they must be explicitely allowed in run-time. See | |
5060 | docs/HOWTO/proxy_certificates.txt for further information. | |
5061 | [Richard Levitte] | |
5062 | ||
b0ef321c | 5063 | Changes between 0.9.7e and 0.9.7f [22 Mar 2005] |
5022e4ec | 5064 | |
59b6836a DSH |
5065 | *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating |
5066 | server and client random values. Previously | |
5067 | (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in | |
5068 | less random data when sizeof(time_t) > 4 (some 64 bit platforms). | |
5069 | ||
5070 | This change has negligible security impact because: | |
5071 | ||
5072 | 1. Server and client random values still have 24 bytes of pseudo random | |
5073 | data. | |
5074 | ||
5075 | 2. Server and client random values are sent in the clear in the initial | |
5076 | handshake. | |
5077 | ||
5078 | 3. The master secret is derived using the premaster secret (48 bytes in | |
5079 | size for static RSA ciphersuites) as well as client server and random | |
5080 | values. | |
5081 | ||
5082 | The OpenSSL team would like to thank the UK NISCC for bringing this issue | |
5083 | to our attention. | |
5084 | ||
5085 | [Stephen Henson, reported by UK NISCC] | |
5086 | ||
130db968 | 5087 | *) Use Windows randomness collection on Cygwin. |
9f0b86c6 | 5088 | [Ulf Möller] |
130db968 | 5089 | |
f69a8aeb LJ |
5090 | *) Fix hang in EGD/PRNGD query when communication socket is closed |
5091 | prematurely by EGD/PRNGD. | |
9f0b86c6 | 5092 | [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014] |
f69a8aeb | 5093 | |
e90fadda DSH |
5094 | *) Prompt for pass phrases when appropriate for PKCS12 input format. |
5095 | [Steve Henson] | |
5096 | ||
b0ef321c BM |
5097 | *) Back-port of selected performance improvements from development |
5098 | branch, as well as improved support for PowerPC platforms. | |
5099 | [Andy Polyakov] | |
5100 | ||
a0e7c8ee DSH |
5101 | *) Add lots of checks for memory allocation failure, error codes to indicate |
5102 | failure and freeing up memory if a failure occurs. | |
5103 | [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson] | |
5104 | ||
5b40d7dd DSH |
5105 | *) Add new -passin argument to dgst. |
5106 | [Steve Henson] | |
5107 | ||
1862dae8 DSH |
5108 | *) Perform some character comparisons of different types in X509_NAME_cmp: |
5109 | this is needed for some certificates that reencode DNs into UTF8Strings | |
5110 | (in violation of RFC3280) and can't or wont issue name rollover | |
5111 | certificates. | |
5112 | [Steve Henson] | |
5113 | ||
5022e4ec RL |
5114 | *) Make an explicit check during certificate validation to see that |
5115 | the CA setting in each certificate on the chain is correct. As a | |
5116 | side effect always do the following basic checks on extensions, | |
5117 | not just when there's an associated purpose to the check: | |
5118 | ||
5119 | - if there is an unhandled critical extension (unless the user | |
5120 | has chosen to ignore this fault) | |
5121 | - if the path length has been exceeded (if one is set at all) | |
5122 | - that certain extensions fit the associated purpose (if one has | |
5123 | been given) | |
5124 | [Richard Levitte] | |
5125 | ||
5126 | Changes between 0.9.7d and 0.9.7e [25 Oct 2004] | |
d5f686d8 | 5127 | |
2f605e8d DSH |
5128 | *) Avoid a race condition when CRLs are checked in a multi threaded |
5129 | environment. This would happen due to the reordering of the revoked | |
5130 | entries during signature checking and serial number lookup. Now the | |
5131 | encoding is cached and the serial number sort performed under a lock. | |
5132 | Add new STACK function sk_is_sorted(). | |
5133 | [Steve Henson] | |
5134 | ||
637ff35e DSH |
5135 | *) Add Delta CRL to the extension code. |
5136 | [Steve Henson] | |
5137 | ||
4843acc8 DSH |
5138 | *) Various fixes to s3_pkt.c so alerts are sent properly. |
5139 | [David Holmes <d.holmes@f5.com>] | |
5140 | ||
d5f686d8 BM |
5141 | *) Reduce the chances of duplicate issuer name and serial numbers (in |
5142 | violation of RFC3280) using the OpenSSL certificate creation utilities. | |
5143 | This is done by creating a random 64 bit value for the initial serial | |
5144 | number when a serial number file is created or when a self signed | |
5145 | certificate is created using 'openssl req -x509'. The initial serial | |
5146 | number file is created using 'openssl x509 -next_serial' in CA.pl | |
5147 | rather than being initialized to 1. | |
5148 | [Steve Henson] | |
5149 | ||
5150 | Changes between 0.9.7c and 0.9.7d [17 Mar 2004] | |
5151 | ||
5152 | *) Fix null-pointer assignment in do_change_cipher_spec() revealed | |
04fac373 | 5153 | by using the Codenomicon TLS Test Tool (CVE-2004-0079) |
d5f686d8 BM |
5154 | [Joe Orton, Steve Henson] |
5155 | ||
5156 | *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites | |
04fac373 | 5157 | (CVE-2004-0112) |
d5f686d8 BM |
5158 | [Joe Orton, Steve Henson] |
5159 | ||
5160 | *) Make it possible to have multiple active certificates with the same | |
5161 | subject in the CA index file. This is done only if the keyword | |
5162 | 'unique_subject' is set to 'no' in the main CA section (default | |
5163 | if 'CA_default') of the configuration file. The value is saved | |
5164 | with the database itself in a separate index attribute file, | |
5165 | named like the index file with '.attr' appended to the name. | |
5166 | [Richard Levitte] | |
cd2e8a6f | 5167 | |
bc501570 DSH |
5168 | *) X509 verify fixes. Disable broken certificate workarounds when |
5169 | X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if | |
5170 | keyUsage extension present. Don't accept CRLs with unhandled critical | |
5171 | extensions: since verify currently doesn't process CRL extensions this | |
5172 | rejects a CRL with *any* critical extensions. Add new verify error codes | |
5173 | for these cases. | |
5174 | [Steve Henson] | |
5175 | ||
dc90f64d DSH |
5176 | *) When creating an OCSP nonce use an OCTET STRING inside the extnValue. |
5177 | A clarification of RFC2560 will require the use of OCTET STRINGs and | |
5178 | some implementations cannot handle the current raw format. Since OpenSSL | |
5179 | copies and compares OCSP nonces as opaque blobs without any attempt at | |
5180 | parsing them this should not create any compatibility issues. | |
5181 | [Steve Henson] | |
5182 | ||
d4575825 DSH |
5183 | *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when |
5184 | calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without | |
5185 | this HMAC (and other) operations are several times slower than OpenSSL | |
5186 | < 0.9.7. | |
5187 | [Steve Henson] | |
5188 | ||
cd2e8a6f DSH |
5189 | *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex(). |
5190 | [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] | |
5191 | ||
caf044cb DSH |
5192 | *) Use the correct content when signing type "other". |
5193 | [Steve Henson] | |
5194 | ||
29902449 DSH |
5195 | Changes between 0.9.7b and 0.9.7c [30 Sep 2003] |
5196 | ||
5197 | *) Fix various bugs revealed by running the NISCC test suite: | |
5198 | ||
5199 | Stop out of bounds reads in the ASN1 code when presented with | |
04fac373 | 5200 | invalid tags (CVE-2003-0543 and CVE-2003-0544). |
29902449 | 5201 | |
04fac373 | 5202 | Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545). |
29902449 DSH |
5203 | |
5204 | If verify callback ignores invalid public key errors don't try to check | |
5205 | certificate signature with the NULL public key. | |
5206 | ||
5207 | [Steve Henson] | |
beab098d | 5208 | |
560dfd2a DSH |
5209 | *) New -ignore_err option in ocsp application to stop the server |
5210 | exiting on the first error in a request. | |
5211 | [Steve Henson] | |
5212 | ||
a9077513 BM |
5213 | *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate |
5214 | if the server requested one: as stated in TLS 1.0 and SSL 3.0 | |
5215 | specifications. | |
5216 | [Steve Henson] | |
5217 | ||
ddc38679 BM |
5218 | *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional |
5219 | extra data after the compression methods not only for TLS 1.0 | |
5220 | but also for SSL 3.0 (as required by the specification). | |
5221 | [Bodo Moeller; problem pointed out by Matthias Loepfe] | |
5222 | ||
5223 | *) Change X509_certificate_type() to mark the key as exported/exportable | |
5224 | when it's 512 *bits* long, not 512 bytes. | |
5225 | [Richard Levitte] | |
5226 | ||
a0694600 RL |
5227 | *) Change AES_cbc_encrypt() so it outputs exact multiple of |
5228 | blocks during encryption. | |
5229 | [Richard Levitte] | |
5230 | ||
63b81558 DSH |
5231 | *) Various fixes to base64 BIO and non blocking I/O. On write |
5232 | flushes were not handled properly if the BIO retried. On read | |
5233 | data was not being buffered properly and had various logic bugs. | |
5234 | This also affects blocking I/O when the data being decoded is a | |
5235 | certain size. | |
5236 | [Steve Henson] | |
5237 | ||
beab098d DSH |
5238 | *) Various S/MIME bugfixes and compatibility changes: |
5239 | output correct application/pkcs7 MIME type if | |
5240 | PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures. | |
5241 | Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening | |
5242 | of files as .eml work). Correctly handle very long lines in MIME | |
5243 | parser. | |
5244 | [Steve Henson] | |
5245 | ||
5246 | Changes between 0.9.7a and 0.9.7b [10 Apr 2003] | |
132eaa59 | 5247 | |
02da5bcd BM |
5248 | *) Countermeasure against the Klima-Pokorny-Rosa extension of |
5249 | Bleichbacher's attack on PKCS #1 v1.5 padding: treat | |
5250 | a protocol version number mismatch like a decryption error | |
5251 | in ssl3_get_client_key_exchange (ssl/s3_srvr.c). | |
5252 | [Bodo Moeller] | |
5253 | ||
c554155b BM |
5254 | *) Turn on RSA blinding by default in the default implementation |
5255 | to avoid a timing attack. Applications that don't want it can call | |
5256 | RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. | |
5257 | They would be ill-advised to do so in most cases. | |
5679bcce | 5258 | [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] |
c554155b BM |
5259 | |
5260 | *) Change RSA blinding code so that it works when the PRNG is not | |
5261 | seeded (in this case, the secret RSA exponent is abused as | |
5262 | an unpredictable seed -- if it is not unpredictable, there | |
5679bcce BM |
5263 | is no point in blinding anyway). Make RSA blinding thread-safe |
5264 | by remembering the creator's thread ID in rsa->blinding and | |
5265 | having all other threads use local one-time blinding factors | |
5266 | (this requires more computation than sharing rsa->blinding, but | |
5267 | avoids excessive locking; and if an RSA object is not shared | |
5268 | between threads, blinding will still be very fast). | |
c554155b BM |
5269 | [Bodo Moeller] |
5270 | ||
d5f686d8 BM |
5271 | *) Fixed a typo bug that would cause ENGINE_set_default() to set an |
5272 | ENGINE as defaults for all supported algorithms irrespective of | |
5273 | the 'flags' parameter. 'flags' is now honoured, so applications | |
5274 | should make sure they are passing it correctly. | |
5275 | [Geoff Thorpe] | |
5276 | ||
63ff3e83 UM |
5277 | *) Target "mingw" now allows native Windows code to be generated in |
5278 | the Cygwin environment as well as with the MinGW compiler. | |
5279 | [Ulf Moeller] | |
132eaa59 | 5280 | |
5b0b0e98 RL |
5281 | Changes between 0.9.7 and 0.9.7a [19 Feb 2003] |
5282 | ||
5283 | *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked | |
5284 | via timing by performing a MAC computation even if incorrrect | |
5285 | block cipher padding has been found. This is a countermeasure | |
5286 | against active attacks where the attacker has to distinguish | |
04fac373 | 5287 | between bad padding and a MAC verification error. (CVE-2003-0078) |
5b0b0e98 RL |
5288 | |
5289 | [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), | |
5290 | Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and | |
5291 | Martin Vuagnoux (EPFL, Ilion)] | |
948dcdb8 | 5292 | |
758f942b RL |
5293 | *) Make the no-err option work as intended. The intention with no-err |
5294 | is not to have the whole error stack handling routines removed from | |
5295 | libcrypto, it's only intended to remove all the function name and | |
5296 | reason texts, thereby removing some of the footprint that may not | |
5297 | be interesting if those errors aren't displayed anyway. | |
5298 | ||
5299 | NOTE: it's still possible for any application or module to have it's | |
5300 | own set of error texts inserted. The routines are there, just not | |
5301 | used by default when no-err is given. | |
5302 | [Richard Levitte] | |
5303 | ||
b7bbac72 RL |
5304 | *) Add support for FreeBSD on IA64. |
5305 | [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454] | |
5306 | ||
9ec1d35f RL |
5307 | *) Adjust DES_cbc_cksum() so it returns the same value as the MIT |
5308 | Kerberos function mit_des_cbc_cksum(). Before this change, | |
5309 | the value returned by DES_cbc_cksum() was like the one from | |
5310 | mit_des_cbc_cksum(), except the bytes were swapped. | |
5311 | [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte] | |
5312 | ||
cf56663f DSH |
5313 | *) Allow an application to disable the automatic SSL chain building. |
5314 | Before this a rather primitive chain build was always performed in | |
5315 | ssl3_output_cert_chain(): an application had no way to send the | |
5316 | correct chain if the automatic operation produced an incorrect result. | |
5317 | ||
5318 | Now the chain builder is disabled if either: | |
5319 | ||
5320 | 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert(). | |
5321 | ||
5322 | 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set. | |
5323 | ||
5324 | The reasoning behind this is that an application would not want the | |
5325 | auto chain building to take place if extra chain certificates are | |
5326 | present and it might also want a means of sending no additional | |
5327 | certificates (for example the chain has two certificates and the | |
5328 | root is omitted). | |
5329 | [Steve Henson] | |
5330 | ||
0b13e9f0 RL |
5331 | *) Add the possibility to build without the ENGINE framework. |
5332 | [Steven Reddie <smr@essemer.com.au> via Richard Levitte] | |
5333 | ||
d3b5cb53 DSH |
5334 | *) Under Win32 gmtime() can return NULL: check return value in |
5335 | OPENSSL_gmtime(). Add error code for case where gmtime() fails. | |
5336 | [Steve Henson] | |
5337 | ||
a74333f9 LJ |
5338 | *) DSA routines: under certain error conditions uninitialized BN objects |
5339 | could be freed. Solution: make sure initialization is performed early | |
5340 | enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>, | |
5341 | Nils Larsch <nla@trustcenter.de> via PR#459) | |
5342 | [Lutz Jaenicke] | |
5343 | ||
8ec16ce7 LJ |
5344 | *) Another fix for SSLv2 session ID handling: the session ID was incorrectly |
5345 | checked on reconnect on the client side, therefore session resumption | |
5346 | could still fail with a "ssl session id is different" error. This | |
5347 | behaviour is masked when SSL_OP_ALL is used due to | |
5348 | SSL_OP_MICROSOFT_SESS_ID_BUG being set. | |
5349 | Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as | |
5350 | followup to PR #377. | |
5351 | [Lutz Jaenicke] | |
5352 | ||
04aff67d RL |
5353 | *) IA-32 assembler support enhancements: unified ELF targets, support |
5354 | for SCO/Caldera platforms, fix for Cygwin shared build. | |
5355 | [Andy Polyakov] | |
5356 | ||
afd41c9f RL |
5357 | *) Add support for FreeBSD on sparc64. As a consequence, support for |
5358 | FreeBSD on non-x86 processors is separate from x86 processors on | |
5359 | the config script, much like the NetBSD support. | |
5360 | [Richard Levitte & Kris Kennaway <kris@obsecurity.org>] | |
948dcdb8 | 5361 | |
02e05594 | 5362 | Changes between 0.9.6h and 0.9.7 [31 Dec 2002] |
3e06fb75 | 5363 | |
ddc38679 BM |
5364 | [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after |
5365 | OpenSSL 0.9.7.] | |
5366 | ||
21cde7a4 LJ |
5367 | *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED |
5368 | code (06) was taken as the first octet of the session ID and the last | |
5369 | octet was ignored consequently. As a result SSLv2 client side session | |
5370 | caching could not have worked due to the session ID mismatch between | |
5371 | client and server. | |
5372 | Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as | |
5373 | PR #377. | |
5374 | [Lutz Jaenicke] | |
5375 | ||
9cd16b1d RL |
5376 | *) Change the declaration of needed Kerberos libraries to use EX_LIBS |
5377 | instead of the special (and badly supported) LIBKRB5. LIBKRB5 is | |
5378 | removed entirely. | |
5379 | [Richard Levitte] | |
5380 | ||
14676ffc | 5381 | *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it |
a1457874 RL |
5382 | seems that in spite of existing for more than a year, many application |
5383 | author have done nothing to provide the necessary callbacks, which | |
14676ffc RL |
5384 | means that this particular engine will not work properly anywhere. |
5385 | This is a very unfortunate situation which forces us, in the name | |
5386 | of usability, to give the hw_ncipher.c a static lock, which is part | |
5387 | of libcrypto. | |
5388 | NOTE: This is for the 0.9.7 series ONLY. This hack will never | |
5389 | appear in 0.9.8 or later. We EXPECT application authors to have | |
5390 | dealt properly with this when 0.9.8 is released (unless we actually | |
5391 | make such changes in the libcrypto locking code that changes will | |
5392 | have to be made anyway). | |
5393 | [Richard Levitte] | |
5394 | ||
2053c43d DSH |
5395 | *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content |
5396 | octets have been read, EOF or an error occurs. Without this change | |
5397 | some truncated ASN1 structures will not produce an error. | |
5398 | [Steve Henson] | |
5399 | ||
17582ccf RL |
5400 | *) Disable Heimdal support, since it hasn't been fully implemented. |
5401 | Still give the possibility to force the use of Heimdal, but with | |
5402 | warnings and a request that patches get sent to openssl-dev. | |
5403 | [Richard Levitte] | |
5404 | ||
0bf23d9b RL |
5405 | *) Add the VC-CE target, introduce the WINCE sysname, and add |
5406 | INSTALL.WCE and appropriate conditionals to make it build. | |
5407 | [Steven Reddie <smr@essemer.com.au> via Richard Levitte] | |
5408 | ||
6f17f16f RL |
5409 | *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and |
5410 | cygssl-x.y.z.dll, where x, y and z are the major, minor and | |
5411 | edit numbers of the version. | |
5412 | [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] | |
5413 | ||
54a656ef BL |
5414 | *) Introduce safe string copy and catenation functions |
5415 | (BUF_strlcpy() and BUF_strlcat()). | |
5416 | [Ben Laurie (CHATS) and Richard Levitte] | |
5417 | ||
5418 | *) Avoid using fixed-size buffers for one-line DNs. | |
5419 | [Ben Laurie (CHATS)] | |
5420 | ||
5421 | *) Add BUF_MEM_grow_clean() to avoid information leakage when | |
5422 | resizing buffers containing secrets, and use where appropriate. | |
5423 | [Ben Laurie (CHATS)] | |
5424 | ||
5425 | *) Avoid using fixed size buffers for configuration file location. | |
5426 | [Ben Laurie (CHATS)] | |
5427 | ||
5428 | *) Avoid filename truncation for various CA files. | |
5429 | [Ben Laurie (CHATS)] | |
5430 | ||
5431 | *) Use sizeof in preference to magic numbers. | |
5432 | [Ben Laurie (CHATS)] | |
5433 | ||
5434 | *) Avoid filename truncation in cert requests. | |
5435 | [Ben Laurie (CHATS)] | |
5436 | ||
54a656ef BL |
5437 | *) Add assertions to check for (supposedly impossible) buffer |
5438 | overflows. | |
5439 | [Ben Laurie (CHATS)] | |
5440 | ||
5441 | *) Don't cache truncated DNS entries in the local cache (this could | |
5442 | potentially lead to a spoofing attack). | |
5443 | [Ben Laurie (CHATS)] | |
5444 | ||
5445 | *) Fix various buffers to be large enough for hex/decimal | |
5446 | representations in a platform independent manner. | |
5447 | [Ben Laurie (CHATS)] | |
5448 | ||
5449 | *) Add CRYPTO_realloc_clean() to avoid information leakage when | |
5450 | resizing buffers containing secrets, and use where appropriate. | |
5451 | [Ben Laurie (CHATS)] | |
5452 | ||
5453 | *) Add BIO_indent() to avoid much slightly worrying code to do | |
5454 | indents. | |
5455 | [Ben Laurie (CHATS)] | |
5456 | ||
5457 | *) Convert sprintf()/BIO_puts() to BIO_printf(). | |
5458 | [Ben Laurie (CHATS)] | |
5459 | ||
5460 | *) buffer_gets() could terminate with the buffer only half | |
5461 | full. Fixed. | |
5462 | [Ben Laurie (CHATS)] | |
5463 | ||
5464 | *) Add assertions to prevent user-supplied crypto functions from | |
5465 | overflowing internal buffers by having large block sizes, etc. | |
5466 | [Ben Laurie (CHATS)] | |
5467 | ||
2b2ab523 BM |
5468 | *) New OPENSSL_assert() macro (similar to assert(), but enabled |
5469 | unconditionally). | |
5470 | [Ben Laurie (CHATS)] | |
5471 | ||
54a656ef BL |
5472 | *) Eliminate unused copy of key in RC4. |
5473 | [Ben Laurie (CHATS)] | |
5474 | ||
5475 | *) Eliminate unused and incorrectly sized buffers for IV in pem.h. | |
5476 | [Ben Laurie (CHATS)] | |
5477 | ||
5478 | *) Fix off-by-one error in EGD path. | |
5479 | [Ben Laurie (CHATS)] | |
5480 | ||
5481 | *) If RANDFILE path is too long, ignore instead of truncating. | |
5482 | [Ben Laurie (CHATS)] | |
5483 | ||
5484 | *) Eliminate unused and incorrectly sized X.509 structure | |
5485 | CBCParameter. | |
5486 | [Ben Laurie (CHATS)] | |
5487 | ||
5488 | *) Eliminate unused and dangerous function knumber(). | |
5489 | [Ben Laurie (CHATS)] | |
5490 | ||
5491 | *) Eliminate unused and dangerous structure, KSSL_ERR. | |
5492 | [Ben Laurie (CHATS)] | |
5493 | ||
5494 | *) Protect against overlong session ID context length in an encoded | |
5495 | session object. Since these are local, this does not appear to be | |
5496 | exploitable. | |
5497 | [Ben Laurie (CHATS)] | |
5498 | ||
3e06fb75 BM |
5499 | *) Change from security patch (see 0.9.6e below) that did not affect |
5500 | the 0.9.6 release series: | |
5501 | ||
5502 | Remote buffer overflow in SSL3 protocol - an attacker could | |
5503 | supply an oversized master key in Kerberos-enabled versions. | |
04fac373 | 5504 | (CVE-2002-0657) |
3e06fb75 | 5505 | [Ben Laurie (CHATS)] |
dc014d43 | 5506 | |
7ba3a4c3 RL |
5507 | *) Change the SSL kerb5 codes to match RFC 2712. |
5508 | [Richard Levitte] | |
5509 | ||
ba111217 BM |
5510 | *) Make -nameopt work fully for req and add -reqopt switch. |
5511 | [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson] | |
5512 | ||
3f6db7f5 DSH |
5513 | *) The "block size" for block ciphers in CFB and OFB mode should be 1. |
5514 | [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>] | |
5515 | ||
f013c7f2 RL |
5516 | *) Make sure tests can be performed even if the corresponding algorithms |
5517 | have been removed entirely. This was also the last step to make | |
5518 | OpenSSL compilable with DJGPP under all reasonable conditions. | |
5519 | [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>] | |
5520 | ||
648765ba | 5521 | *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT |
c6ccf055 LJ |
5522 | to allow version independent disabling of normally unselected ciphers, |
5523 | which may be activated as a side-effect of selecting a single cipher. | |
648765ba BM |
5524 | |
5525 | (E.g., cipher list string "RSA" enables ciphersuites that are left | |
5526 | out of "ALL" because they do not provide symmetric encryption. | |
5527 | "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.) | |
c6ccf055 LJ |
5528 | [Lutz Jaenicke, Bodo Moeller] |
5529 | ||
041843e4 RL |
5530 | *) Add appropriate support for separate platform-dependent build |
5531 | directories. The recommended way to make a platform-dependent | |
5532 | build directory is the following (tested on Linux), maybe with | |
5533 | some local tweaks: | |
5534 | ||
5535 | # Place yourself outside of the OpenSSL source tree. In | |
5536 | # this example, the environment variable OPENSSL_SOURCE | |
5537 | # is assumed to contain the absolute OpenSSL source directory. | |
3e06fb75 BM |
5538 | mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`" |
5539 | cd objtree/"`uname -s`-`uname -r`-`uname -m`" | |
4a9476dd | 5540 | (cd $OPENSSL_SOURCE; find . -type f) | while read F; do |
041843e4 RL |
5541 | mkdir -p `dirname $F` |
5542 | ln -s $OPENSSL_SOURCE/$F $F | |
5543 | done | |
5544 | ||
5545 | To be absolutely sure not to disturb the source tree, a "make clean" | |
5546 | is a good thing. If it isn't successfull, don't worry about it, | |
5547 | it probably means the source directory is very clean. | |
5548 | [Richard Levitte] | |
5549 | ||
a6c6874a GT |
5550 | *) Make sure any ENGINE control commands make local copies of string |
5551 | pointers passed to them whenever necessary. Otherwise it is possible | |
5552 | the caller may have overwritten (or deallocated) the original string | |
5553 | data when a later ENGINE operation tries to use the stored values. | |
9f0b86c6 | 5554 | [Götz Babin-Ebell <babinebell@trustcenter.de>] |
a6c6874a | 5555 | |
d15711ef BL |
5556 | *) Improve diagnostics in file reading and command-line digests. |
5557 | [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>] | |
5558 | ||
fbb56e5b RL |
5559 | *) Add AES modes CFB and OFB to the object database. Correct an |
5560 | error in AES-CFB decryption. | |
5561 | [Richard Levitte] | |
5562 | ||
544a2aea DSH |
5563 | *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this |
5564 | allows existing EVP_CIPHER_CTX structures to be reused after | |
5565 | calling EVP_*Final(). This behaviour is used by encryption | |
5566 | BIOs and some applications. This has the side effect that | |
5567 | applications must explicitly clean up cipher contexts with | |
5568 | EVP_CIPHER_CTX_cleanup() or they will leak memory. | |
5569 | [Steve Henson] | |
5570 | ||
dc014d43 DSH |
5571 | *) Check the values of dna and dnb in bn_mul_recursive before calling |
5572 | bn_mul_comba (a non zero value means the a or b arrays do not contain | |
5573 | n2 elements) and fallback to bn_mul_normal if either is not zero. | |
5574 | [Steve Henson] | |
4d94ae00 | 5575 | |
c0455cbb LJ |
5576 | *) Fix escaping of non-ASCII characters when using the -subj option |
5577 | of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>) | |
5578 | [Lutz Jaenicke] | |
5579 | ||
85fb12d5 | 5580 | *) Make object definitions compliant to LDAP (RFC2256): SN is the short |
e9cbcb1d LJ |
5581 | form for "surname", serialNumber has no short form. |
5582 | Use "mail" as the short name for "rfc822Mailbox" according to RFC2798; | |
5583 | therefore remove "mail" short name for "internet 7". | |
e1f7ea25 LJ |
5584 | The OID for unique identifiers in X509 certificates is |
5585 | x500UniqueIdentifier, not uniqueIdentifier. | |
e9cbcb1d | 5586 | Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>) |
08b977b5 | 5587 | [Lutz Jaenicke] |
ffbe98b7 | 5588 | |
85fb12d5 | 5589 | *) Add an "init" command to the ENGINE config module and auto initialize |
0dc09233 DSH |
5590 | ENGINEs. Without any "init" command the ENGINE will be initialized |
5591 | after all ctrl commands have been executed on it. If init=1 the | |
5592 | ENGINE is initailized at that point (ctrls before that point are run | |
5593 | on the uninitialized ENGINE and after on the initialized one). If | |
5594 | init=0 then the ENGINE will not be iniatialized at all. | |
5595 | [Steve Henson] | |
5596 | ||
85fb12d5 | 5597 | *) Fix the 'app_verify_callback' interface so that the user-defined |
023ec151 BM |
5598 | argument is actually passed to the callback: In the |
5599 | SSL_CTX_set_cert_verify_callback() prototype, the callback | |
5600 | declaration has been changed from | |
5601 | int (*cb)() | |
5602 | into | |
5603 | int (*cb)(X509_STORE_CTX *,void *); | |
5604 | in ssl_verify_cert_chain (ssl/ssl_cert.c), the call | |
5605 | i=s->ctx->app_verify_callback(&ctx) | |
5606 | has been changed into | |
5607 | i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg). | |
5608 | ||
5609 | To update applications using SSL_CTX_set_cert_verify_callback(), | |
5610 | a dummy argument can be added to their callback functions. | |
5611 | [D. K. Smetters <smetters@parc.xerox.com>] | |
5612 | ||
85fb12d5 | 5613 | *) Added the '4758cca' ENGINE to support IBM 4758 cards. |
92d1bc09 GT |
5614 | [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe] |
5615 | ||
85fb12d5 | 5616 | *) Add and OPENSSL_LOAD_CONF define which will cause |
e84be9b4 DSH |
5617 | OpenSSL_add_all_algorithms() to load the openssl.cnf config file. |
5618 | This allows older applications to transparently support certain | |
5619 | OpenSSL features: such as crypto acceleration and dynamic ENGINE loading. | |
5620 | Two new functions OPENSSL_add_all_algorithms_noconf() which will never | |
5621 | load the config file and OPENSSL_add_all_algorithms_conf() which will | |
5622 | always load it have also been added. | |
5623 | [Steve Henson] | |
5624 | ||
85fb12d5 | 5625 | *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES. |
0d22b5da RL |
5626 | Adjust NIDs and EVP layer. |
5627 | [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] | |
5628 | ||
85fb12d5 | 5629 | *) Config modules support in openssl utility. |
3647bee2 DSH |
5630 | |
5631 | Most commands now load modules from the config file, | |
5632 | though in a few (such as version) this isn't done | |
5633 | because it couldn't be used for anything. | |
5634 | ||
5635 | In the case of ca and req the config file used is | |
5636 | the same as the utility itself: that is the -config | |
5637 | command line option can be used to specify an | |
5638 | alternative file. | |
5639 | [Steve Henson] | |
5640 | ||
85fb12d5 | 5641 | *) Move default behaviour from OPENSSL_config(). If appname is NULL |
92f91ff4 DSH |
5642 | use "openssl_conf" if filename is NULL use default openssl config file. |
5643 | [Steve Henson] | |
5644 | ||
85fb12d5 | 5645 | *) Add an argument to OPENSSL_config() to allow the use of an alternative |
92f91ff4 DSH |
5646 | config section name. Add a new flag to tolerate a missing config file |
5647 | and move code to CONF_modules_load_file(). | |
5648 | [Steve Henson] | |
5649 | ||
85fb12d5 | 5650 | *) Support for crypto accelerator cards from Accelerated Encryption |
3cd039dd RL |
5651 | Processing, www.aep.ie. (Use engine 'aep') |
5652 | The support was copied from 0.9.6c [engine] and adapted/corrected | |
5653 | to work with the new engine framework. | |
5654 | [AEP Inc. and Richard Levitte] | |
5655 | ||
85fb12d5 | 5656 | *) Support for SureWare crypto accelerator cards from Baltimore |
3cd039dd RL |
5657 | Technologies. (Use engine 'sureware') |
5658 | The support was copied from 0.9.6c [engine] and adapted | |
5659 | to work with the new engine framework. | |
5660 | [Richard Levitte] | |
5661 | ||
85fb12d5 | 5662 | *) Have the CHIL engine fork-safe (as defined by nCipher) and actually |
1199e2d8 RL |
5663 | make the newer ENGINE framework commands for the CHIL engine work. |
5664 | [Toomas Kiisk <vix@cyber.ee> and Richard Levitte] | |
5665 | ||
85fb12d5 | 5666 | *) Make it possible to produce shared libraries on ReliantUNIX. |
a3fffd64 RL |
5667 | [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte] |
5668 | ||
85fb12d5 | 5669 | *) Add the configuration target debug-linux-ppro. |
80bb905d RL |
5670 | Make 'openssl rsa' use the general key loading routines |
5671 | implemented in apps.c, and make those routines able to | |
5672 | handle the key format FORMAT_NETSCAPE and the variant | |
5673 | FORMAT_IISSGC. | |
5674 | [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] | |
5675 | ||
381a146d | 5676 | *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). |
80bb905d RL |
5677 | [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] |
5678 | ||
85fb12d5 | 5679 | *) Add -keyform to rsautl, and document -engine. |
8242a6a9 RL |
5680 | [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>] |
5681 | ||
85fb12d5 | 5682 | *) Change BIO_new_file (crypto/bio/bss_file.c) to use new |
a14e2d9d BM |
5683 | BIO_R_NO_SUCH_FILE error code rather than the generic |
5684 | ERR_R_SYS_LIB error code if fopen() fails with ENOENT. | |
5685 | [Ben Laurie] | |
5686 | ||
85fb12d5 | 5687 | *) Add new functions |
a14e2d9d BM |
5688 | ERR_peek_last_error |
5689 | ERR_peek_last_error_line | |
5690 | ERR_peek_last_error_line_data. | |
5691 | These are similar to | |
5692 | ERR_peek_error | |
5693 | ERR_peek_error_line | |
5694 | ERR_peek_error_line_data, | |
5695 | but report on the latest error recorded rather than the first one | |
5696 | still in the error queue. | |
5697 | [Ben Laurie, Bodo Moeller] | |
5698 | ||
85fb12d5 | 5699 | *) default_algorithms option in ENGINE config module. This allows things |
df5eaa8a DSH |
5700 | like: |
5701 | default_algorithms = ALL | |
5702 | default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS | |
5703 | [Steve Henson] | |
5704 | ||
85fb12d5 | 5705 | *) Prelminary ENGINE config module. |
c9501c22 DSH |
5706 | [Steve Henson] |
5707 | ||
85fb12d5 | 5708 | *) New experimental application configuration code. |
bc37d996 DSH |
5709 | [Steve Henson] |
5710 | ||
85fb12d5 | 5711 | *) Change the AES code to follow the same name structure as all other |
6f9079fd RL |
5712 | symmetric ciphers, and behave the same way. Move everything to |
5713 | the directory crypto/aes, thereby obsoleting crypto/rijndael. | |
5714 | [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte] | |
5715 | ||
85fb12d5 | 5716 | *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c. |
7c517a04 BL |
5717 | [Ben Laurie and Theo de Raadt] |
5718 | ||
85fb12d5 | 5719 | *) Add option to output public keys in req command. |
21a85f19 DSH |
5720 | [Massimiliano Pala madwolf@openca.org] |
5721 | ||
85fb12d5 | 5722 | *) Use wNAFs in EC_POINTs_mul() for improved efficiency |
76c4336c | 5723 | (up to about 10% better than before for P-192 and P-224). |
3ba1f111 BM |
5724 | [Bodo Moeller] |
5725 | ||
85fb12d5 | 5726 | *) New functions/macros |
7aa983c6 BM |
5727 | |
5728 | SSL_CTX_set_msg_callback(ctx, cb) | |
5729 | SSL_CTX_set_msg_callback_arg(ctx, arg) | |
5730 | SSL_set_msg_callback(ssl, cb) | |
5731 | SSL_set_msg_callback_arg(ssl, arg) | |
5732 | ||
5733 | to request calling a callback function | |
5734 | ||
5735 | void cb(int write_p, int version, int content_type, | |
5736 | const void *buf, size_t len, SSL *ssl, void *arg) | |
5737 | ||
5738 | whenever a protocol message has been completely received | |
5739 | (write_p == 0) or sent (write_p == 1). Here 'version' is the | |
5740 | protocol version according to which the SSL library interprets | |
5741 | the current protocol message (SSL2_VERSION, SSL3_VERSION, or | |
5742 | TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or | |
5743 | the content type as defined in the SSL 3.0/TLS 1.0 protocol | |
5744 | specification (change_cipher_spec(20), alert(21), handshake(22)). | |
5745 | 'buf' and 'len' point to the actual message, 'ssl' to the | |
5746 | SSL object, and 'arg' is the application-defined value set by | |
5747 | SSL[_CTX]_set_msg_callback_arg(). | |
5748 | ||
5749 | 'openssl s_client' and 'openssl s_server' have new '-msg' options | |
5750 | to enable a callback that displays all protocol messages. | |
5751 | [Bodo Moeller] | |
5752 | ||
85fb12d5 | 5753 | *) Change the shared library support so shared libraries are built as |
a7b42009 RL |
5754 | soon as the corresponding static library is finished, and thereby get |
5755 | openssl and the test programs linked against the shared library. | |
5756 | This still only happens when the keyword "shard" has been given to | |
5757 | the configuration scripts. | |
5758 | ||
5759 | NOTE: shared library support is still an experimental thing, and | |
5760 | backward binary compatibility is still not guaranteed. | |
5761 | ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte] | |
5762 | ||
85fb12d5 | 5763 | *) Add support for Subject Information Access extension. |
7d5b04db DSH |
5764 | [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>] |
5765 | ||
85fb12d5 | 5766 | *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero |
48b0cf8b BM |
5767 | additional bytes when new memory had to be allocated, not just |
5768 | when reusing an existing buffer. | |
5769 | [Bodo Moeller] | |
5770 | ||
85fb12d5 | 5771 | *) New command line and configuration option 'utf8' for the req command. |
1fc6d41b DSH |
5772 | This allows field values to be specified as UTF8 strings. |
5773 | [Steve Henson] | |
5774 | ||
85fb12d5 | 5775 | *) Add -multi and -mr options to "openssl speed" - giving multiple parallel |
0e211563 BL |
5776 | runs for the former and machine-readable output for the latter. |
5777 | [Ben Laurie] | |
5778 | ||
85fb12d5 | 5779 | *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion |
89da653f BM |
5780 | of the e-mail address in the DN (i.e., it will go into a certificate |
5781 | extension only). The new configuration file option 'email_in_dn = no' | |
5782 | has the same effect. | |
5783 | [Massimiliano Pala madwolf@openca.org] | |
5784 | ||
85fb12d5 | 5785 | *) Change all functions with names starting with des_ to be starting |
12852213 | 5786 | with DES_ instead. Add wrappers that are compatible with libdes, |
2d57b73a | 5787 | but are named _ossl_old_des_*. Finally, add macros that map the |
0d81c69b RL |
5788 | des_* symbols to the corresponding _ossl_old_des_* if libdes |
5789 | compatibility is desired. If OpenSSL 0.9.6c compatibility is | |
5790 | desired, the des_* symbols will be mapped to DES_*, with one | |
5791 | exception. | |
12852213 | 5792 | |
0d81c69b RL |
5793 | Since we provide two compatibility mappings, the user needs to |
5794 | define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes | |
5795 | compatibility is desired. The default (i.e., when that macro | |
5796 | isn't defined) is OpenSSL 0.9.6c compatibility. | |
5797 | ||
5798 | There are also macros that enable and disable the support of old | |
5799 | des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT | |
5800 | and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those | |
5801 | are defined, the default will apply: to support the old des routines. | |
5802 | ||
5803 | In either case, one must include openssl/des.h to get the correct | |
5804 | definitions. Do not try to just include openssl/des_old.h, that | |
5805 | won't work. | |
c2e4f17c RL |
5806 | |
5807 | NOTE: This is a major break of an old API into a new one. Software | |
5808 | authors are encouraged to switch to the DES_ style functions. Some | |
5809 | time in the future, des_old.h and the libdes compatibility functions | |
0d81c69b RL |
5810 | will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the |
5811 | default), and then completely removed. | |
c2e4f17c RL |
5812 | [Richard Levitte] |
5813 | ||
85fb12d5 | 5814 | *) Test for certificates which contain unsupported critical extensions. |
f1558bb4 DSH |
5815 | If such a certificate is found during a verify operation it is |
5816 | rejected by default: this behaviour can be overridden by either | |
5817 | handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or | |
5818 | by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function | |
5819 | X509_supported_extension() has also been added which returns 1 if a | |
5820 | particular extension is supported. | |
5821 | [Steve Henson] | |
5822 | ||
85fb12d5 | 5823 | *) Modify the behaviour of EVP cipher functions in similar way to digests |
581f1c84 DSH |
5824 | to retain compatibility with existing code. |
5825 | [Steve Henson] | |
5826 | ||
85fb12d5 | 5827 | *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain |
50d194af DSH |
5828 | compatibility with existing code. In particular the 'ctx' parameter does |
5829 | not have to be to be initialized before the call to EVP_DigestInit() and | |
5830 | it is tidied up after a call to EVP_DigestFinal(). New function | |
5831 | EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function | |
5832 | EVP_MD_CTX_copy() changed to not require the destination to be | |
5833 | initialized valid and new function EVP_MD_CTX_copy_ex() added which | |
5834 | requires the destination to be valid. | |
5835 | ||
5836 | Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(), | |
5837 | EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex(). | |
20d2186c DSH |
5838 | [Steve Henson] |
5839 | ||
85fb12d5 | 5840 | *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it |
48948d53 BM |
5841 | so that complete 'Handshake' protocol structures are kept in memory |
5842 | instead of overwriting 'msg_type' and 'length' with 'body' data. | |
5843 | [Bodo Moeller] | |
5844 | ||
85fb12d5 | 5845 | *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32. |
285046ec RL |
5846 | [Massimo Santin via Richard Levitte] |
5847 | ||
85fb12d5 | 5848 | *) Major restructuring to the underlying ENGINE code. This includes |
07cee702 GT |
5849 | reduction of linker bloat, separation of pure "ENGINE" manipulation |
5850 | (initialisation, etc) from functionality dealing with implementations | |
5851 | of specific crypto iterfaces. This change also introduces integrated | |
5852 | support for symmetric ciphers and digest implementations - so ENGINEs | |
5853 | can now accelerate these by providing EVP_CIPHER and EVP_MD | |
5854 | implementations of their own. This is detailed in crypto/engine/README | |
5855 | as it couldn't be adequately described here. However, there are a few | |
5856 | API changes worth noting - some RSA, DSA, DH, and RAND functions that | |
5857 | were changed in the original introduction of ENGINE code have now | |
5858 | reverted back - the hooking from this code to ENGINE is now a good | |
5859 | deal more passive and at run-time, operations deal directly with | |
5860 | RSA_METHODs, DSA_METHODs (etc) as they did before, rather than | |
5861 | dereferencing through an ENGINE pointer any more. Also, the ENGINE | |
5862 | functions dealing with BN_MOD_EXP[_CRT] handlers have been removed - | |
5863 | they were not being used by the framework as there is no concept of a | |
5864 | BIGNUM_METHOD and they could not be generalised to the new | |
5865 | 'ENGINE_TABLE' mechanism that underlies the new code. Similarly, | |
5866 | ENGINE_cpy() has been removed as it cannot be consistently defined in | |
5867 | the new code. | |
5868 | [Geoff Thorpe] | |
5869 | ||
85fb12d5 | 5870 | *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds. |
d46c1a81 DSH |
5871 | [Steve Henson] |
5872 | ||
85fb12d5 | 5873 | *) Change mkdef.pl to sort symbols that get the same entry number, |
89eeccac RL |
5874 | and make sure the automatically generated functions ERR_load_* |
5875 | become part of libeay.num as well. | |
5876 | [Richard Levitte] | |
5877 | ||
85fb12d5 | 5878 | *) New function SSL_renegotiate_pending(). This returns true once |
6b0e9fac BM |
5879 | renegotiation has been requested (either SSL_renegotiate() call |
5880 | or HelloRequest/ClientHello receveived from the peer) and becomes | |
5881 | false once a handshake has been completed. | |
5882 | (For servers, SSL_renegotiate() followed by SSL_do_handshake() | |
5883 | sends a HelloRequest, but does not ensure that a handshake takes | |
5884 | place. SSL_renegotiate_pending() is useful for checking if the | |
5885 | client has followed the request.) | |
5886 | [Bodo Moeller] | |
5887 | ||
85fb12d5 | 5888 | *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION. |
6b0e9fac BM |
5889 | By default, clients may request session resumption even during |
5890 | renegotiation (if session ID contexts permit); with this option, | |
5891 | session resumption is possible only in the first handshake. | |
c21506ba BM |
5892 | |
5893 | SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes | |
5894 | more bits available for options that should not be part of | |
5895 | SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION). | |
6b0e9fac BM |
5896 | [Bodo Moeller] |
5897 | ||
85fb12d5 | 5898 | *) Add some demos for certificate and certificate request creation. |
96bd6f73 DSH |
5899 | [Steve Henson] |
5900 | ||
85fb12d5 | 5901 | *) Make maximum certificate chain size accepted from the peer application |
c0f5dd07 LJ |
5902 | settable (SSL*_get/set_max_cert_list()), as proposed by |
5903 | "Douglas E. Engert" <deengert@anl.gov>. | |
5904 | [Lutz Jaenicke] | |
5905 | ||
85fb12d5 | 5906 | *) Add support for shared libraries for Unixware-7 |
b26ca340 | 5907 | (Boyd Lynn Gerber <gerberb@zenez.com>). |
6c36f7a9 LJ |
5908 | [Lutz Jaenicke] |
5909 | ||
85fb12d5 | 5910 | *) Add a "destroy" handler to ENGINEs that allows structural cleanup to |
908efd3b GT |
5911 | be done prior to destruction. Use this to unload error strings from |
5912 | ENGINEs that load their own error strings. NB: This adds two new API | |
5913 | functions to "get" and "set" this destroy handler in an ENGINE. | |
a9ed4da8 | 5914 | [Geoff Thorpe] |
908efd3b | 5915 | |
85fb12d5 | 5916 | *) Alter all existing ENGINE implementations (except "openssl" and |
541814c4 GT |
5917 | "openbsd") to dynamically instantiate their own error strings. This |
5918 | makes them more flexible to be built both as statically-linked ENGINEs | |
5919 | and self-contained shared-libraries loadable via the "dynamic" ENGINE. | |
5920 | Also, add stub code to each that makes building them as self-contained | |
5921 | shared-libraries easier (see README.ENGINE). | |
5922 | [Geoff Thorpe] | |
5923 | ||
85fb12d5 | 5924 | *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE |
541814c4 GT |
5925 | implementations into applications that are completely implemented in |
5926 | self-contained shared-libraries. The "dynamic" ENGINE exposes control | |
5927 | commands that can be used to configure what shared-library to load and | |
5928 | to control aspects of the way it is handled. Also, made an update to | |
5929 | the README.ENGINE file that brings its information up-to-date and | |
5930 | provides some information and instructions on the "dynamic" ENGINE | |
5931 | (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc). | |
5932 | [Geoff Thorpe] | |
5933 | ||
85fb12d5 | 5934 | *) Make it possible to unload ranges of ERR strings with a new |
5b166395 GT |
5935 | "ERR_unload_strings" function. |
5936 | [Geoff Thorpe] | |
5937 | ||
85fb12d5 | 5938 | *) Add a copy() function to EVP_MD. |
26188931 BL |
5939 | [Ben Laurie] |
5940 | ||
85fb12d5 | 5941 | *) Make EVP_MD routines take a context pointer instead of just the |
e3fefbfd | 5942 | md_data void pointer. |
26188931 BL |
5943 | [Ben Laurie] |
5944 | ||
85fb12d5 | 5945 | *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates |
26188931 BL |
5946 | that the digest can only process a single chunk of data |
5947 | (typically because it is provided by a piece of | |
5948 | hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application | |
5949 | is only going to provide a single chunk of data, and hence the | |
5950 | framework needn't accumulate the data for oneshot drivers. | |
5951 | [Ben Laurie] | |
5952 | ||
85fb12d5 | 5953 | *) As with "ERR", make it possible to replace the underlying "ex_data" |
36026dfc GT |
5954 | functions. This change also alters the storage and management of global |
5955 | ex_data state - it's now all inside ex_data.c and all "class" code (eg. | |
5956 | RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class | |
5957 | index counters. The API functions that use this state have been changed | |
5958 | to take a "class_index" rather than pointers to the class's local STACK | |
5959 | and counter, and there is now an API function to dynamically create new | |
5960 | classes. This centralisation allows us to (a) plug a lot of the | |
5961 | thread-safety problems that existed, and (b) makes it possible to clean | |
5962 | up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b) | |
5963 | such data would previously have always leaked in application code and | |
5964 | workarounds were in place to make the memory debugging turn a blind eye | |
5965 | to it. Application code that doesn't use this new function will still | |
5966 | leak as before, but their memory debugging output will announce it now | |
5967 | rather than letting it slide. | |
6ee2a136 BM |
5968 | |
5969 | Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change | |
5970 | induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now | |
5971 | has a return value to indicate success or failure. | |
36026dfc GT |
5972 | [Geoff Thorpe] |
5973 | ||
85fb12d5 | 5974 | *) Make it possible to replace the underlying "ERR" functions such that the |
0783bf15 GT |
5975 | global state (2 LHASH tables and 2 locks) is only used by the "default" |
5976 | implementation. This change also adds two functions to "get" and "set" | |
5977 | the implementation prior to it being automatically set the first time | |
5978 | any other ERR function takes place. Ie. an application can call "get", | |
5979 | pass the return value to a module it has just loaded, and that module | |
5980 | can call its own "set" function using that value. This means the | |
5981 | module's "ERR" operations will use (and modify) the error state in the | |
5982 | application and not in its own statically linked copy of OpenSSL code. | |
5983 | [Geoff Thorpe] | |
5984 | ||
85fb12d5 | 5985 | *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment |
eb6dc02b GT |
5986 | reference counts. This performs normal REF_PRINT/REF_CHECK macros on |
5987 | the operation, and provides a more encapsulated way for external code | |
5988 | (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code | |
5989 | to use these functions rather than manually incrementing the counts. | |
3cad81f6 BM |
5990 | |
5991 | Also rename "DSO_up()" function to more descriptive "DSO_up_ref()". | |
eb6dc02b GT |
5992 | [Geoff Thorpe] |
5993 | ||
85fb12d5 | 5994 | *) Add EVP test program. |
0e360199 BL |
5995 | [Ben Laurie] |
5996 | ||
85fb12d5 | 5997 | *) Add symmetric cipher support to ENGINE. Expect the API to change! |
354c3ace BL |
5998 | [Ben Laurie] |
5999 | ||
85fb12d5 | 6000 | *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name() |
35bf3541 DSH |
6001 | X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(), |
6002 | X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate(). | |
6003 | These allow a CRL to be built without having to access X509_CRL fields | |
6004 | directly. Modify 'ca' application to use new functions. | |
6005 | [Steve Henson] | |
6006 | ||
85fb12d5 | 6007 | *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended |
06da6e49 | 6008 | bug workarounds. Rollback attack detection is a security feature. |
6383bbe5 | 6009 | The problem will only arise on OpenSSL servers when TLSv1 is not |
06da6e49 LJ |
6010 | available (sslv3_server_method() or SSL_OP_NO_TLSv1). |
6011 | Software authors not wanting to support TLSv1 will have special reasons | |
6012 | for their choice and can explicitly enable this option. | |
6013 | [Bodo Moeller, Lutz Jaenicke] | |
6014 | ||
85fb12d5 | 6015 | *) Rationalise EVP so it can be extended: don't include a union of |
76f8a1f5 BM |
6016 | cipher/digest structures, add init/cleanup functions for EVP_MD_CTX |
6017 | (similar to those existing for EVP_CIPHER_CTX). | |
4f4b1924 BM |
6018 | Usage example: |
6019 | ||
6020 | EVP_MD_CTX md; | |
6021 | ||
6022 | EVP_MD_CTX_init(&md); /* new function call */ | |
6023 | EVP_DigestInit(&md, EVP_sha1()); | |
6024 | EVP_DigestUpdate(&md, in, len); | |
6025 | EVP_DigestFinal(&md, out, NULL); | |
6026 | EVP_MD_CTX_cleanup(&md); /* new function call */ | |
6027 | ||
dbad1690 BL |
6028 | [Ben Laurie] |
6029 | ||
85fb12d5 | 6030 | *) Make DES key schedule conform to the usual scheme, as well as |
8408f4fb BL |
6031 | correcting its structure. This means that calls to DES functions |
6032 | now have to pass a pointer to a des_key_schedule instead of a | |
6033 | plain des_key_schedule (which was actually always a pointer | |
4f4b1924 BM |
6034 | anyway): E.g., |
6035 | ||
6036 | des_key_schedule ks; | |
6037 | ||
6038 | des_set_key_checked(..., &ks); | |
6039 | des_ncbc_encrypt(..., &ks, ...); | |
6040 | ||
6041 | (Note that a later change renames 'des_...' into 'DES_...'.) | |
dbad1690 BL |
6042 | [Ben Laurie] |
6043 | ||
85fb12d5 | 6044 | *) Initial reduction of linker bloat: the use of some functions, such as |
19da1300 DSH |
6045 | PEM causes large amounts of unused functions to be linked in due to |
6046 | poor organisation. For example pem_all.c contains every PEM function | |
6047 | which has a knock on effect of linking in large amounts of (unused) | |
6048 | ASN1 code. Grouping together similar functions and splitting unrelated | |
6049 | functions prevents this. | |
6050 | [Steve Henson] | |
6051 | ||
85fb12d5 | 6052 | *) Cleanup of EVP macros. |
381a146d | 6053 | [Ben Laurie] |
6aecef81 | 6054 | |
85fb12d5 | 6055 | *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the |
381a146d LJ |
6056 | correct _ecb suffix. |
6057 | [Ben Laurie] | |
c518ade1 | 6058 | |
85fb12d5 | 6059 | *) Add initial OCSP responder support to ocsp application. The |
ee306a13 DSH |
6060 | revocation information is handled using the text based index |
6061 | use by the ca application. The responder can either handle | |
6062 | requests generated internally, supplied in files (for example | |
6063 | via a CGI script) or using an internal minimal server. | |
6064 | [Steve Henson] | |
6065 | ||
85fb12d5 | 6066 | *) Add configuration choices to get zlib compression for TLS. |
e452de9d RL |
6067 | [Richard Levitte] |
6068 | ||
85fb12d5 | 6069 | *) Changes to Kerberos SSL for RFC 2712 compliance: |
0665dd68 RL |
6070 | 1. Implemented real KerberosWrapper, instead of just using |
6071 | KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>] | |
6072 | 2. Implemented optional authenticator field of KerberosWrapper. | |
6073 | ||
6074 | Added openssl-style ASN.1 macros for Kerberos ticket, ap_req, | |
6075 | and authenticator structs; see crypto/krb5/. | |
6076 | ||
6077 | Generalized Kerberos calls to support multiple Kerberos libraries. | |
6078 | [Vern Staats <staatsvr@asc.hpc.mil>, | |
6079 | Jeffrey Altman <jaltman@columbia.edu> | |
6080 | via Richard Levitte] | |
6081 | ||
85fb12d5 | 6082 | *) Cause 'openssl speed' to use fully hard-coded DSA keys as it |
af436bc1 GT |
6083 | already does with RSA. testdsa.h now has 'priv_key/pub_key' |
6084 | values for each of the key sizes rather than having just | |
6085 | parameters (and 'speed' generating keys each time). | |
6086 | [Geoff Thorpe] | |
6087 | ||
85fb12d5 | 6088 | *) Speed up EVP routines. |
f31b1250 BL |
6089 | Before: |
6090 | encrypt | |
6091 | type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes | |
6092 | des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k | |
6093 | des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k | |
6094 | des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k | |
6095 | decrypt | |
6096 | des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k | |
6097 | des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k | |
6098 | des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k | |
6099 | After: | |
6100 | encrypt | |
c148d709 | 6101 | des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k |
f31b1250 | 6102 | decrypt |
c148d709 | 6103 | des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k |
f31b1250 BL |
6104 | [Ben Laurie] |
6105 | ||
85fb12d5 | 6106 | *) Added the OS2-EMX target. |
c80410c5 RL |
6107 | ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte] |
6108 | ||
85fb12d5 | 6109 | *) Rewrite apps to use NCONF routines instead of the old CONF. New functions |
b7a26e6d DSH |
6110 | to support NCONF routines in extension code. New function CONF_set_nconf() |
6111 | to allow functions which take an NCONF to also handle the old LHASH | |
6112 | structure: this means that the old CONF compatible routines can be | |
6113 | retained (in particular wrt extensions) without having to duplicate the | |
6114 | code. New function X509V3_add_ext_nconf_sk to add extensions to a stack. | |
6115 | [Steve Henson] | |
6116 | ||
85fb12d5 | 6117 | *) Enhance the general user interface with mechanisms for inner control |
e3fefbfd | 6118 | and with possibilities to have yes/no kind of prompts. |
235dd0a2 RL |
6119 | [Richard Levitte] |
6120 | ||
85fb12d5 | 6121 | *) Change all calls to low level digest routines in the library and |
323f289c DSH |
6122 | applications to use EVP. Add missing calls to HMAC_cleanup() and |
6123 | don't assume HMAC_CTX can be copied using memcpy(). | |
6124 | [Verdon Walker <VWalker@novell.com>, Steve Henson] | |
6125 | ||
85fb12d5 | 6126 | *) Add the possibility to control engines through control names but with |
839590f5 RL |
6127 | arbitrary arguments instead of just a string. |
6128 | Change the key loaders to take a UI_METHOD instead of a callback | |
6129 | function pointer. NOTE: this breaks binary compatibility with earlier | |
6130 | versions of OpenSSL [engine]. | |
e3fefbfd | 6131 | Adapt the nCipher code for these new conditions and add a card insertion |
839590f5 RL |
6132 | callback. |
6133 | [Richard Levitte] | |
6134 | ||
85fb12d5 | 6135 | *) Enhance the general user interface with mechanisms to better support |
9ad0f681 RL |
6136 | dialog box interfaces, application-defined prompts, the possibility |
6137 | to use defaults (for example default passwords from somewhere else) | |
e3fefbfd | 6138 | and interrupts/cancellations. |
9ad0f681 RL |
6139 | [Richard Levitte] |
6140 | ||
85fb12d5 | 6141 | *) Tidy up PKCS#12 attribute handling. Add support for the CSP name |
f2a253e0 DSH |
6142 | attribute in PKCS#12 files, add new -CSP option to pkcs12 utility. |
6143 | [Steve Henson] | |
6144 | ||
85fb12d5 | 6145 | *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also |
e3fefbfd | 6146 | tidy up some unnecessarily weird code in 'sk_new()'). |
d918f851 GT |
6147 | [Geoff, reported by Diego Tartara <dtartara@novamens.com>] |
6148 | ||
85fb12d5 | 6149 | *) Change the key loading routines for ENGINEs to use the same kind |
79bb8d00 RL |
6150 | callback (pem_password_cb) as all other routines that need this |
6151 | kind of callback. | |
6152 | [Richard Levitte] | |
6153 | ||
85fb12d5 | 6154 | *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with |
e8734731 LJ |
6155 | 256 bit (=32 byte) keys. Of course seeding with more entropy bytes |
6156 | than this minimum value is recommended. | |
7e978372 | 6157 | [Lutz Jaenicke] |
4831e626 | 6158 | |
85fb12d5 | 6159 | *) New random seeder for OpenVMS, using the system process statistics |
496da8b9 RL |
6160 | that are easily reachable. |
6161 | [Richard Levitte] | |
6162 | ||
85fb12d5 | 6163 | *) Windows apparently can't transparently handle global |
4831e626 DSH |
6164 | variables defined in DLLs. Initialisations such as: |
6165 | ||
6166 | const ASN1_ITEM *it = &ASN1_INTEGER_it; | |
6167 | ||
6168 | wont compile. This is used by the any applications that need to | |
e3fefbfd | 6169 | declare their own ASN1 modules. This was fixed by adding the option |
4831e626 DSH |
6170 | EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly |
6171 | needed for static libraries under Win32. | |
6172 | [Steve Henson] | |
6173 | ||
85fb12d5 | 6174 | *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle |
926a56bf DSH |
6175 | setting of purpose and trust fields. New X509_STORE trust and |
6176 | purpose functions and tidy up setting in other SSL functions. | |
6177 | [Steve Henson] | |
6178 | ||
85fb12d5 | 6179 | *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE |
bdee69f7 DSH |
6180 | structure. These are inherited by X509_STORE_CTX when it is |
6181 | initialised. This allows various defaults to be set in the | |
6182 | X509_STORE structure (such as flags for CRL checking and custom | |
6183 | purpose or trust settings) for functions which only use X509_STORE_CTX | |
6184 | internally such as S/MIME. | |
6185 | ||
6186 | Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and | |
6187 | trust settings if they are not set in X509_STORE. This allows X509_STORE | |
6188 | purposes and trust (in S/MIME for example) to override any set by default. | |
6189 | ||
6190 | Add command line options for CRL checking to smime, s_client and s_server | |
6191 | applications. | |
6192 | [Steve Henson] | |
6193 | ||
85fb12d5 | 6194 | *) Initial CRL based revocation checking. If the CRL checking flag(s) |
b545dc67 DSH |
6195 | are set then the CRL is looked up in the X509_STORE structure and |
6196 | its validity and signature checked, then if the certificate is found | |
6197 | in the CRL the verify fails with a revoked error. | |
6198 | ||
6199 | Various new CRL related callbacks added to X509_STORE_CTX structure. | |
6200 | ||
6201 | Command line options added to 'verify' application to support this. | |
6202 | ||
6203 | This needs some additional work, such as being able to handle multiple | |
6204 | CRLs with different times, extension based lookup (rather than just | |
6205 | by subject name) and ultimately more complete V2 CRL extension | |
6206 | handling. | |
6207 | [Steve Henson] | |
6208 | ||
85fb12d5 | 6209 | *) Add a general user interface API (crypto/ui/). This is designed |
8a774dc9 BM |
6210 | to replace things like des_read_password and friends (backward |
6211 | compatibility functions using this new API are provided). | |
6212 | The purpose is to remove prompting functions from the DES code | |
6213 | section as well as provide for prompting through dialog boxes in | |
6214 | a window system and the like. | |
a63d5eaa RL |
6215 | [Richard Levitte] |
6216 | ||
85fb12d5 | 6217 | *) Add "ex_data" support to ENGINE so implementations can add state at a |
e5a77633 GT |
6218 | per-structure level rather than having to store it globally. |
6219 | [Geoff] | |
6220 | ||
85fb12d5 | 6221 | *) Make it possible for ENGINE structures to be copied when retrieved by |
e5a77633 GT |
6222 | ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY. |
6223 | This causes the "original" ENGINE structure to act like a template, | |
6224 | analogous to the RSA vs. RSA_METHOD type of separation. Because of this | |
6225 | operational state can be localised to each ENGINE structure, despite the | |
6226 | fact they all share the same "methods". New ENGINE structures returned in | |
6227 | this case have no functional references and the return value is the single | |
6228 | structural reference. This matches the single structural reference returned | |
6229 | by ENGINE_by_id() normally, when it is incremented on the pre-existing | |
6230 | ENGINE structure. | |
6231 | [Geoff] | |
6232 | ||
85fb12d5 | 6233 | *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this |
c962479b DSH |
6234 | needs to match any other type at all we need to manually clear the |
6235 | tag cache. | |
6236 | [Steve Henson] | |
6237 | ||
85fb12d5 | 6238 | *) Changes to the "openssl engine" utility to include; |
2a8a10ed GT |
6239 | - verbosity levels ('-v', '-vv', and '-vvv') that provide information |
6240 | about an ENGINE's available control commands. | |
6241 | - executing control commands from command line arguments using the | |
6242 | '-pre' and '-post' switches. '-post' is only used if '-t' is | |
6243 | specified and the ENGINE is successfully initialised. The syntax for | |
6244 | the individual commands are colon-separated, for example; | |
6245 | openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so | |
6246 | [Geoff] | |
6247 | ||
85fb12d5 | 6248 | *) New dynamic control command support for ENGINEs. ENGINEs can now |
2a8a10ed GT |
6249 | declare their own commands (numbers), names (strings), descriptions, |
6250 | and input types for run-time discovery by calling applications. A | |
6251 | subset of these commands are implicitly classed as "executable" | |
6252 | depending on their input type, and only these can be invoked through | |
6253 | the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this | |
6254 | can be based on user input, config files, etc). The distinction is | |
6255 | that "executable" commands cannot return anything other than a boolean | |
6256 | result and can only support numeric or string input, whereas some | |
6257 | discoverable commands may only be for direct use through | |
6258 | ENGINE_ctrl(), eg. supporting the exchange of binary data, function | |
6259 | pointers, or other custom uses. The "executable" commands are to | |
6260 | support parameterisations of ENGINE behaviour that can be | |
6261 | unambiguously defined by ENGINEs and used consistently across any | |
6262 | OpenSSL-based application. Commands have been added to all the | |
6263 | existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow | |
6264 | control over shared-library paths without source code alterations. | |
6265 | [Geoff] | |
6266 | ||
85fb12d5 | 6267 | *) Changed all ENGINE implementations to dynamically allocate their |
2a8a10ed GT |
6268 | ENGINEs rather than declaring them statically. Apart from this being |
6269 | necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction, | |
6270 | this also allows the implementations to compile without using the | |
6271 | internal engine_int.h header. | |
6272 | [Geoff] | |
6273 | ||
85fb12d5 | 6274 | *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a |
4d6115a5 GT |
6275 | 'const' value. Any code that should be able to modify a RAND_METHOD |
6276 | should already have non-const pointers to it (ie. they should only | |
6277 | modify their own ones). | |
6278 | [Geoff] | |
6279 | ||
85fb12d5 | 6280 | *) Made a variety of little tweaks to the ENGINE code. |
4d6115a5 GT |
6281 | - "atalla" and "ubsec" string definitions were moved from header files |
6282 | to C code. "nuron" string definitions were placed in variables | |
6283 | rather than hard-coded - allowing parameterisation of these values | |
6284 | later on via ctrl() commands. | |
6285 | - Removed unused "#if 0"'d code. | |
6286 | - Fixed engine list iteration code so it uses ENGINE_free() to release | |
6287 | structural references. | |
6288 | - Constified the RAND_METHOD element of ENGINE structures. | |
6289 | - Constified various get/set functions as appropriate and added | |
6290 | missing functions (including a catch-all ENGINE_cpy that duplicates | |
6291 | all ENGINE values onto a new ENGINE except reference counts/state). | |
6292 | - Removed NULL parameter checks in get/set functions. Setting a method | |
e13ae96d | 6293 | or function to NULL is a way of cancelling out a previously set |
4d6115a5 GT |
6294 | value. Passing a NULL ENGINE parameter is just plain stupid anyway |
6295 | and doesn't justify the extra error symbols and code. | |
6296 | - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for | |
6297 | flags from engine_int.h to engine.h. | |
6298 | - Changed prototypes for ENGINE handler functions (init(), finish(), | |
6299 | ctrl(), key-load functions, etc) to take an (ENGINE*) parameter. | |
6300 | [Geoff] | |
6301 | ||
85fb12d5 | 6302 | *) Implement binary inversion algorithm for BN_mod_inverse in addition |
e3fefbfd | 6303 | to the algorithm using long division. The binary algorithm can be |
1f224bf0 BM |
6304 | used only if the modulus is odd. On 32-bit systems, it is faster |
6305 | only for relatively small moduli (roughly 20-30% for 128-bit moduli, | |
6306 | roughly 5-15% for 256-bit moduli), so we use it only for moduli | |
6307 | up to 450 bits. In 64-bit environments, the binary algorithm | |
6308 | appears to be advantageous for much longer moduli; here we use it | |
6309 | for moduli up to 2048 bits. | |
7d0d0996 BM |
6310 | [Bodo Moeller] |
6311 | ||
85fb12d5 | 6312 | *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code |
722ca278 DSH |
6313 | could not support the combine flag in choice fields. |
6314 | [Steve Henson] | |
6315 | ||
85fb12d5 | 6316 | *) Add a 'copy_extensions' option to the 'ca' utility. This copies |
791bd0cd DSH |
6317 | extensions from a certificate request to the certificate. |
6318 | [Steve Henson] | |
6319 | ||
85fb12d5 | 6320 | *) Allow multiple 'certopt' and 'nameopt' options to be separated |
535d79da DSH |
6321 | by commas. Add 'namopt' and 'certopt' options to the 'ca' config |
6322 | file: this allows the display of the certificate about to be | |
6323 | signed to be customised, to allow certain fields to be included | |
6324 | or excluded and extension details. The old system didn't display | |
6325 | multicharacter strings properly, omitted fields not in the policy | |
6326 | and couldn't display additional details such as extensions. | |
6327 | [Steve Henson] | |
6328 | ||
85fb12d5 | 6329 | *) Function EC_POINTs_mul for multiple scalar multiplication |
3ba1f111 BM |
6330 | of an arbitrary number of elliptic curve points |
6331 | \sum scalars[i]*points[i], | |
6332 | optionally including the generator defined for the EC_GROUP: | |
6333 | scalar*generator + \sum scalars[i]*points[i]. | |
6334 | ||
38374911 BM |
6335 | EC_POINT_mul is a simple wrapper function for the typical case |
6336 | that the point list has just one item (besides the optional | |
6337 | generator). | |
48fe4d62 BM |
6338 | [Bodo Moeller] |
6339 | ||
85fb12d5 | 6340 | *) First EC_METHODs for curves over GF(p): |
48fe4d62 BM |
6341 | |
6342 | EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr | |
6343 | operations and provides various method functions that can also | |
6344 | operate with faster implementations of modular arithmetic. | |
6345 | ||
6346 | EC_GFp_mont_method() reuses most functions that are part of | |
6347 | EC_GFp_simple_method, but uses Montgomery arithmetic. | |
6348 | ||
6349 | [Bodo Moeller; point addition and point doubling | |
6350 | implementation directly derived from source code provided by | |
6351 | Lenka Fibikova <fibikova@exp-math.uni-essen.de>] | |
6352 | ||
85fb12d5 | 6353 | *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h, |
48fe4d62 BM |
6354 | crypto/ec/ec_lib.c): |
6355 | ||
6f8f4431 BM |
6356 | Curves are EC_GROUP objects (with an optional group generator) |
6357 | based on EC_METHODs that are built into the library. | |
48fe4d62 BM |
6358 | |
6359 | Points are EC_POINT objects based on EC_GROUP objects. | |
6360 | ||
6361 | Most of the framework would be able to handle curves over arbitrary | |
6f8f4431 BM |
6362 | finite fields, but as there are no obvious types for fields other |
6363 | than GF(p), some functions are limited to that for now. | |
48fe4d62 BM |
6364 | [Bodo Moeller] |
6365 | ||
85fb12d5 | 6366 | *) Add the -HTTP option to s_server. It is similar to -WWW, but requires |
251cb4cf RL |
6367 | that the file contains a complete HTTP response. |
6368 | [Richard Levitte] | |
6369 | ||
85fb12d5 | 6370 | *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl |
b4f682d3 DSH |
6371 | change the def and num file printf format specifier from "%-40sXXX" |
6372 | to "%-39s XXX". The latter will always guarantee a space after the | |
6373 | field while the former will cause them to run together if the field | |
6374 | is 40 of more characters long. | |
6375 | [Steve Henson] | |
6376 | ||
85fb12d5 | 6377 | *) Constify the cipher and digest 'method' functions and structures |
13588350 DSH |
6378 | and modify related functions to take constant EVP_MD and EVP_CIPHER |
6379 | pointers. | |
6380 | [Steve Henson] | |
6381 | ||
85fb12d5 | 6382 | *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them |
48fe4d62 | 6383 | in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32. |
c62b26fd BM |
6384 | [Bodo Moeller] |
6385 | ||
85fb12d5 | 6386 | *) Modify EVP_Digest*() routines so they now return values. Although the |
2dc769a1 DSH |
6387 | internal software routines can never fail additional hardware versions |
6388 | might. | |
6389 | [Steve Henson] | |
6390 | ||
85fb12d5 | 6391 | *) Clean up crypto/err/err.h and change some error codes to avoid conflicts: |
5277d7cb BM |
6392 | |
6393 | Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7 | |
6394 | (= ERR_R_PKCS7_LIB); it is now 64 instead of 32. | |
6395 | ||
6396 | ASN1 error codes | |
6397 | ERR_R_NESTED_ASN1_ERROR | |
6398 | ... | |
6399 | ERR_R_MISSING_ASN1_EOS | |
6400 | were 4 .. 9, conflicting with | |
6401 | ERR_LIB_RSA (= ERR_R_RSA_LIB) | |
6402 | ... | |
6403 | ERR_LIB_PEM (= ERR_R_PEM_LIB). | |
6404 | They are now 58 .. 63 (i.e., just below ERR_R_FATAL). | |
6405 | ||
6406 | Add new error code 'ERR_R_INTERNAL_ERROR'. | |
6407 | [Bodo Moeller] | |
6408 | ||
85fb12d5 | 6409 | *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock |
5277d7cb BM |
6410 | suffices. |
6411 | [Bodo Moeller] | |
6412 | ||
85fb12d5 | 6413 | *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This |
bad40585 BM |
6414 | sets the subject name for a new request or supersedes the |
6415 | subject name in a given request. Formats that can be parsed are | |
6416 | 'CN=Some Name, OU=myOU, C=IT' | |
6417 | and | |
6418 | 'CN=Some Name/OU=myOU/C=IT'. | |
6419 | ||
6420 | Add options '-batch' and '-verbose' to 'openssl req'. | |
6421 | [Massimiliano Pala <madwolf@hackmasters.net>] | |
6422 | ||
85fb12d5 | 6423 | *) Introduce the possibility to access global variables through |
62dc5aad RL |
6424 | functions on platform were that's the best way to handle exporting |
6425 | global variables in shared libraries. To enable this functionality, | |
6426 | one must configure with "EXPORT_VAR_AS_FN" or defined the C macro | |
6427 | "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter | |
6428 | is normally done by Configure or something similar). | |
6429 | ||
6430 | To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL | |
6431 | in the source file (foo.c) like this: | |
6432 | ||
6433 | OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1; | |
6434 | OPENSSL_IMPLEMENT_GLOBAL(double,bar); | |
6435 | ||
6436 | To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL | |
6437 | and OPENSSL_GLOBAL_REF in the header file (foo.h) like this: | |
6438 | ||
6439 | OPENSSL_DECLARE_GLOBAL(int,foo); | |
6440 | #define foo OPENSSL_GLOBAL_REF(foo) | |
6441 | OPENSSL_DECLARE_GLOBAL(double,bar); | |
6442 | #define bar OPENSSL_GLOBAL_REF(bar) | |
6443 | ||
6444 | The #defines are very important, and therefore so is including the | |
e3fefbfd | 6445 | header file everywhere where the defined globals are used. |
62dc5aad RL |
6446 | |
6447 | The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition | |
e3fefbfd | 6448 | of ASN.1 items, but that structure is a bit different. |
62dc5aad RL |
6449 | |
6450 | The largest change is in util/mkdef.pl which has been enhanced with | |
6451 | better and easier to understand logic to choose which symbols should | |
6452 | go into the Windows .def files as well as a number of fixes and code | |
6453 | cleanup (among others, algorithm keywords are now sorted | |
6454 | lexicographically to avoid constant rewrites). | |
6455 | [Richard Levitte] | |
6456 | ||
85fb12d5 | 6457 | *) In BN_div() keep a copy of the sign of 'num' before writing the |
3d2e469c DSH |
6458 | result to 'rm' because if rm==num the value will be overwritten |
6459 | and produce the wrong result if 'num' is negative: this caused | |
6460 | problems with BN_mod() and BN_nnmod(). | |
6461 | [Steve Henson] | |
6462 | ||
85fb12d5 | 6463 | *) Function OCSP_request_verify(). This checks the signature on an |
fafc7f98 DSH |
6464 | OCSP request and verifies the signer certificate. The signer |
6465 | certificate is just checked for a generic purpose and OCSP request | |
6466 | trust settings. | |
6467 | [Steve Henson] | |
6468 | ||
85fb12d5 | 6469 | *) Add OCSP_check_validity() function to check the validity of OCSP |
f1965221 DSH |
6470 | responses. OCSP responses are prepared in real time and may only |
6471 | be a few seconds old. Simply checking that the current time lies | |
6472 | between thisUpdate and nextUpdate max reject otherwise valid responses | |
e3fefbfd | 6473 | caused by either OCSP responder or client clock inaccuracy. Instead |
f1965221 DSH |
6474 | we allow thisUpdate and nextUpdate to fall within a certain period of |
6475 | the current time. The age of the response can also optionally be | |
6476 | checked. Two new options -validity_period and -status_age added to | |
6477 | ocsp utility. | |
6478 | [Steve Henson] | |
6479 | ||
85fb12d5 | 6480 | *) If signature or public key algorithm is unrecognized print out its |
e3fefbfd | 6481 | OID rather that just UNKNOWN. |
4ff18c8c DSH |
6482 | [Steve Henson] |
6483 | ||
85fb12d5 | 6484 | *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and |
d7c06e9e DSH |
6485 | OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate |
6486 | ID to be generated from the issuer certificate alone which can then be | |
6487 | passed to OCSP_id_issuer_cmp(). | |
6488 | [Steve Henson] | |
6489 | ||
85fb12d5 | 6490 | *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new |
386828d0 DSH |
6491 | ASN1 modules to export functions returning ASN1_ITEM pointers |
6492 | instead of the ASN1_ITEM structures themselves. This adds several | |
6493 | new macros which allow the underlying ASN1 function/structure to | |
6494 | be accessed transparently. As a result code should not use ASN1_ITEM | |
6495 | references directly (such as &X509_it) but instead use the relevant | |
6496 | macros (such as ASN1_ITEM_rptr(X509)). This option is to allow | |
6497 | use of the new ASN1 code on platforms where exporting structures | |
6498 | is problematical (for example in shared libraries) but exporting | |
6499 | functions returning pointers to structures is not. | |
6500 | [Steve Henson] | |
6501 | ||
85fb12d5 | 6502 | *) Add support for overriding the generation of SSL/TLS session IDs. |
fa2b8db4 GT |
6503 | These callbacks can be registered either in an SSL_CTX or per SSL. |
6504 | The purpose of this is to allow applications to control, if they wish, | |
6505 | the arbitrary values chosen for use as session IDs, particularly as it | |
6506 | can be useful for session caching in multiple-server environments. A | |
6507 | command-line switch for testing this (and any client code that wishes | |
6508 | to use such a feature) has been added to "s_server". | |
6509 | [Geoff Thorpe, Lutz Jaenicke] | |
6510 | ||
85fb12d5 | 6511 | *) Modify mkdef.pl to recognise and parse preprocessor conditionals |
d399fdf8 RL |
6512 | of the form '#if defined(...) || defined(...) || ...' and |
6513 | '#if !defined(...) && !defined(...) && ...'. This also avoids | |
6514 | the growing number of special cases it was previously handling. | |
6515 | [Richard Levitte] | |
6516 | ||
85fb12d5 | 6517 | *) Make all configuration macros available for application by making |
cf1b7d96 RL |
6518 | sure they are available in opensslconf.h, by giving them names starting |
6519 | with "OPENSSL_" to avoid conflicts with other packages and by making | |
6520 | sure e_os2.h will cover all platform-specific cases together with | |
6521 | opensslconf.h. | |
2affbab9 RL |
6522 | Additionally, it is now possible to define configuration/platform- |
6523 | specific names (called "system identities"). In the C code, these | |
6524 | are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another | |
6525 | macro with the name beginning with "OPENSSL_SYS_", which is determined | |
6526 | from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on | |
6527 | what is available. | |
cf1b7d96 RL |
6528 | [Richard Levitte] |
6529 | ||
85fb12d5 | 6530 | *) New option -set_serial to 'req' and 'x509' this allows the serial |
acba75c5 DSH |
6531 | number to use to be specified on the command line. Previously self |
6532 | signed certificates were hard coded with serial number 0 and the | |
6533 | CA options of 'x509' had to use a serial number in a file which was | |
6534 | auto incremented. | |
6535 | [Steve Henson] | |
6536 | ||
85fb12d5 | 6537 | *) New options to 'ca' utility to support V2 CRL entry extensions. |
a6b7ffdd DSH |
6538 | Currently CRL reason, invalidity date and hold instruction are |
6539 | supported. Add new CRL extensions to V3 code and some new objects. | |
6540 | [Steve Henson] | |
6541 | ||
85fb12d5 | 6542 | *) New function EVP_CIPHER_CTX_set_padding() this is used to |
f2e5ca84 DSH |
6543 | disable standard block padding (aka PKCS#5 padding) in the EVP |
6544 | API, which was previously mandatory. This means that the data is | |
6545 | not padded in any way and so the total length much be a multiple | |
6546 | of the block size, otherwise an error occurs. | |
6547 | [Steve Henson] | |
6548 | ||
85fb12d5 | 6549 | *) Initial (incomplete) OCSP SSL support. |
cdc7b8cc DSH |
6550 | [Steve Henson] |
6551 | ||
85fb12d5 | 6552 | *) New function OCSP_parse_url(). This splits up a URL into its host, |
67c18019 DSH |
6553 | port and path components: primarily to parse OCSP URLs. New -url |
6554 | option to ocsp utility. | |
6555 | [Steve Henson] | |
6556 | ||
85fb12d5 | 6557 | *) New nonce behavior. The return value of OCSP_check_nonce() now |
46a58ab9 DSH |
6558 | reflects the various checks performed. Applications can decide |
6559 | whether to tolerate certain situations such as an absent nonce | |
6560 | in a response when one was present in a request: the ocsp application | |
6561 | just prints out a warning. New function OCSP_add1_basic_nonce() | |
6562 | this is to allow responders to include a nonce in a response even if | |
6563 | the request is nonce-less. | |
6564 | [Steve Henson] | |
6565 | ||
85fb12d5 | 6566 | *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are |
620cea37 BM |
6567 | skipped when using openssl x509 multiple times on a single input file, |
6568 | e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs". | |
6569 | [Bodo Moeller] | |
6570 | ||
85fb12d5 | 6571 | *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string() |
ccb08f98 DSH |
6572 | set string type: to handle setting ASN1_TIME structures. Fix ca |
6573 | utility to correctly initialize revocation date of CRLs. | |
6574 | [Steve Henson] | |
6575 | ||
85fb12d5 | 6576 | *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override |
836f9960 LJ |
6577 | the clients preferred ciphersuites and rather use its own preferences. |
6578 | Should help to work around M$ SGC (Server Gated Cryptography) bug in | |
6579 | Internet Explorer by ensuring unchanged hash method during stepup. | |
b72faddc | 6580 | (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.) |
836f9960 LJ |
6581 | [Lutz Jaenicke] |
6582 | ||
85fb12d5 | 6583 | *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael |
c47c6196 DSH |
6584 | to aes and add a new 'exist' option to print out symbols that don't |
6585 | appear to exist. | |
6586 | [Steve Henson] | |
6587 | ||
85fb12d5 | 6588 | *) Additional options to ocsp utility to allow flags to be set and |
8c950429 DSH |
6589 | additional certificates supplied. |
6590 | [Steve Henson] | |
6591 | ||
85fb12d5 | 6592 | *) Add the option -VAfile to 'openssl ocsp', so the user can give the |
9235adbf RL |
6593 | OCSP client a number of certificate to only verify the response |
6594 | signature against. | |
6595 | [Richard Levitte] | |
6596 | ||
85fb12d5 | 6597 | *) Update Rijndael code to version 3.0 and change EVP AES ciphers to |
deb2c1a1 | 6598 | handle the new API. Currently only ECB, CBC modes supported. Add new |
47234cd3 BM |
6599 | AES OIDs. |
6600 | ||
ea4f109c BM |
6601 | Add TLS AES ciphersuites as described in RFC3268, "Advanced |
6602 | Encryption Standard (AES) Ciphersuites for Transport Layer | |
6603 | Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were | |
6604 | not enabled by default and were not part of the "ALL" ciphersuite | |
6605 | alias because they were not yet official; they could be | |
6606 | explicitly requested by specifying the "AESdraft" ciphersuite | |
6607 | group alias. In the final release of OpenSSL 0.9.7, the group | |
6608 | alias is called "AES" and is part of "ALL".) | |
6609 | [Ben Laurie, Steve Henson, Bodo Moeller] | |
deb2c1a1 | 6610 | |
85fb12d5 | 6611 | *) New function OCSP_copy_nonce() to copy nonce value (if present) from |
26e083cc DSH |
6612 | request to response. |
6613 | [Steve Henson] | |
6614 | ||
85fb12d5 | 6615 | *) Functions for OCSP responders. OCSP_request_onereq_count(), |
02e4fbed DSH |
6616 | OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info() |
6617 | extract information from a certificate request. OCSP_response_create() | |
6618 | creates a response and optionally adds a basic response structure. | |
6619 | OCSP_basic_add1_status() adds a complete single response to a basic | |
e3fefbfd | 6620 | response and returns the OCSP_SINGLERESP structure just added (to allow |
02e4fbed DSH |
6621 | extensions to be included for example). OCSP_basic_add1_cert() adds a |
6622 | certificate to a basic response and OCSP_basic_sign() signs a basic | |
6623 | response with various flags. New helper functions ASN1_TIME_check() | |
6624 | (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime() | |
6625 | (converts ASN1_TIME to GeneralizedTime). | |
6626 | [Steve Henson] | |
6627 | ||
85fb12d5 | 6628 | *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() |
88ce56f8 | 6629 | in a single operation. X509_get0_pubkey_bitstr() extracts the public_key |
e3fefbfd | 6630 | structure from a certificate. X509_pubkey_digest() digests the public_key |
88ce56f8 DSH |
6631 | contents: this is used in various key identifiers. |
6632 | [Steve Henson] | |
6633 | ||
85fb12d5 | 6634 | *) Make sk_sort() tolerate a NULL argument. |
b8470240 DSH |
6635 | [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] |
6636 | ||
85fb12d5 | 6637 | *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates |
50d51991 | 6638 | passed by the function are trusted implicitly. If any of them signed the |
e3fefbfd | 6639 | response then it is assumed to be valid and is not verified. |
50d51991 DSH |
6640 | [Steve Henson] |
6641 | ||
85fb12d5 | 6642 | *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT |
a43cf9fa DSH |
6643 | to data. This was previously part of the PKCS7 ASN1 code. This |
6644 | was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. | |
6645 | [Steve Henson, reported by Kenneth R. Robinette | |
6646 | <support@securenetterm.com>] | |
6647 | ||
85fb12d5 | 6648 | *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 |
a43cf9fa DSH |
6649 | routines: without these tracing memory leaks is very painful. |
6650 | Fix leaks in PKCS12 and PKCS7 routines. | |
6651 | [Steve Henson] | |
6652 | ||
85fb12d5 | 6653 | *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). |
ba8e2824 DSH |
6654 | Previously it initialised the 'type' argument to V_ASN1_UTCTIME which |
6655 | effectively meant GeneralizedTime would never be used. Now it | |
6656 | is initialised to -1 but X509_time_adj() now has to check the value | |
6657 | and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or | |
6658 | V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. | |
6659 | [Steve Henson, reported by Kenneth R. Robinette | |
6660 | <support@securenetterm.com>] | |
6661 | ||
85fb12d5 | 6662 | *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously |
8e8972bb DSH |
6663 | result in a zero length in the ASN1_INTEGER structure which was |
6664 | not consistent with the structure when d2i_ASN1_INTEGER() was used | |
6665 | and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() | |
6666 | to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() | |
6667 | where it did not print out a minus for negative ASN1_INTEGER. | |
6668 | [Steve Henson] | |
6669 | ||
85fb12d5 | 6670 | *) Add summary printout to ocsp utility. The various functions which |
73758d43 DSH |
6671 | convert status values to strings have been renamed to: |
6672 | OCSP_response_status_str(), OCSP_cert_status_str() and | |
6673 | OCSP_crl_reason_str() and are no longer static. New options | |
6674 | to verify nonce values and to disable verification. OCSP response | |
6675 | printout format cleaned up. | |
6676 | [Steve Henson] | |
6677 | ||
85fb12d5 | 6678 | *) Add additional OCSP certificate checks. These are those specified |
e8af92fc DSH |
6679 | in RFC2560. This consists of two separate checks: the CA of the |
6680 | certificate being checked must either be the OCSP signer certificate | |
6681 | or the issuer of the OCSP signer certificate. In the latter case the | |
6682 | OCSP signer certificate must contain the OCSP signing extended key | |
6683 | usage. This check is performed by attempting to match the OCSP | |
6684 | signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash | |
6685 | in the OCSP_CERTID structures of the response. | |
6686 | [Steve Henson] | |
6687 | ||
85fb12d5 | 6688 | *) Initial OCSP certificate verification added to OCSP_basic_verify() |
81f169e9 DSH |
6689 | and related routines. This uses the standard OpenSSL certificate |
6690 | verify routines to perform initial checks (just CA validity) and | |
6691 | to obtain the certificate chain. Then additional checks will be | |
6692 | performed on the chain. Currently the root CA is checked to see | |
6693 | if it is explicitly trusted for OCSP signing. This is used to set | |
6694 | a root CA as a global signing root: that is any certificate that | |
6695 | chains to that CA is an acceptable OCSP signing certificate. | |
6696 | [Steve Henson] | |
6697 | ||
85fb12d5 | 6698 | *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 |
dfebac32 BM |
6699 | extensions from a separate configuration file. |
6700 | As when reading extensions from the main configuration file, | |
6701 | the '-extensions ...' option may be used for specifying the | |
6702 | section to use. | |
6703 | [Massimiliano Pala <madwolf@comune.modena.it>] | |
6704 | ||
85fb12d5 | 6705 | *) New OCSP utility. Allows OCSP requests to be generated or |
5782ceb2 DSH |
6706 | read. The request can be sent to a responder and the output |
6707 | parsed, outputed or printed in text form. Not complete yet: | |
6708 | still needs to check the OCSP response validity. | |
6709 | [Steve Henson] | |
6710 | ||
85fb12d5 | 6711 | *) New subcommands for 'openssl ca': |
c67cdb50 BM |
6712 | 'openssl ca -status <serial>' prints the status of the cert with |
6713 | the given serial number (according to the index file). | |
6714 | 'openssl ca -updatedb' updates the expiry status of certificates | |
6715 | in the index file. | |
6716 | [Massimiliano Pala <madwolf@comune.modena.it>] | |
6717 | ||
85fb12d5 | 6718 | *) New '-newreq-nodes' command option to CA.pl. This is like |
d199858e BM |
6719 | '-newreq', but calls 'openssl req' with the '-nodes' option |
6720 | so that the resulting key is not encrypted. | |
6721 | [Damien Miller <djm@mindrot.org>] | |
6722 | ||
85fb12d5 | 6723 | *) New configuration for the GNU Hurd. |
10a2975a RL |
6724 | [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] |
6725 | ||
85fb12d5 | 6726 | *) Initial code to implement OCSP basic response verify. This |
9b4dc830 DSH |
6727 | is currently incomplete. Currently just finds the signer's |
6728 | certificate and verifies the signature on the response. | |
6729 | [Steve Henson] | |
6730 | ||
85fb12d5 | 6731 | *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in |
673b3fde BM |
6732 | value of OPENSSLDIR. This is available via the new '-d' option |
6733 | to 'openssl version', and is also included in 'openssl version -a'. | |
6734 | [Bodo Moeller] | |
6735 | ||
85fb12d5 | 6736 | *) Allowing defining memory allocation callbacks that will be given |
a5435e8b BM |
6737 | file name and line number information in additional arguments |
6738 | (a const char* and an int). The basic functionality remains, as | |
6739 | well as the original possibility to just replace malloc(), | |
6740 | realloc() and free() by functions that do not know about these | |
6741 | additional arguments. To register and find out the current | |
6742 | settings for extended allocation functions, the following | |
6743 | functions are provided: | |
65a22e8e RL |
6744 | |
6745 | CRYPTO_set_mem_ex_functions | |
6746 | CRYPTO_set_locked_mem_ex_functions | |
6747 | CRYPTO_get_mem_ex_functions | |
6748 | CRYPTO_get_locked_mem_ex_functions | |
6749 | ||
a5435e8b BM |
6750 | These work the same way as CRYPTO_set_mem_functions and friends. |
6751 | CRYPTO_get_[locked_]mem_functions now writes 0 where such an | |
6752 | extended allocation function is enabled. | |
6753 | Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where | |
6754 | a conventional allocation function is enabled. | |
6755 | [Richard Levitte, Bodo Moeller] | |
65a22e8e | 6756 | |
85fb12d5 | 6757 | *) Finish off removing the remaining LHASH function pointer casts. |
3c914840 | 6758 | There should no longer be any prototype-casting required when using |
56a67adb GT |
6759 | the LHASH abstraction, and any casts that remain are "bugs". See |
6760 | the callback types and macros at the head of lhash.h for details | |
6761 | (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). | |
3c914840 GT |
6762 | [Geoff Thorpe] |
6763 | ||
85fb12d5 | 6764 | *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. |
3351b8d0 LJ |
6765 | If /dev/[u]random devices are not available or do not return enough |
6766 | entropy, EGD style sockets (served by EGD or PRNGD) will automatically | |
6767 | be queried. | |
6768 | The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and | |
6769 | /etc/entropy will be queried once each in this sequence, quering stops | |
6770 | when enough entropy was collected without querying more sockets. | |
599c0353 LJ |
6771 | [Lutz Jaenicke] |
6772 | ||
85fb12d5 | 6773 | *) Change the Unix RAND_poll() variant to be able to poll several |
361ef5f4 RL |
6774 | random devices, as specified by DEVRANDOM, until a sufficient amount |
6775 | of data has been collected. We spend at most 10 ms on each file | |
6776 | (select timeout) and read in non-blocking mode. DEVRANDOM now | |
6777 | defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" | |
6778 | (previously it was just the string "/dev/urandom"), so on typical | |
6779 | platforms the 10 ms delay will never occur. | |
6780 | Also separate out the Unix variant to its own file, rand_unix.c. | |
6781 | For VMS, there's a currently-empty rand_vms.c. | |
0c61e299 RL |
6782 | [Richard Levitte] |
6783 | ||
85fb12d5 | 6784 | *) Move OCSP client related routines to ocsp_cl.c. These |
0b33bc65 DSH |
6785 | provide utility functions which an application needing |
6786 | to issue a request to an OCSP responder and analyse the | |
6787 | response will typically need: as opposed to those which an | |
6788 | OCSP responder itself would need which will be added later. | |
6789 | ||
6790 | OCSP_request_sign() signs an OCSP request with an API similar | |
6791 | to PKCS7_sign(). OCSP_response_status() returns status of OCSP | |
6792 | response. OCSP_response_get1_basic() extracts basic response | |
6793 | from response. OCSP_resp_find_status(): finds and extracts status | |
6794 | information from an OCSP_CERTID structure (which will be created | |
6795 | when the request structure is built). These are built from lower | |
6796 | level functions which work on OCSP_SINGLERESP structures but | |
6797 | wont normally be used unless the application wishes to examine | |
6798 | extensions in the OCSP response for example. | |
6799 | ||
6800 | Replace nonce routines with a pair of functions. | |
6801 | OCSP_request_add1_nonce() adds a nonce value and optionally | |
6802 | generates a random value. OCSP_check_nonce() checks the | |
6803 | validity of the nonce in an OCSP response. | |
6804 | [Steve Henson] | |
6805 | ||
85fb12d5 | 6806 | *) Change function OCSP_request_add() to OCSP_request_add0_id(). |
8e961835 DSH |
6807 | This doesn't copy the supplied OCSP_CERTID and avoids the |
6808 | need to free up the newly created id. Change return type | |
6809 | to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. | |
6810 | This can then be used to add extensions to the request. | |
6811 | Deleted OCSP_request_new(), since most of its functionality | |
6812 | is now in OCSP_REQUEST_new() (and the case insensitive name | |
6813 | clash) apart from the ability to set the request name which | |
6814 | will be added elsewhere. | |
6815 | [Steve Henson] | |
6816 | ||
85fb12d5 | 6817 | *) Update OCSP API. Remove obsolete extensions argument from |
bf0d176e DSH |
6818 | various functions. Extensions are now handled using the new |
6819 | OCSP extension code. New simple OCSP HTTP function which | |
6820 | can be used to send requests and parse the response. | |
6821 | [Steve Henson] | |
6822 | ||
85fb12d5 | 6823 | *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new |
ec5add87 DSH |
6824 | ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN |
6825 | uses the special reorder version of SET OF to sort the attributes | |
6826 | and reorder them to match the encoded order. This resolves a long | |
6827 | standing problem: a verify on a PKCS7 structure just after signing | |
6828 | it used to fail because the attribute order did not match the | |
6829 | encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: | |
6830 | it uses the received order. This is necessary to tolerate some broken | |
6831 | software that does not order SET OF. This is handled by encoding | |
6832 | as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) | |
6833 | to produce the required SET OF. | |
6834 | [Steve Henson] | |
6835 | ||
85fb12d5 | 6836 | *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and |
a6574c21 RL |
6837 | OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header |
6838 | files to get correct declarations of the ASN.1 item variables. | |
6839 | [Richard Levitte] | |
6840 | ||
85fb12d5 | 6841 | *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many |
ecbe0781 DSH |
6842 | PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: |
6843 | asn1_check_tlen() would sometimes attempt to use 'ctx' when it was | |
6844 | NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). | |
6845 | New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant | |
6846 | ASN1_ITEM and no wrapper functions. | |
6847 | [Steve Henson] | |
6848 | ||
85fb12d5 | 6849 | *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These |
4e1209eb DSH |
6850 | replace the old function pointer based I/O routines. Change most of |
6851 | the *_d2i_bio() and *_d2i_fp() functions to use these. | |
6852 | [Steve Henson] | |
6853 | ||
85fb12d5 | 6854 | *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor |
3f07fe09 RL |
6855 | lines, recognice more "algorithms" that can be deselected, and make |
6856 | it complain about algorithm deselection that isn't recognised. | |
6857 | [Richard Levitte] | |
6858 | ||
85fb12d5 | 6859 | *) New ASN1 functions to handle dup, sign, verify, digest, pack and |
73e92de5 DSH |
6860 | unpack operations in terms of ASN1_ITEM. Modify existing wrappers |
6861 | to use new functions. Add NO_ASN1_OLD which can be set to remove | |
6862 | some old style ASN1 functions: this can be used to determine if old | |
6863 | code will still work when these eventually go away. | |
09ab755c DSH |
6864 | [Steve Henson] |
6865 | ||
85fb12d5 | 6866 | *) New extension functions for OCSP structures, these follow the |
ec558b65 DSH |
6867 | same conventions as certificates and CRLs. |
6868 | [Steve Henson] | |
6869 | ||
85fb12d5 | 6870 | *) New function X509V3_add1_i2d(). This automatically encodes and |
57d2f217 DSH |
6871 | adds an extension. Its behaviour can be customised with various |
6872 | flags to append, replace or delete. Various wrappers added for | |
6873 | certifcates and CRLs. | |
6874 | [Steve Henson] | |
6875 | ||
85fb12d5 | 6876 | *) Fix to avoid calling the underlying ASN1 print routine when |
5755cab4 DSH |
6877 | an extension cannot be parsed. Correct a typo in the |
6878 | OCSP_SERVICELOC extension. Tidy up print OCSP format. | |
6879 | [Steve Henson] | |
6880 | ||
85fb12d5 | 6881 | *) Make mkdef.pl parse some of the ASN1 macros and add apropriate |
9c67ab2f | 6882 | entries for variables. |
5755cab4 | 6883 | [Steve Henson] |
9c67ab2f | 6884 | |
85fb12d5 | 6885 | *) Add functionality to apps/openssl.c for detecting locking |
3ac82faa BM |
6886 | problems: As the program is single-threaded, all we have |
6887 | to do is register a locking callback using an array for | |
6888 | storing which locks are currently held by the program. | |
3ac82faa BM |
6889 | [Bodo Moeller] |
6890 | ||
85fb12d5 | 6891 | *) Use a lock around the call to CRYPTO_get_ex_new_index() in |
3ac82faa BM |
6892 | SSL_get_ex_data_X509_STORE_idx(), which is used in |
6893 | ssl_verify_cert_chain() and thus can be called at any time | |
6894 | during TLS/SSL handshakes so that thread-safety is essential. | |
6895 | Unfortunately, the ex_data design is not at all suited | |
6896 | for multi-threaded use, so it probably should be abolished. | |
6897 | [Bodo Moeller] | |
6898 | ||
85fb12d5 | 6899 | *) Added Broadcom "ubsec" ENGINE to OpenSSL. |
2a86064f GT |
6900 | [Broadcom, tweaked and integrated by Geoff Thorpe] |
6901 | ||
85fb12d5 | 6902 | *) Move common extension printing code to new function |
2c15d426 | 6903 | X509V3_print_extensions(). Reorganise OCSP print routines and |
c08523d8 | 6904 | implement some needed OCSP ASN1 functions. Add OCSP extensions. |
2c15d426 DSH |
6905 | [Steve Henson] |
6906 | ||
85fb12d5 | 6907 | *) New function X509_signature_print() to remove duplication in some |
de487514 DSH |
6908 | print routines. |
6909 | [Steve Henson] | |
6910 | ||
85fb12d5 | 6911 | *) Add a special meaning when SET OF and SEQUENCE OF flags are both |
06db4253 DSH |
6912 | set (this was treated exactly the same as SET OF previously). This |
6913 | is used to reorder the STACK representing the structure to match the | |
6914 | encoding. This will be used to get round a problem where a PKCS7 | |
6915 | structure which was signed could not be verified because the STACK | |
6916 | order did not reflect the encoded order. | |
6917 | [Steve Henson] | |
6918 | ||
85fb12d5 | 6919 | *) Reimplement the OCSP ASN1 module using the new code. |
36f554d4 DSH |
6920 | [Steve Henson] |
6921 | ||
85fb12d5 | 6922 | *) Update the X509V3 code to permit the use of an ASN1_ITEM structure |
2aff7727 DSH |
6923 | for its ASN1 operations. The old style function pointers still exist |
6924 | for now but they will eventually go away. | |
6925 | [Steve Henson] | |
6926 | ||
85fb12d5 | 6927 | *) Merge in replacement ASN1 code from the ASN1 branch. This almost |
5755cab4 DSH |
6928 | completely replaces the old ASN1 functionality with a table driven |
6929 | encoder and decoder which interprets an ASN1_ITEM structure describing | |
6930 | the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is | |
6931 | largely maintained. Almost all of the old asn1_mac.h macro based ASN1 | |
6932 | has also been converted to the new form. | |
9d6b1ce6 DSH |
6933 | [Steve Henson] |
6934 | ||
85fb12d5 | 6935 | *) Change BN_mod_exp_recp so that negative moduli are tolerated |
8dea52fa BM |
6936 | (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set |
6937 | so that BN_mod_exp_mont and BN_mod_exp_mont_word work | |
6938 | for negative moduli. | |
6939 | [Bodo Moeller] | |
6940 | ||
85fb12d5 | 6941 | *) Fix BN_uadd and BN_usub: Always return non-negative results instead |
8dea52fa BM |
6942 | of not touching the result's sign bit. |
6943 | [Bodo Moeller] | |
6944 | ||
85fb12d5 | 6945 | *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be |
80d89e6a BM |
6946 | set. |
6947 | [Bodo Moeller] | |
6948 | ||
85fb12d5 | 6949 | *) Changed the LHASH code to use prototypes for callbacks, and created |
f1919c3d GT |
6950 | macros to declare and implement thin (optionally static) functions |
6951 | that provide type-safety and avoid function pointer casting for the | |
6952 | type-specific callbacks. | |
6953 | [Geoff Thorpe] | |
6954 | ||
85fb12d5 | 6955 | *) Added Kerberos Cipher Suites to be used with TLS, as written in |
a47b505e | 6956 | RFC 2712. |
33479d27 | 6957 | [Veers Staats <staatsvr@asc.hpc.mil>, |
1946cd8b | 6958 | Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte] |
33479d27 | 6959 | |
85fb12d5 | 6960 | *) Reformat the FAQ so the different questions and answers can be divided |
4b757c83 | 6961 | in sections depending on the subject. |
0ae485dc RL |
6962 | [Richard Levitte] |
6963 | ||
85fb12d5 | 6964 | *) Have the zlib compression code load ZLIB.DLL dynamically under |
20f88b9b RL |
6965 | Windows. |
6966 | [Richard Levitte] | |
6967 | ||
85fb12d5 | 6968 | *) New function BN_mod_sqrt for computing square roots modulo a prime |
aa66eba7 BM |
6969 | (using the probabilistic Tonelli-Shanks algorithm unless |
6970 | p == 3 (mod 4) or p == 5 (mod 8), which are cases that can | |
6971 | be handled deterministically). | |
6b5d39e8 BM |
6972 | [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] |
6973 | ||
85fb12d5 | 6974 | *) Make BN_mod_inverse faster by explicitly handling small quotients |
bdec3c53 BM |
6975 | in the Euclid loop. (Speed gain about 20% for small moduli [256 or |
6976 | 512 bits], about 30% for larger ones [1024 or 2048 bits].) | |
499e167f BM |
6977 | [Bodo Moeller] |
6978 | ||
85fb12d5 | 6979 | *) New function BN_kronecker. |
dcbd0d74 BM |
6980 | [Bodo Moeller] |
6981 | ||
85fb12d5 | 6982 | *) Fix BN_gcd so that it works on negative inputs; the result is |
dcbd0d74 BM |
6983 | positive unless both parameters are zero. |
6984 | Previously something reasonably close to an infinite loop was | |
6985 | possible because numbers could be growing instead of shrinking | |
6986 | in the implementation of Euclid's algorithm. | |
6987 | [Bodo Moeller] | |
6988 | ||
85fb12d5 | 6989 | *) Fix BN_is_word() and BN_is_one() macros to take into account the |
dcbd0d74 BM |
6990 | sign of the number in question. |
6991 | ||
6992 | Fix BN_is_word(a,w) to work correctly for w == 0. | |
6993 | ||
6994 | The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w) | |
6995 | because its test if the absolute value of 'a' equals 'w'. | |
6996 | Note that BN_abs_is_word does *not* handle w == 0 reliably; | |
6997 | it exists mostly for use in the implementations of BN_is_zero(), | |
6998 | BN_is_one(), and BN_is_word(). | |
6999 | [Bodo Moeller] | |
7000 | ||
85fb12d5 | 7001 | *) New function BN_swap. |
78a0c1f1 BM |
7002 | [Bodo Moeller] |
7003 | ||
85fb12d5 | 7004 | *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that |
78a0c1f1 BM |
7005 | the exponentiation functions are more likely to produce reasonable |
7006 | results on negative inputs. | |
7007 | [Bodo Moeller] | |
7008 | ||
85fb12d5 | 7009 | *) Change BN_mod_mul so that the result is always non-negative. |
78a0c1f1 BM |
7010 | Previously, it could be negative if one of the factors was negative; |
7011 | I don't think anyone really wanted that behaviour. | |
7012 | [Bodo Moeller] | |
7013 | ||
85fb12d5 | 7014 | *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c |
1946cd8b | 7015 | (except for exponentiation, which stays in crypto/bn/bn_exp.c, |
78a0c1f1 BM |
7016 | and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c) |
7017 | and add new functions: | |
5acaa495 | 7018 | |
78a0c1f1 BM |
7019 | BN_nnmod |
7020 | BN_mod_sqr | |
7021 | BN_mod_add | |
5acaa495 | 7022 | BN_mod_add_quick |
78a0c1f1 | 7023 | BN_mod_sub |
5acaa495 BM |
7024 | BN_mod_sub_quick |
7025 | BN_mod_lshift1 | |
7026 | BN_mod_lshift1_quick | |
7027 | BN_mod_lshift | |
7028 | BN_mod_lshift_quick | |
7029 | ||
78a0c1f1 | 7030 | These functions always generate non-negative results. |
5acaa495 | 7031 | |
78a0c1f1 BM |
7032 | BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r |
7033 | such that |m| < r < 0, BN_nnmod will output rem + |m| instead). | |
5acaa495 BM |
7034 | |
7035 | BN_mod_XXX_quick(r, a, [b,] m) generates the same result as | |
7036 | BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b] | |
7037 | be reduced modulo m. | |
78a0c1f1 BM |
7038 | [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller] |
7039 | ||
c1862f91 BM |
7040 | #if 0 |
7041 | The following entry accidentily appeared in the CHANGES file | |
7042 | distributed with OpenSSL 0.9.7. The modifications described in | |
7043 | it do *not* apply to OpenSSL 0.9.7. | |
7044 | ||
85fb12d5 | 7045 | *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there |
baa257f1 RL |
7046 | was actually never needed) and in BN_mul(). The removal in BN_mul() |
7047 | required a small change in bn_mul_part_recursive() and the addition | |
1946cd8b UM |
7048 | of the functions bn_cmp_part_words(), bn_sub_part_words() and |
7049 | bn_add_part_words(), which do the same thing as bn_cmp_words(), | |
baa257f1 RL |
7050 | bn_sub_words() and bn_add_words() except they take arrays with |
7051 | differing sizes. | |
7052 | [Richard Levitte] | |
c1862f91 | 7053 | #endif |
baa257f1 | 7054 | |
85fb12d5 | 7055 | *) In 'openssl passwd', verify passwords read from the terminal |
db70a3fd BM |
7056 | unless the '-salt' option is used (which usually means that |
7057 | verification would just waste user's time since the resulting | |
7058 | hash is going to be compared with some given password hash) | |
7059 | or the new '-noverify' option is used. | |
7060 | ||
7061 | This is an incompatible change, but it does not affect | |
7062 | non-interactive use of 'openssl passwd' (passwords on the command | |
7063 | line, '-stdin' option, '-in ...' option) and thus should not | |
7064 | cause any problems. | |
7065 | [Bodo Moeller] | |
7066 | ||
85fb12d5 | 7067 | *) Remove all references to RSAref, since there's no more need for it. |
ccb9643f RL |
7068 | [Richard Levitte] |
7069 | ||
85fb12d5 | 7070 | *) Make DSO load along a path given through an environment variable |
e06433d9 RL |
7071 | (SHLIB_PATH) with shl_load(). |
7072 | [Richard Levitte] | |
7073 | ||
85fb12d5 | 7074 | *) Constify the ENGINE code as a result of BIGNUM constification. |
55b3c877 RL |
7075 | Also constify the RSA code and most things related to it. In a |
7076 | few places, most notable in the depth of the ASN.1 code, ugly | |
7077 | casts back to non-const were required (to be solved at a later | |
7078 | time) | |
10e473e9 RL |
7079 | [Richard Levitte] |
7080 | ||
85fb12d5 | 7081 | *) Make it so the openssl application has all engines loaded by default. |
e7ef1a56 RL |
7082 | [Richard Levitte] |
7083 | ||
85fb12d5 | 7084 | *) Constify the BIGNUM routines a little more. |
020fc820 RL |
7085 | [Richard Levitte] |
7086 | ||
85fb12d5 | 7087 | *) Add the following functions: |
11c0f120 RL |
7088 | |
7089 | ENGINE_load_cswift() | |
7090 | ENGINE_load_chil() | |
7091 | ENGINE_load_atalla() | |
7092 | ENGINE_load_nuron() | |
7093 | ENGINE_load_builtin_engines() | |
7094 | ||
7095 | That way, an application can itself choose if external engines that | |
7096 | are built-in in OpenSSL shall ever be used or not. The benefit is | |
7097 | that applications won't have to be linked with libdl or other dso | |
7098 | libraries unless it's really needed. | |
7099 | ||
7100 | Changed 'openssl engine' to load all engines on demand. | |
7101 | Changed the engine header files to avoid the duplication of some | |
7102 | declarations (they differed!). | |
7103 | [Richard Levitte] | |
7104 | ||
85fb12d5 | 7105 | *) 'openssl engine' can now list capabilities. |
69e7805f RL |
7106 | [Richard Levitte] |
7107 | ||
85fb12d5 | 7108 | *) Better error reporting in 'openssl engine'. |
e264cfe1 RL |
7109 | [Richard Levitte] |
7110 | ||
85fb12d5 | 7111 | *) Never call load_dh_param(NULL) in s_server. |
15d52ddb BM |
7112 | [Bodo Moeller] |
7113 | ||
85fb12d5 | 7114 | *) Add engine application. It can currently list engines by name and |
14c6d27d RL |
7115 | identity, and test if they are actually available. |
7116 | [Richard Levitte] | |
7117 | ||
85fb12d5 | 7118 | *) Improve RPM specification file by forcing symbolic linking and making |
501ebf16 RL |
7119 | sure the installed documentation is also owned by root.root. |
7120 | [Damien Miller <djm@mindrot.org>] | |
7121 | ||
85fb12d5 | 7122 | *) Give the OpenSSL applications more possibilities to make use of |
3aba98e7 RL |
7123 | keys (public as well as private) handled by engines. |
7124 | [Richard Levitte] | |
7125 | ||
85fb12d5 | 7126 | *) Add OCSP code that comes from CertCo. |
7c155330 RL |
7127 | [Richard Levitte] |
7128 | ||
85fb12d5 | 7129 | *) Add VMS support for the Rijndael code. |
5270e702 RL |
7130 | [Richard Levitte] |
7131 | ||
85fb12d5 | 7132 | *) Added untested support for Nuron crypto accelerator. |
5270e702 RL |
7133 | [Ben Laurie] |
7134 | ||
85fb12d5 | 7135 | *) Add support for external cryptographic devices. This code was |
5270e702 RL |
7136 | previously distributed separately as the "engine" branch. |
7137 | [Geoff Thorpe, Richard Levitte] | |
7138 | ||
85fb12d5 | 7139 | *) Rework the filename-translation in the DSO code. It is now possible to |
1df586be GT |
7140 | have far greater control over how a "name" is turned into a filename |
7141 | depending on the operating environment and any oddities about the | |
7142 | different shared library filenames on each system. | |
7143 | [Geoff Thorpe] | |
7144 | ||
85fb12d5 | 7145 | *) Support threads on FreeBSD-elf in Configure. |
53400da7 RL |
7146 | [Richard Levitte] |
7147 | ||
85fb12d5 | 7148 | *) Fix for SHA1 assembly problem with MASM: it produces |
627ec355 DSH |
7149 | warnings about corrupt line number information when assembling |
7150 | with debugging information. This is caused by the overlapping | |
7151 | of two sections. | |
7152 | [Bernd Matthes <mainbug@celocom.de>, Steve Henson] | |
7153 | ||
85fb12d5 | 7154 | *) NCONF changes. |
567f17cf RL |
7155 | NCONF_get_number() has no error checking at all. As a replacement, |
7156 | NCONF_get_number_e() is defined (_e for "error checking") and is | |
7157 | promoted strongly. The old NCONF_get_number is kept around for | |
7158 | binary backward compatibility. | |
7159 | Make it possible for methods to load from something other than a BIO, | |
7160 | by providing a function pointer that is given a name instead of a BIO. | |
7161 | For example, this could be used to load configuration data from an | |
7162 | LDAP server. | |
7163 | [Richard Levitte] | |
7164 | ||
85fb12d5 | 7165 | *) Fix for non blocking accept BIOs. Added new I/O special reason |
924046ce DSH |
7166 | BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs |
7167 | with non blocking I/O was not possible because no retry code was | |
7168 | implemented. Also added new SSL code SSL_WANT_ACCEPT to cover | |
7169 | this case. | |
7170 | [Steve Henson] | |
7171 | ||
85fb12d5 | 7172 | *) Added the beginnings of Rijndael support. |
3ab56511 RL |
7173 | [Ben Laurie] |
7174 | ||
85fb12d5 | 7175 | *) Fix for bug in DirectoryString mask setting. Add support for |
8ca533e3 DSH |
7176 | X509_NAME_print_ex() in 'req' and X509_print_ex() function |
7177 | to allow certificate printing to more controllable, additional | |
7178 | 'certopt' option to 'x509' to allow new printing options to be | |
7179 | set. | |
d0c98589 DSH |
7180 | [Steve Henson] |
7181 | ||
85fb12d5 | 7182 | *) Clean old EAY MD5 hack from e_os.h. |
ef71cb6d RL |
7183 | [Richard Levitte] |
7184 | ||
d5f686d8 | 7185 | Changes between 0.9.6l and 0.9.6m [17 Mar 2004] |
29902449 | 7186 | |
d5f686d8 | 7187 | *) Fix null-pointer assignment in do_change_cipher_spec() revealed |
04fac373 | 7188 | by using the Codenomicon TLS Test Tool (CVE-2004-0079) |
d5f686d8 | 7189 | [Joe Orton, Steve Henson] |
29902449 | 7190 | |
d5f686d8 BM |
7191 | Changes between 0.9.6k and 0.9.6l [04 Nov 2003] |
7192 | ||
7193 | *) Fix additional bug revealed by the NISCC test suite: | |
29902449 | 7194 | |
d5f686d8 | 7195 | Stop bug triggering large recursion when presented with |
04fac373 | 7196 | certain ASN.1 tags (CVE-2003-0851) |
29902449 DSH |
7197 | [Steve Henson] |
7198 | ||
d5f686d8 BM |
7199 | Changes between 0.9.6j and 0.9.6k [30 Sep 2003] |
7200 | ||
29902449 DSH |
7201 | *) Fix various bugs revealed by running the NISCC test suite: |
7202 | ||
7203 | Stop out of bounds reads in the ASN1 code when presented with | |
04fac373 | 7204 | invalid tags (CVE-2003-0543 and CVE-2003-0544). |
29902449 DSH |
7205 | |
7206 | If verify callback ignores invalid public key errors don't try to check | |
7207 | certificate signature with the NULL public key. | |
7208 | ||
7209 | [Steve Henson] | |
ed7f1d0b | 7210 | |
14f3d7c5 DSH |
7211 | *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate |
7212 | if the server requested one: as stated in TLS 1.0 and SSL 3.0 | |
7213 | specifications. | |
7214 | [Steve Henson] | |
7215 | ||
ddc38679 BM |
7216 | *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional |
7217 | extra data after the compression methods not only for TLS 1.0 | |
7218 | but also for SSL 3.0 (as required by the specification). | |
7219 | [Bodo Moeller; problem pointed out by Matthias Loepfe] | |
7220 | ||
02e05594 | 7221 | *) Change X509_certificate_type() to mark the key as exported/exportable |
cf9a88ca RL |
7222 | when it's 512 *bits* long, not 512 bytes. |
7223 | [Richard Levitte] | |
ed7f1d0b | 7224 | |
7a04fdd8 BM |
7225 | Changes between 0.9.6i and 0.9.6j [10 Apr 2003] |
7226 | ||
7227 | *) Countermeasure against the Klima-Pokorny-Rosa extension of | |
7228 | Bleichbacher's attack on PKCS #1 v1.5 padding: treat | |
7229 | a protocol version number mismatch like a decryption error | |
7230 | in ssl3_get_client_key_exchange (ssl/s3_srvr.c). | |
7231 | [Bodo Moeller] | |
7232 | ||
7233 | *) Turn on RSA blinding by default in the default implementation | |
7234 | to avoid a timing attack. Applications that don't want it can call | |
7235 | RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING. | |
7236 | They would be ill-advised to do so in most cases. | |
7237 | [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller] | |
7238 | ||
7239 | *) Change RSA blinding code so that it works when the PRNG is not | |
7240 | seeded (in this case, the secret RSA exponent is abused as | |
7241 | an unpredictable seed -- if it is not unpredictable, there | |
7242 | is no point in blinding anyway). Make RSA blinding thread-safe | |
7243 | by remembering the creator's thread ID in rsa->blinding and | |
7244 | having all other threads use local one-time blinding factors | |
7245 | (this requires more computation than sharing rsa->blinding, but | |
7246 | avoids excessive locking; and if an RSA object is not shared | |
7247 | between threads, blinding will still be very fast). | |
7248 | [Bodo Moeller] | |
7249 | ||
5b0b0e98 RL |
7250 | Changes between 0.9.6h and 0.9.6i [19 Feb 2003] |
7251 | ||
7252 | *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked | |
7253 | via timing by performing a MAC computation even if incorrrect | |
7254 | block cipher padding has been found. This is a countermeasure | |
7255 | against active attacks where the attacker has to distinguish | |
04fac373 | 7256 | between bad padding and a MAC verification error. (CVE-2003-0078) |
5b0b0e98 RL |
7257 | |
7258 | [Bodo Moeller; problem pointed out by Brice Canvel (EPFL), | |
7259 | Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and | |
7260 | Martin Vuagnoux (EPFL, Ilion)] | |
7261 | ||
43ecece5 | 7262 | Changes between 0.9.6g and 0.9.6h [5 Dec 2002] |
2af52de7 | 7263 | |
df29cc8f RL |
7264 | *) New function OPENSSL_cleanse(), which is used to cleanse a section of |
7265 | memory from it's contents. This is done with a counter that will | |
7266 | place alternating values in each byte. This can be used to solve | |
7267 | two issues: 1) the removal of calls to memset() by highly optimizing | |
7268 | compilers, and 2) cleansing with other values than 0, since those can | |
7269 | be read through on certain media, for example a swap space on disk. | |
7270 | [Geoff Thorpe] | |
7271 | ||
6a8afe22 LJ |
7272 | *) Bugfix: client side session caching did not work with external caching, |
7273 | because the session->cipher setting was not restored when reloading | |
7274 | from the external cache. This problem was masked, when | |
7275 | SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set. | |
7276 | (Found by Steve Haslam <steve@araqnid.ddts.net>.) | |
7277 | [Lutz Jaenicke] | |
7278 | ||
0a594209 RL |
7279 | *) Fix client_certificate (ssl/s2_clnt.c): The permissible total |
7280 | length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33. | |
7281 | [Zeev Lieber <zeev-l@yahoo.com>] | |
7282 | ||
84034f7a RL |
7283 | *) Undo an undocumented change introduced in 0.9.6e which caused |
7284 | repeated calls to OpenSSL_add_all_ciphers() and | |
7285 | OpenSSL_add_all_digests() to be ignored, even after calling | |
7286 | EVP_cleanup(). | |
7287 | [Richard Levitte] | |
7288 | ||
83411793 RL |
7289 | *) Change the default configuration reader to deal with last line not |
7290 | being properly terminated. | |
7291 | [Richard Levitte] | |
7292 | ||
c81a1509 RL |
7293 | *) Change X509_NAME_cmp() so it applies the special rules on handling |
7294 | DN values that are of type PrintableString, as well as RDNs of type | |
7295 | emailAddress where the value has the type ia5String. | |
7296 | [stefank@valicert.com via Richard Levitte] | |
7297 | ||
9c3db400 GT |
7298 | *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half |
7299 | the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently | |
7300 | doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be | |
7301 | the bitwise-OR of the two for use by the majority of applications | |
7302 | wanting this behaviour, and update the docs. The documented | |
7303 | behaviour and actual behaviour were inconsistent and had been | |
7304 | changing anyway, so this is more a bug-fix than a behavioural | |
7305 | change. | |
7306 | [Geoff Thorpe, diagnosed by Nadav Har'El] | |
7307 | ||
a4f53a1c BM |
7308 | *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c |
7309 | (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes). | |
7310 | [Bodo Moeller] | |
7311 | ||
e78f1378 | 7312 | *) Fix initialization code race conditions in |
929f1167 BM |
7313 | SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(), |
7314 | SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(), | |
7315 | SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(), | |
7316 | TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(), | |
b8565a9a BM |
7317 | ssl2_get_cipher_by_char(), |
7318 | ssl3_get_cipher_by_char(). | |
e78f1378 | 7319 | [Patrick McCormick <patrick@tellme.com>, Bodo Moeller] |
a90ae024 | 7320 | |
82a20fb0 LJ |
7321 | *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after |
7322 | the cached sessions are flushed, as the remove_cb() might use ex_data | |
7323 | contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com> | |
7324 | (see [openssl.org #212]). | |
7325 | [Geoff Thorpe, Lutz Jaenicke] | |
7326 | ||
2af52de7 DSH |
7327 | *) Fix typo in OBJ_txt2obj which incorrectly passed the content |
7328 | length, instead of the encoding length to d2i_ASN1_OBJECT. | |
7329 | [Steve Henson] | |
7330 | ||
8e28c671 | 7331 | Changes between 0.9.6f and 0.9.6g [9 Aug 2002] |
fbe792f0 | 7332 | |
8e28c671 BM |
7333 | *) [In 0.9.6g-engine release:] |
7334 | Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall'). | |
7335 | [Lynn Gazis <lgazis@rainbow.com>] | |
fbe792f0 RL |
7336 | |
7337 | Changes between 0.9.6e and 0.9.6f [8 Aug 2002] | |
5574e0ed | 7338 | |
f9082268 DSH |
7339 | *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX |
7340 | and get fix the header length calculation. | |
7341 | [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>, | |
7342 | Alon Kantor <alonk@checkpoint.com> (and others), | |
7343 | Steve Henson] | |
7344 | ||
5574e0ed BM |
7345 | *) Use proper error handling instead of 'assertions' in buffer |
7346 | overflow checks added in 0.9.6e. This prevents DoS (the | |
7347 | assertions could call abort()). | |
7348 | [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller] | |
3e06fb75 | 7349 | |
c046fffa LJ |
7350 | Changes between 0.9.6d and 0.9.6e [30 Jul 2002] |
7351 | ||
7352 | *) Add various sanity checks to asn1_get_length() to reject | |
7353 | the ASN1 length bytes if they exceed sizeof(long), will appear | |
7354 | negative or the content length exceeds the length of the | |
7355 | supplied buffer. | |
7356 | [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] | |
46ffee47 | 7357 | |
063a8905 LJ |
7358 | *) Fix cipher selection routines: ciphers without encryption had no flags |
7359 | for the cipher strength set and where therefore not handled correctly | |
7360 | by the selection routines (PR #130). | |
7361 | [Lutz Jaenicke] | |
7362 | ||
46ffee47 BM |
7363 | *) Fix EVP_dsa_sha macro. |
7364 | [Nils Larsch] | |
7365 | ||
c21506ba BM |
7366 | *) New option |
7367 | SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS | |
7368 | for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure | |
7369 | that was added in OpenSSL 0.9.6d. | |
7370 | ||
7371 | As the countermeasure turned out to be incompatible with some | |
7372 | broken SSL implementations, the new option is part of SSL_OP_ALL. | |
7373 | SSL_OP_ALL is usually employed when compatibility with weird SSL | |
7374 | implementations is desired (e.g. '-bugs' option to 's_client' and | |
7375 | 's_server'), so the new option is automatically set in many | |
7376 | applications. | |
7377 | [Bodo Moeller] | |
7378 | ||
c046fffa LJ |
7379 | *) Changes in security patch: |
7380 | ||
7381 | Changes marked "(CHATS)" were sponsored by the Defense Advanced | |
7382 | Research Projects Agency (DARPA) and Air Force Research Laboratory, | |
7383 | Air Force Materiel Command, USAF, under agreement number | |
7384 | F30602-01-2-0537. | |
7385 | ||
7386 | *) Add various sanity checks to asn1_get_length() to reject | |
7387 | the ASN1 length bytes if they exceed sizeof(long), will appear | |
7388 | negative or the content length exceeds the length of the | |
04fac373 | 7389 | supplied buffer. (CVE-2002-0659) |
c046fffa LJ |
7390 | [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>] |
7391 | ||
7392 | *) Assertions for various potential buffer overflows, not known to | |
7393 | happen in practice. | |
7394 | [Ben Laurie (CHATS)] | |
7395 | ||
7396 | *) Various temporary buffers to hold ASCII versions of integers were | |
04fac373 | 7397 | too small for 64 bit platforms. (CVE-2002-0655) |
c046fffa LJ |
7398 | [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)> |
7399 | ||
c046fffa | 7400 | *) Remote buffer overflow in SSL3 protocol - an attacker could |
04fac373 | 7401 | supply an oversized session ID to a client. (CVE-2002-0656) |
c046fffa LJ |
7402 | [Ben Laurie (CHATS)] |
7403 | ||
7404 | *) Remote buffer overflow in SSL2 protocol - an attacker could | |
04fac373 | 7405 | supply an oversized client master key. (CVE-2002-0656) |
c046fffa LJ |
7406 | [Ben Laurie (CHATS)] |
7407 | ||
46ffee47 | 7408 | Changes between 0.9.6c and 0.9.6d [9 May 2002] |
381a146d | 7409 | |
8df61b50 BM |
7410 | *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not |
7411 | encoded as NULL) with id-dsa-with-sha1. | |
7412 | [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller] | |
7413 | ||
1064acaf BM |
7414 | *) Check various X509_...() return values in apps/req.c. |
7415 | [Nils Larsch <nla@trustcenter.de>] | |
7416 | ||
2940a129 LJ |
7417 | *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines: |
7418 | an end-of-file condition would erronously be flagged, when the CRLF | |
7419 | was just at the end of a processed block. The bug was discovered when | |
7420 | processing data through a buffering memory BIO handing the data to a | |
7421 | BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov | |
7422 | <ptsekov@syntrex.com> and Nedelcho Stanev. | |
7423 | [Lutz Jaenicke] | |
7424 | ||
82b0bf0b BM |
7425 | *) Implement a countermeasure against a vulnerability recently found |
7426 | in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment | |
7427 | before application data chunks to avoid the use of known IVs | |
7428 | with data potentially chosen by the attacker. | |
7429 | [Bodo Moeller] | |
7430 | ||
7431 | *) Fix length checks in ssl3_get_client_hello(). | |
7432 | [Bodo Moeller] | |
7433 | ||
7434 | *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently | |
7435 | to prevent ssl3_read_internal() from incorrectly assuming that | |
7436 | ssl3_read_bytes() found application data while handshake | |
7437 | processing was enabled when in fact s->s3->in_read_app_data was | |
7438 | merely automatically cleared during the initial handshake. | |
7439 | [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>] | |
7440 | ||
381a146d LJ |
7441 | *) Fix object definitions for Private and Enterprise: they were not |
7442 | recognized in their shortname (=lowercase) representation. Extend | |
7443 | obj_dat.pl to issue an error when using undefined keywords instead | |
7444 | of silently ignoring the problem (Svenning Sorensen | |
7445 | <sss@sss.dnsalias.net>). | |
7446 | [Lutz Jaenicke] | |
7447 | ||
7448 | *) Fix DH_generate_parameters() so that it works for 'non-standard' | |
7449 | generators, i.e. generators other than 2 and 5. (Previously, the | |
7450 | code did not properly initialise the 'add' and 'rem' values to | |
7451 | BN_generate_prime().) | |
7452 | ||
7453 | In the new general case, we do not insist that 'generator' is | |
7454 | actually a primitive root: This requirement is rather pointless; | |
7455 | a generator of the order-q subgroup is just as good, if not | |
7456 | better. | |
7457 | [Bodo Moeller] | |
7458 | ||
7459 | *) Map new X509 verification errors to alerts. Discovered and submitted by | |
7460 | Tom Wu <tom@arcot.com>. | |
7461 | [Lutz Jaenicke] | |
7462 | ||
7463 | *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from | |
7464 | returning non-zero before the data has been completely received | |
7465 | when using non-blocking I/O. | |
7466 | [Bodo Moeller; problem pointed out by John Hughes] | |
7467 | ||
7468 | *) Some of the ciphers missed the strength entry (SSL_LOW etc). | |
7469 | [Ben Laurie, Lutz Jaenicke] | |
7470 | ||
7471 | *) Fix bug in SSL_clear(): bad sessions were not removed (found by | |
7472 | Yoram Zahavi <YoramZ@gilian.com>). | |
7473 | [Lutz Jaenicke] | |
7474 | ||
7475 | *) Add information about CygWin 1.3 and on, and preserve proper | |
7476 | configuration for the versions before that. | |
7477 | [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte] | |
7478 | ||
7479 | *) Make removal from session cache (SSL_CTX_remove_session()) more robust: | |
7480 | check whether we deal with a copy of a session and do not delete from | |
7481 | the cache in this case. Problem reported by "Izhar Shoshani Levi" | |
7482 | <izhar@checkpoint.com>. | |
7483 | [Lutz Jaenicke] | |
7484 | ||
7485 | *) Do not store session data into the internal session cache, if it | |
7486 | is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP | |
7487 | flag is set). Proposed by Aslam <aslam@funk.com>. | |
7488 | [Lutz Jaenicke] | |
7489 | ||
7490 | *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested | |
7491 | value is 0. | |
7492 | [Richard Levitte] | |
7493 | ||
381a146d LJ |
7494 | *) [In 0.9.6d-engine release:] |
7495 | Fix a crashbug and a logic bug in hwcrhk_load_pubkey(). | |
7496 | [Toomas Kiisk <vix@cyber.ee> via Richard Levitte] | |
7497 | ||
3e06fb75 BM |
7498 | *) Add the configuration target linux-s390x. |
7499 | [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte] | |
7500 | ||
381a146d LJ |
7501 | *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of |
7502 | ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag | |
7503 | variable as an indication that a ClientHello message has been | |
7504 | received. As the flag value will be lost between multiple | |
7505 | invocations of ssl3_accept when using non-blocking I/O, the | |
7506 | function may not be aware that a handshake has actually taken | |
7507 | place, thus preventing a new session from being added to the | |
7508 | session cache. | |
7509 | ||
7510 | To avoid this problem, we now set s->new_session to 2 instead of | |
7511 | using a local variable. | |
7512 | [Lutz Jaenicke, Bodo Moeller] | |
7513 | ||
7514 | *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c) | |
7515 | if the SSL_R_LENGTH_MISMATCH error is detected. | |
7516 | [Geoff Thorpe, Bodo Moeller] | |
7517 | ||
7518 | *) New 'shared_ldflag' column in Configure platform table. | |
7519 | [Richard Levitte] | |
7520 | ||
7521 | *) Fix EVP_CIPHER_mode macro. | |
7522 | ["Dan S. Camper" <dan@bti.net>] | |
7523 | ||
7524 | *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown | |
7525 | type, we must throw them away by setting rr->length to 0. | |
7526 | [D P Chang <dpc@qualys.com>] | |
7527 | ||
7528 | Changes between 0.9.6b and 0.9.6c [21 dec 2001] | |
7529 | ||
7530 | *) Fix BN_rand_range bug pointed out by Dominikus Scherkl | |
7531 | <Dominikus.Scherkl@biodata.com>. (The previous implementation | |
7532 | worked incorrectly for those cases where range = 10..._2 and | |
7533 | 3*range is two bits longer than range.) | |
7534 | [Bodo Moeller] | |
7535 | ||
7536 | *) Only add signing time to PKCS7 structures if it is not already | |
7537 | present. | |
7538 | [Steve Henson] | |
7539 | ||
7540 | *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce", | |
7541 | OBJ_ld_ce should be OBJ_id_ce. | |
7542 | Also some ip-pda OIDs in crypto/objects/objects.txt were | |
7543 | incorrect (cf. RFC 3039). | |
7544 | [Matt Cooper, Frederic Giudicelli, Bodo Moeller] | |
7545 | ||
7546 | *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid() | |
7547 | returns early because it has nothing to do. | |
7548 | [Andy Schneider <andy.schneider@bjss.co.uk>] | |
7549 | ||
7550 | *) [In 0.9.6c-engine release:] | |
7551 | Fix mutex callback return values in crypto/engine/hw_ncipher.c. | |
7552 | [Andy Schneider <andy.schneider@bjss.co.uk>] | |
7553 | ||
7554 | *) [In 0.9.6c-engine release:] | |
7555 | Add support for Cryptographic Appliance's keyserver technology. | |
7556 | (Use engine 'keyclient') | |
7557 | [Cryptographic Appliances and Geoff Thorpe] | |
7558 | ||
7559 | *) Add a configuration entry for OS/390 Unix. The C compiler 'c89' | |
7560 | is called via tools/c89.sh because arguments have to be | |
7561 | rearranged (all '-L' options must appear before the first object | |
7562 | modules). | |
7563 | [Richard Shapiro <rshapiro@abinitio.com>] | |
7564 | ||
7565 | *) [In 0.9.6c-engine release:] | |
7566 | Add support for Broadcom crypto accelerator cards, backported | |
7567 | from 0.9.7. | |
7568 | [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox] | |
7569 | ||
7570 | *) [In 0.9.6c-engine release:] | |
7571 | Add support for SureWare crypto accelerator cards from | |
7572 | Baltimore Technologies. (Use engine 'sureware') | |
7573 | [Baltimore Technologies and Mark Cox] | |
7574 | ||
7575 | *) [In 0.9.6c-engine release:] | |
7576 | Add support for crypto accelerator cards from Accelerated | |
7577 | Encryption Processing, www.aep.ie. (Use engine 'aep') | |
7578 | [AEP Inc. and Mark Cox] | |
7579 | ||
7580 | *) Add a configuration entry for gcc on UnixWare. | |
7581 | [Gary Benson <gbenson@redhat.com>] | |
7582 | ||
7583 | *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake | |
7584 | messages are stored in a single piece (fixed-length part and | |
7585 | variable-length part combined) and fix various bugs found on the way. | |
7586 | [Bodo Moeller] | |
7587 | ||
7588 | *) Disable caching in BIO_gethostbyname(), directly use gethostbyname() | |
7589 | instead. BIO_gethostbyname() does not know what timeouts are | |
7590 | appropriate, so entries would stay in cache even when they have | |
7591 | become invalid. | |
7592 | [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com> | |
7593 | ||
7594 | *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when | |
7595 | faced with a pathologically small ClientHello fragment that does | |
7596 | not contain client_version: Instead of aborting with an error, | |
7597 | simply choose the highest available protocol version (i.e., | |
7598 | TLS 1.0 unless it is disabled). In practice, ClientHello | |
7599 | messages are never sent like this, but this change gives us | |
7600 | strictly correct behaviour at least for TLS. | |
7601 | [Bodo Moeller] | |
7602 | ||
7603 | *) Fix SSL handshake functions and SSL_clear() such that SSL_clear() | |
7604 | never resets s->method to s->ctx->method when called from within | |
7605 | one of the SSL handshake functions. | |
7606 | [Bodo Moeller; problem pointed out by Niko Baric] | |
7607 | ||
7608 | *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert | |
7609 | (sent using the client's version number) if client_version is | |
7610 | smaller than the protocol version in use. Also change | |
7611 | ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if | |
7612 | the client demanded SSL 3.0 but only TLS 1.0 is enabled; then | |
7613 | the client will at least see that alert. | |
7614 | [Bodo Moeller] | |
7615 | ||
7616 | *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation | |
7617 | correctly. | |
7618 | [Bodo Moeller] | |
7619 | ||
7620 | *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a | |
7621 | client receives HelloRequest while in a handshake. | |
7622 | [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>] | |
7623 | ||
7624 | *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C | |
7625 | should end in 'break', not 'goto end' which circuments various | |
7626 | cleanups done in state SSL_ST_OK. But session related stuff | |
7627 | must be disabled for SSL_ST_OK in the case that we just sent a | |
7628 | HelloRequest. | |
7629 | ||
7630 | Also avoid some overhead by not calling ssl_init_wbio_buffer() | |
7631 | before just sending a HelloRequest. | |
7632 | [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>] | |
7633 | ||
7634 | *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't | |
7635 | reveal whether illegal block cipher padding was found or a MAC | |
7636 | verification error occured. (Neither SSLerr() codes nor alerts | |
7637 | are directly visible to potential attackers, but the information | |
7638 | may leak via logfiles.) | |
7639 | ||
7640 | Similar changes are not required for the SSL 2.0 implementation | |
7641 | because the number of padding bytes is sent in clear for SSL 2.0, | |
7642 | and the extra bytes are just ignored. However ssl/s2_pkt.c | |
7643 | failed to verify that the purported number of padding bytes is in | |
7644 | the legal range. | |
7645 | [Bodo Moeller] | |
7646 | ||
7647 | *) Add OpenUNIX-8 support including shared libraries | |
7648 | (Boyd Lynn Gerber <gerberb@zenez.com>). | |
7649 | [Lutz Jaenicke] | |
7650 | ||
7651 | *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid | |
7652 | 'wristwatch attack' using huge encoding parameters (cf. | |
7653 | James H. Manger's CRYPTO 2001 paper). Note that the | |
7654 | RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use | |
7655 | encoding parameters and hence was not vulnerable. | |
7656 | [Bodo Moeller] | |
7657 | ||
7658 | *) BN_sqr() bug fix. | |
9f0b86c6 | 7659 | [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>] |
381a146d LJ |
7660 | |
7661 | *) Rabin-Miller test analyses assume uniformly distributed witnesses, | |
7662 | so use BN_pseudo_rand_range() instead of using BN_pseudo_rand() | |
7663 | followed by modular reduction. | |
7664 | [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>] | |
7665 | ||
7666 | *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range() | |
7667 | equivalent based on BN_pseudo_rand() instead of BN_rand(). | |
7668 | [Bodo Moeller] | |
7669 | ||
7670 | *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB). | |
7671 | This function was broken, as the check for a new client hello message | |
7672 | to handle SGC did not allow these large messages. | |
7673 | (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.) | |
7674 | [Lutz Jaenicke] | |
7675 | ||
7676 | *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long](). | |
7677 | [Lutz Jaenicke] | |
7678 | ||
7679 | *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl() | |
7680 | for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>). | |
7681 | [Lutz Jaenicke] | |
7682 | ||
7683 | *) Rework the configuration and shared library support for Tru64 Unix. | |
7684 | The configuration part makes use of modern compiler features and | |
7685 | still retains old compiler behavior for those that run older versions | |
7686 | of the OS. The shared library support part includes a variant that | |
7687 | uses the RPATH feature, and is available through the special | |
7688 | configuration target "alpha-cc-rpath", which will never be selected | |
7689 | automatically. | |
7690 | [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte] | |
7691 | ||
7692 | *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message() | |
7693 | with the same message size as in ssl3_get_certificate_request(). | |
7694 | Otherwise, if no ServerKeyExchange message occurs, CertificateRequest | |
7695 | messages might inadvertently be reject as too long. | |
7696 | [Petr Lampa <lampa@fee.vutbr.cz>] | |
7697 | ||
7698 | *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX). | |
7699 | [Andy Polyakov] | |
7700 | ||
7701 | *) Modified SSL library such that the verify_callback that has been set | |
7702 | specificly for an SSL object with SSL_set_verify() is actually being | |
7703 | used. Before the change, a verify_callback set with this function was | |
7704 | ignored and the verify_callback() set in the SSL_CTX at the time of | |
7705 | the call was used. New function X509_STORE_CTX_set_verify_cb() introduced | |
7706 | to allow the necessary settings. | |
7707 | [Lutz Jaenicke] | |
7708 | ||
7709 | *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c | |
7710 | explicitly to NULL, as at least on Solaris 8 this seems not always to be | |
7711 | done automatically (in contradiction to the requirements of the C | |
7712 | standard). This made problems when used from OpenSSH. | |
7713 | [Lutz Jaenicke] | |
7714 | ||
7715 | *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored | |
7716 | dh->length and always used | |
7717 | ||
7718 | BN_rand_range(priv_key, dh->p). | |
7719 | ||
7720 | BN_rand_range() is not necessary for Diffie-Hellman, and this | |
7721 | specific range makes Diffie-Hellman unnecessarily inefficient if | |
7722 | dh->length (recommended exponent length) is much smaller than the | |
7723 | length of dh->p. We could use BN_rand_range() if the order of | |
7724 | the subgroup was stored in the DH structure, but we only have | |
7725 | dh->length. | |
7726 | ||
7727 | So switch back to | |
7728 | ||
7729 | BN_rand(priv_key, l, ...) | |
7730 | ||
7731 | where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1 | |
7732 | otherwise. | |
7733 | [Bodo Moeller] | |
7734 | ||
7735 | *) In | |
7736 | ||
7737 | RSA_eay_public_encrypt | |
7738 | RSA_eay_private_decrypt | |
7739 | RSA_eay_private_encrypt (signing) | |
7740 | RSA_eay_public_decrypt (signature verification) | |
7741 | ||
7742 | (default implementations for RSA_public_encrypt, | |
7743 | RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt), | |
7744 | always reject numbers >= n. | |
7745 | [Bodo Moeller] | |
7746 | ||
7747 | *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2 | |
7748 | to synchronize access to 'locking_thread'. This is necessary on | |
7749 | systems where access to 'locking_thread' (an 'unsigned long' | |
7750 | variable) is not atomic. | |
7751 | [Bodo Moeller] | |
7752 | ||
7753 | *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID | |
7754 | *before* setting the 'crypto_lock_rand' flag. The previous code had | |
7755 | a race condition if 0 is a valid thread ID. | |
7756 | [Travis Vitek <vitek@roguewave.com>] | |
7757 | ||
7758 | *) Add support for shared libraries under Irix. | |
7759 | [Albert Chin-A-Young <china@thewrittenword.com>] | |
7760 | ||
7761 | *) Add configuration option to build on Linux on both big-endian and | |
7762 | little-endian MIPS. | |
7763 | [Ralf Baechle <ralf@uni-koblenz.de>] | |
7764 | ||
7765 | *) Add the possibility to create shared libraries on HP-UX. | |
7766 | [Richard Levitte] | |
7767 | ||
7768 | Changes between 0.9.6a and 0.9.6b [9 Jul 2001] | |
7769 | ||
7770 | *) Change ssleay_rand_bytes (crypto/rand/md_rand.c) | |
7771 | to avoid a SSLeay/OpenSSL PRNG weakness pointed out by | |
7772 | Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>: | |
7773 | PRNG state recovery was possible based on the output of | |
7774 | one PRNG request appropriately sized to gain knowledge on | |
7775 | 'md' followed by enough consecutive 1-byte PRNG requests | |
7776 | to traverse all of 'state'. | |
7777 | ||
7778 | 1. When updating 'md_local' (the current thread's copy of 'md') | |
7779 | during PRNG output generation, hash all of the previous | |
7780 | 'md_local' value, not just the half used for PRNG output. | |
7781 | ||
7782 | 2. Make the number of bytes from 'state' included into the hash | |
7783 | independent from the number of PRNG bytes requested. | |
7784 | ||
7785 | The first measure alone would be sufficient to avoid | |
7786 | Markku-Juhani's attack. (Actually it had never occurred | |
7787 | to me that the half of 'md_local' used for chaining was the | |
7788 | half from which PRNG output bytes were taken -- I had always | |
7789 | assumed that the secret half would be used.) The second | |
7790 | measure makes sure that additional data from 'state' is never | |
7791 | mixed into 'md_local' in small portions; this heuristically | |
7792 | further strengthens the PRNG. | |
7793 | [Bodo Moeller] | |
7794 | ||
7795 | *) Fix crypto/bn/asm/mips3.s. | |
7796 | [Andy Polyakov] | |
7797 | ||
7798 | *) When only the key is given to "enc", the IV is undefined. Print out | |
7799 | an error message in this case. | |
7800 | [Lutz Jaenicke] | |
7801 | ||
7802 | *) Handle special case when X509_NAME is empty in X509 printing routines. | |
7803 | [Steve Henson] | |
7804 | ||
7805 | *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are | |
7806 | positive and less than q. | |
7807 | [Bodo Moeller] | |
7808 | ||
7809 | *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is | |
7810 | used: it isn't thread safe and the add_lock_callback should handle | |
7811 | that itself. | |
7812 | [Paul Rose <Paul.Rose@bridge.com>] | |
7813 | ||
7814 | *) Verify that incoming data obeys the block size in | |
7815 | ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c). | |
7816 | [Bodo Moeller] | |
7817 | ||
7818 | *) Fix OAEP check. | |
9f0b86c6 | 7819 | [Ulf Möller, Bodo Möller] |
381a146d LJ |
7820 | |
7821 | *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5 | |
7822 | RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5 | |
7823 | when fixing the server behaviour for backwards-compatible 'client | |
7824 | hello' messages. (Note that the attack is impractical against | |
7825 | SSL 3.0 and TLS 1.0 anyway because length and version checking | |
7826 | means that the probability of guessing a valid ciphertext is | |
7827 | around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98 | |
7828 | paper.) | |
7829 | ||
7830 | Before 0.9.5, the countermeasure (hide the error by generating a | |
7831 | random 'decryption result') did not work properly because | |
7832 | ERR_clear_error() was missing, meaning that SSL_get_error() would | |
7833 | detect the supposedly ignored error. | |
7834 | ||
7835 | Both problems are now fixed. | |
7836 | [Bodo Moeller] | |
7837 | ||
7838 | *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096 | |
7839 | (previously it was 1024). | |
7840 | [Bodo Moeller] | |
7841 | ||
7842 | *) Fix for compatibility mode trust settings: ignore trust settings | |
7843 | unless some valid trust or reject settings are present. | |
7844 | [Steve Henson] | |
7845 | ||
7846 | *) Fix for blowfish EVP: its a variable length cipher. | |
7847 | [Steve Henson] | |
7848 | ||
7849 | *) Fix various bugs related to DSA S/MIME verification. Handle missing | |
7850 | parameters in DSA public key structures and return an error in the | |
7851 | DSA routines if parameters are absent. | |
7852 | [Steve Henson] | |
7853 | ||
7854 | *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd" | |
7855 | in the current directory if neither $RANDFILE nor $HOME was set. | |
7856 | RAND_file_name() in 0.9.6a returned NULL in this case. This has | |
7857 | caused some confusion to Windows users who haven't defined $HOME. | |
7858 | Thus RAND_file_name() is changed again: e_os.h can define a | |
7859 | DEFAULT_HOME, which will be used if $HOME is not set. | |
7860 | For Windows, we use "C:"; on other platforms, we still require | |
7861 | environment variables. | |
7862 | ||
7863 | *) Move 'if (!initialized) RAND_poll()' into regions protected by | |
7864 | CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids | |
7865 | having multiple threads call RAND_poll() concurrently. | |
7866 | [Bodo Moeller] | |
7867 | ||
7868 | *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a | |
7869 | combination of a flag and a thread ID variable. | |
7870 | Otherwise while one thread is in ssleay_rand_bytes (which sets the | |
7871 | flag), *other* threads can enter ssleay_add_bytes without obeying | |
7872 | the CRYPTO_LOCK_RAND lock (and may even illegally release the lock | |
7873 | that they do not hold after the first thread unsets add_do_not_lock). | |
7874 | [Bodo Moeller] | |
7875 | ||
7876 | *) Change bctest again: '-x' expressions are not available in all | |
7877 | versions of 'test'. | |
7878 | [Bodo Moeller] | |
7879 | ||
7880 | Changes between 0.9.6 and 0.9.6a [5 Apr 2001] | |
7881 | ||
7882 | *) Fix a couple of memory leaks in PKCS7_dataDecode() | |
7883 | [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>] | |
7884 | ||
7885 | *) Change Configure and Makefiles to provide EXE_EXT, which will contain | |
7886 | the default extension for executables, if any. Also, make the perl | |
7887 | scripts that use symlink() to test if it really exists and use "cp" | |
7888 | if it doesn't. All this made OpenSSL compilable and installable in | |
7889 | CygWin. | |
7890 | [Richard Levitte] | |
7891 | ||
7892 | *) Fix for asn1_GetSequence() for indefinite length constructed data. | |
7893 | If SEQUENCE is length is indefinite just set c->slen to the total | |
7894 | amount of data available. | |
7895 | [Steve Henson, reported by shige@FreeBSD.org] | |
7896 | [This change does not apply to 0.9.7.] | |
7897 | ||
7898 | *) Change bctest to avoid here-documents inside command substitution | |
7899 | (workaround for FreeBSD /bin/sh bug). | |
7900 | For compatibility with Ultrix, avoid shell functions (introduced | |
7901 | in the bctest version that searches along $PATH). | |
7902 | [Bodo Moeller] | |
7903 | ||
7904 | *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes | |
7905 | with des_encrypt() defined on some operating systems, like Solaris | |
7906 | and UnixWare. | |
7907 | [Richard Levitte] | |
7908 | ||
7909 | *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton: | |
7910 | On the Importance of Eliminating Errors in Cryptographic | |
7911 | Computations, J. Cryptology 14 (2001) 2, 101-119, | |
7912 | http://theory.stanford.edu/~dabo/papers/faults.ps.gz). | |
7913 | [Ulf Moeller] | |
7914 | ||
7915 | *) MIPS assembler BIGNUM division bug fix. | |
7916 | [Andy Polyakov] | |
7917 | ||
7918 | *) Disabled incorrect Alpha assembler code. | |
7919 | [Richard Levitte] | |
7920 | ||
7921 | *) Fix PKCS#7 decode routines so they correctly update the length | |
7922 | after reading an EOC for the EXPLICIT tag. | |
7923 | [Steve Henson] | |
7924 | [This change does not apply to 0.9.7.] | |
7925 | ||
7926 | *) Fix bug in PKCS#12 key generation routines. This was triggered | |
7927 | if a 3DES key was generated with a 0 initial byte. Include | |
7928 | PKCS12_BROKEN_KEYGEN compilation option to retain the old | |
7929 | (but broken) behaviour. | |
7930 | [Steve Henson] | |
7931 | ||
7932 | *) Enhance bctest to search for a working bc along $PATH and print | |
7933 | it when found. | |
7934 | [Tim Rice <tim@multitalents.net> via Richard Levitte] | |
7935 | ||
7936 | *) Fix memory leaks in err.c: free err_data string if necessary; | |
7937 | don't write to the wrong index in ERR_set_error_data. | |
7938 | [Bodo Moeller] | |
7939 | ||
7940 | *) Implement ssl23_peek (analogous to ssl23_read), which previously | |
7941 | did not exist. | |
7942 | [Bodo Moeller] | |
7943 | ||
7944 | *) Replace rdtsc with _emit statements for VC++ version 5. | |
7945 | [Jeremy Cooper <jeremy@baymoo.org>] | |
7946 | ||
7947 | *) Make it possible to reuse SSLv2 sessions. | |
7948 | [Richard Levitte] | |
7949 | ||
7950 | *) In copy_email() check for >= 0 as a return value for | |
7951 | X509_NAME_get_index_by_NID() since 0 is a valid index. | |
7952 | [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>] | |
7953 | ||
7954 | *) Avoid coredump with unsupported or invalid public keys by checking if | |
7955 | X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when | |
7956 | PKCS7_verify() fails with non detached data. | |
7957 | [Steve Henson] | |
7958 | ||
7959 | *) Don't use getenv in library functions when run as setuid/setgid. | |
7960 | New function OPENSSL_issetugid(). | |
7961 | [Ulf Moeller] | |
7962 | ||
7963 | *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c) | |
7964 | due to incorrect handling of multi-threading: | |
7965 | ||
7966 | 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl(). | |
7967 | ||
7968 | 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on(). | |
7969 | ||
7970 | 3. Count how many times MemCheck_off() has been called so that | |
7971 | nested use can be treated correctly. This also avoids | |
7972 | inband-signalling in the previous code (which relied on the | |
7973 | assumption that thread ID 0 is impossible). | |
7974 | [Bodo Moeller] | |
7975 | ||
7976 | *) Add "-rand" option also to s_client and s_server. | |
7977 | [Lutz Jaenicke] | |
7978 | ||
7979 | *) Fix CPU detection on Irix 6.x. | |
7980 | [Kurt Hockenbury <khockenb@stevens-tech.edu> and | |
7981 | "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] | |
7982 | ||
7983 | *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME | |
7984 | was empty. | |
7985 | [Steve Henson] | |
7986 | [This change does not apply to 0.9.7.] | |
7987 | ||
7988 | *) Use the cached encoding of an X509_NAME structure rather than | |
7989 | copying it. This is apparently the reason for the libsafe "errors" | |
7990 | but the code is actually correct. | |
7991 | [Steve Henson] | |
7992 | ||
7993 | *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent | |
7994 | Bleichenbacher's DSA attack. | |
7995 | Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits | |
7996 | to be set and top=0 forces the highest bit to be set; top=-1 is new | |
7997 | and leaves the highest bit random. | |
7998 | [Ulf Moeller, Bodo Moeller] | |
7999 | ||
8000 | *) In the NCONF_...-based implementations for CONF_... queries | |
8001 | (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using | |
8002 | a temporary CONF structure with the data component set to NULL | |
8003 | (which gives segmentation faults in lh_retrieve). | |
8004 | Instead, use NULL for the CONF pointer in CONF_get_string and | |
8005 | CONF_get_number (which may use environment variables) and directly | |
8006 | return NULL from CONF_get_section. | |
8007 | [Bodo Moeller] | |
8008 | ||
8009 | *) Fix potential buffer overrun for EBCDIC. | |
8010 | [Ulf Moeller] | |
8011 | ||
8012 | *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign | |
8013 | keyUsage if basicConstraints absent for a CA. | |
8014 | [Steve Henson] | |
8015 | ||
8016 | *) Make SMIME_write_PKCS7() write mail header values with a format that | |
8017 | is more generally accepted (no spaces before the semicolon), since | |
8018 | some programs can't parse those values properly otherwise. Also make | |
8019 | sure BIO's that break lines after each write do not create invalid | |
8020 | headers. | |
8021 | [Richard Levitte] | |
8022 | ||
8023 | *) Make the CRL encoding routines work with empty SEQUENCE OF. The | |
8024 | macros previously used would not encode an empty SEQUENCE OF | |
8025 | and break the signature. | |
8026 | [Steve Henson] | |
8027 | [This change does not apply to 0.9.7.] | |
8028 | ||
8029 | *) Zero the premaster secret after deriving the master secret in | |
8030 | DH ciphersuites. | |
8031 | [Steve Henson] | |
8032 | ||
8033 | *) Add some EVP_add_digest_alias registrations (as found in | |
8034 | OpenSSL_add_all_digests()) to SSL_library_init() | |
8035 | aka OpenSSL_add_ssl_algorithms(). This provides improved | |
8036 | compatibility with peers using X.509 certificates | |
8037 | with unconventional AlgorithmIdentifier OIDs. | |
8038 | [Bodo Moeller] | |
8039 | ||
8040 | *) Fix for Irix with NO_ASM. | |
8041 | ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] | |
8042 | ||
8043 | *) ./config script fixes. | |
8044 | [Ulf Moeller, Richard Levitte] | |
8045 | ||
8046 | *) Fix 'openssl passwd -1'. | |
8047 | [Bodo Moeller] | |
8048 | ||
8049 | *) Change PKCS12_key_gen_asc() so it can cope with non null | |
8050 | terminated strings whose length is passed in the passlen | |
8051 | parameter, for example from PEM callbacks. This was done | |
8052 | by adding an extra length parameter to asc2uni(). | |
8053 | [Steve Henson, reported by <oddissey@samsung.co.kr>] | |
8054 | ||
8055 | *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn | |
8056 | call failed, free the DSA structure. | |
8057 | [Bodo Moeller] | |
8058 | ||
8059 | *) Fix to uni2asc() to cope with zero length Unicode strings. | |
8060 | These are present in some PKCS#12 files. | |
8061 | [Steve Henson] | |
8062 | ||
8063 | *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). | |
8064 | Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits | |
8065 | when writing a 32767 byte record. | |
8066 | [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] | |
8067 | ||
8068 | *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), | |
8069 | obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. | |
8070 | ||
8071 | (RSA objects have a reference count access to which is protected | |
8072 | by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], | |
8073 | so they are meant to be shared between threads.) | |
8074 | [Bodo Moeller, Geoff Thorpe; original patch submitted by | |
8075 | "Reddie, Steven" <Steven.Reddie@ca.com>] | |
8076 | ||
8077 | *) Fix a deadlock in CRYPTO_mem_leaks(). | |
8078 | [Bodo Moeller] | |
8079 | ||
8080 | *) Use better test patterns in bntest. | |
9f0b86c6 | 8081 | [Ulf Möller] |
381a146d LJ |
8082 | |
8083 | *) rand_win.c fix for Borland C. | |
9f0b86c6 | 8084 | [Ulf Möller] |
381a146d LJ |
8085 | |
8086 | *) BN_rshift bugfix for n == 0. | |
8087 | [Bodo Moeller] | |
8088 | ||
8089 | *) Add a 'bctest' script that checks for some known 'bc' bugs | |
8090 | so that 'make test' does not abort just because 'bc' is broken. | |
8091 | [Bodo Moeller] | |
8092 | ||
8093 | *) Store verify_result within SSL_SESSION also for client side to | |
8094 | avoid potential security hole. (Re-used sessions on the client side | |
8095 | always resulted in verify_result==X509_V_OK, not using the original | |
8096 | result of the server certificate verification.) | |
8097 | [Lutz Jaenicke] | |
8098 | ||
8099 | *) Fix ssl3_pending: If the record in s->s3->rrec is not of type | |
8100 | SSL3_RT_APPLICATION_DATA, return 0. | |
8101 | Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true. | |
8102 | [Bodo Moeller] | |
8103 | ||
8104 | *) Fix SSL_peek: | |
8105 | Both ssl2_peek and ssl3_peek, which were totally broken in earlier | |
8106 | releases, have been re-implemented by renaming the previous | |
8107 | implementations of ssl2_read and ssl3_read to ssl2_read_internal | |
8108 | and ssl3_read_internal, respectively, and adding 'peek' parameters | |
8109 | to them. The new ssl[23]_{read,peek} functions are calls to | |
8110 | ssl[23]_read_internal with the 'peek' flag set appropriately. | |
8111 | A 'peek' parameter has also been added to ssl3_read_bytes, which | |
8112 | does the actual work for ssl3_read_internal. | |
8113 | [Bodo Moeller] | |
8114 | ||
8115 | *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling | |
8116 | the method-specific "init()" handler. Also clean up ex_data after | |
8117 | calling the method-specific "finish()" handler. Previously, this was | |
8118 | happening the other way round. | |
8119 | [Geoff Thorpe] | |
8120 | ||
8121 | *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16. | |
8122 | The previous value, 12, was not always sufficient for BN_mod_exp(). | |
8123 | [Bodo Moeller] | |
8124 | ||
8125 | *) Make sure that shared libraries get the internal name engine with | |
8126 | the full version number and not just 0. This should mark the | |
8127 | shared libraries as not backward compatible. Of course, this should | |
8128 | be changed again when we can guarantee backward binary compatibility. | |
8129 | [Richard Levitte] | |
8130 | ||
8131 | *) Fix typo in get_cert_by_subject() in by_dir.c | |
8132 | [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>] | |
8133 | ||
8134 | *) Rework the system to generate shared libraries: | |
8135 | ||
8136 | - Make note of the expected extension for the shared libraries and | |
8137 | if there is a need for symbolic links from for example libcrypto.so.0 | |
8138 | to libcrypto.so.0.9.7. There is extended info in Configure for | |
8139 | that. | |
8140 | ||
8141 | - Make as few rebuilds of the shared libraries as possible. | |
8142 | ||
8143 | - Still avoid linking the OpenSSL programs with the shared libraries. | |
8144 | ||
8145 | - When installing, install the shared libraries separately from the | |
8146 | static ones. | |
8147 | [Richard Levitte] | |
8148 | ||
3a0afe1e BM |
8149 | *) Fix SSL_CTX_set_read_ahead macro to actually use its argument. |
8150 | ||
8151 | Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new | |
8152 | and not in SSL_clear because the latter is also used by the | |
8153 | accept/connect functions; previously, the settings made by | |
8154 | SSL_set_read_ahead would be lost during the handshake. | |
8155 | [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>] | |
8156 | ||
88aeb646 RL |
8157 | *) Correct util/mkdef.pl to be selective about disabled algorithms. |
8158 | Previously, it would create entries for disableed algorithms no | |
8159 | matter what. | |
8160 | [Richard Levitte] | |
c5e8580e | 8161 | |
81a6c781 BM |
8162 | *) Added several new manual pages for SSL_* function. |
8163 | [Lutz Jaenicke] | |
8164 | ||
0e8f2fdf | 8165 | Changes between 0.9.5a and 0.9.6 [24 Sep 2000] |
bbb72003 | 8166 | |
f1192b7f BM |
8167 | *) In ssl23_get_client_hello, generate an error message when faced |
8168 | with an initial SSL 3.0/TLS record that is too small to contain the | |
8169 | first two bytes of the ClientHello message, i.e. client_version. | |
8170 | (Note that this is a pathologic case that probably has never happened | |
8171 | in real life.) The previous approach was to use the version number | |
5a5accdd | 8172 | from the record header as a substitute; but our protocol choice |
f1192b7f BM |
8173 | should not depend on that one because it is not authenticated |
8174 | by the Finished messages. | |
8175 | [Bodo Moeller] | |
8176 | ||
d49da3aa UM |
8177 | *) More robust randomness gathering functions for Windows. |
8178 | [Jeffrey Altman <jaltman@columbia.edu>] | |
8179 | ||
dbba890c DSH |
8180 | *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is |
8181 | not set then we don't setup the error code for issuer check errors | |
8182 | to avoid possibly overwriting other errors which the callback does | |
8183 | handle. If an application does set the flag then we assume it knows | |
8184 | what it is doing and can handle the new informational codes | |
8185 | appropriately. | |
8186 | [Steve Henson] | |
8187 | ||
6cffb201 DSH |
8188 | *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for |
8189 | a general "ANY" type, as such it should be able to decode anything | |
8190 | including tagged types. However it didn't check the class so it would | |
8191 | wrongly interpret tagged types in the same way as their universal | |
8192 | counterpart and unknown types were just rejected. Changed so that the | |
8193 | tagged and unknown types are handled in the same way as a SEQUENCE: | |
8194 | that is the encoding is stored intact. There is also a new type | |
8195 | "V_ASN1_OTHER" which is used when the class is not universal, in this | |
8196 | case we have no idea what the actual type is so we just lump them all | |
8197 | together. | |
8198 | [Steve Henson] | |
8199 | ||
645749ef RL |
8200 | *) On VMS, stdout may very well lead to a file that is written to |
8201 | in a record-oriented fashion. That means that every write() will | |
8202 | write a separate record, which will be read separately by the | |
8203 | programs trying to read from it. This can be very confusing. | |
8204 | ||
8205 | The solution is to put a BIO filter in the way that will buffer | |
8206 | text until a linefeed is reached, and then write everything a | |
8207 | line at a time, so every record written will be an actual line, | |
8208 | not chunks of lines and not (usually doesn't happen, but I've | |
8209 | seen it once) several lines in one record. BIO_f_linebuffer() is | |
8210 | the answer. | |
8211 | ||
8212 | Currently, it's a VMS-only method, because that's where it has | |
8213 | been tested well enough. | |
8214 | [Richard Levitte] | |
8215 | ||
fe035197 | 8216 | *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery, |
a45bd295 | 8217 | it can return incorrect results. |
cb1fbf8e BM |
8218 | (Note: The buggy variant was not enabled in OpenSSL 0.9.5a, |
8219 | but it was in 0.9.6-beta[12].) | |
a45bd295 BM |
8220 | [Bodo Moeller] |
8221 | ||
730e37ed DSH |
8222 | *) Disable the check for content being present when verifying detached |
8223 | signatures in pk7_smime.c. Some versions of Netscape (wrongly) | |
8224 | include zero length content when signing messages. | |
8225 | [Steve Henson] | |
8226 | ||
07fcf422 BM |
8227 | *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR |
8228 | BIO_ctrl (for BIO pairs). | |
9f0b86c6 | 8229 | [Bodo Möller] |
07fcf422 | 8230 | |
0e05f545 RL |
8231 | *) Add DSO method for VMS. |
8232 | [Richard Levitte] | |
8233 | ||
1d84fd64 UM |
8234 | *) Bug fix: Montgomery multiplication could produce results with the |
8235 | wrong sign. | |
9f0b86c6 | 8236 | [Ulf Möller] |
1d84fd64 | 8237 | |
775bcebd RL |
8238 | *) Add RPM specification openssl.spec and modify it to build three |
8239 | packages. The default package contains applications, application | |
8240 | documentation and run-time libraries. The devel package contains | |
8241 | include files, static libraries and function documentation. The | |
8242 | doc package contains the contents of the doc directory. The original | |
8243 | openssl.spec was provided by Damien Miller <djm@mindrot.org>. | |
8244 | [Richard Levitte] | |
8245 | ||
cc99526d RL |
8246 | *) Add a large number of documentation files for many SSL routines. |
8247 | [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] | |
8248 | ||
72660f5f RL |
8249 | *) Add a configuration entry for Sony News 4. |
8250 | [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>] | |
8251 | ||
5401c4c2 UM |
8252 | *) Don't set the two most significant bits to one when generating a |
8253 | random number < q in the DSA library. | |
9f0b86c6 | 8254 | [Ulf Möller] |
5401c4c2 | 8255 | |
54f10e6a BM |
8256 | *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default |
8257 | behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if | |
8258 | the underlying transport is blocking) if a handshake took place. | |
8259 | (The default behaviour is needed by applications such as s_client | |
8260 | and s_server that use select() to determine when to use SSL_read; | |
8261 | but for applications that know in advance when to expect data, it | |
8262 | just makes things more complicated.) | |
8263 | [Bodo Moeller] | |
8264 | ||
2959f292 BL |
8265 | *) Add RAND_egd_bytes(), which gives control over the number of bytes read |
8266 | from EGD. | |
8267 | [Ben Laurie] | |
8268 | ||
97d8e82c RL |
8269 | *) Add a few more EBCDIC conditionals that make `req' and `x509' |
8270 | work better on such systems. | |
8271 | [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] | |
8272 | ||
84b65340 DSH |
8273 | *) Add two demo programs for PKCS12_parse() and PKCS12_create(). |
8274 | Update PKCS12_parse() so it copies the friendlyName and the | |
8275 | keyid to the certificates aux info. | |
8276 | [Steve Henson] | |
8277 | ||
f50c11ca DSH |
8278 | *) Fix bug in PKCS7_verify() which caused an infinite loop |
8279 | if there was more than one signature. | |
8280 | [Sven Uszpelkat <su@celocom.de>] | |
8281 | ||
948d0125 RL |
8282 | *) Major change in util/mkdef.pl to include extra information |
8283 | about each symbol, as well as presentig variables as well | |
8284 | as functions. This change means that there's n more need | |
8285 | to rebuild the .num files when some algorithms are excluded. | |
8286 | [Richard Levitte] | |
8287 | ||
bbb72003 DSH |
8288 | *) Allow the verify time to be set by an application, |
8289 | rather than always using the current time. | |
8290 | [Steve Henson] | |
2f043896 | 8291 | |
bbb72003 DSH |
8292 | *) Phase 2 verify code reorganisation. The certificate |
8293 | verify code now looks up an issuer certificate by a | |
8294 | number of criteria: subject name, authority key id | |
8295 | and key usage. It also verifies self signed certificates | |
8296 | by the same criteria. The main comparison function is | |
8297 | X509_check_issued() which performs these checks. | |
2f043896 | 8298 | |
bbb72003 DSH |
8299 | Lot of changes were necessary in order to support this |
8300 | without completely rewriting the lookup code. | |
2f043896 | 8301 | |
bbb72003 | 8302 | Authority and subject key identifier are now cached. |
2f043896 | 8303 | |
bbb72003 DSH |
8304 | The LHASH 'certs' is X509_STORE has now been replaced |
8305 | by a STACK_OF(X509_OBJECT). This is mainly because an | |
8306 | LHASH can't store or retrieve multiple objects with | |
8307 | the same hash value. | |
c90341a1 | 8308 | |
bbb72003 DSH |
8309 | As a result various functions (which were all internal |
8310 | use only) have changed to handle the new X509_STORE | |
8311 | structure. This will break anything that messed round | |
8312 | with X509_STORE internally. | |
2f043896 | 8313 | |
bbb72003 DSH |
8314 | The functions X509_STORE_add_cert() now checks for an |
8315 | exact match, rather than just subject name. | |
2f043896 | 8316 | |
bbb72003 DSH |
8317 | The X509_STORE API doesn't directly support the retrieval |
8318 | of multiple certificates matching a given criteria, however | |
8319 | this can be worked round by performing a lookup first | |
8320 | (which will fill the cache with candidate certificates) | |
8321 | and then examining the cache for matches. This is probably | |
8322 | the best we can do without throwing out X509_LOOKUP | |
8323 | entirely (maybe later...). | |
2f043896 | 8324 | |
bbb72003 | 8325 | The X509_VERIFY_CTX structure has been enhanced considerably. |
2f043896 | 8326 | |
bbb72003 DSH |
8327 | All certificate lookup operations now go via a get_issuer() |
8328 | callback. Although this currently uses an X509_STORE it | |
8329 | can be replaced by custom lookups. This is a simple way | |
8330 | to bypass the X509_STORE hackery necessary to make this | |
8331 | work and makes it possible to use more efficient techniques | |
8332 | in future. A very simple version which uses a simple | |
8333 | STACK for its trusted certificate store is also provided | |
8334 | using X509_STORE_CTX_trusted_stack(). | |
2f043896 | 8335 | |
bbb72003 DSH |
8336 | The verify_cb() and verify() callbacks now have equivalents |
8337 | in the X509_STORE_CTX structure. | |
2f043896 | 8338 | |
bbb72003 DSH |
8339 | X509_STORE_CTX also has a 'flags' field which can be used |
8340 | to customise the verify behaviour. | |
8341 | [Steve Henson] | |
2f043896 | 8342 | |
34216c04 DSH |
8343 | *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which |
8344 | excludes S/MIME capabilities. | |
8345 | [Steve Henson] | |
8346 | ||
8347 | *) When a certificate request is read in keep a copy of the | |
8348 | original encoding of the signed data and use it when outputing | |
8349 | again. Signatures then use the original encoding rather than | |
8350 | a decoded, encoded version which may cause problems if the | |
8351 | request is improperly encoded. | |
8352 | [Steve Henson] | |
8353 | ||
affadbef BM |
8354 | *) For consistency with other BIO_puts implementations, call |
8355 | buffer_write(b, ...) directly in buffer_puts instead of calling | |
8356 | BIO_write(b, ...). | |
22c7ea40 BM |
8357 | |
8358 | In BIO_puts, increment b->num_write as in BIO_write. | |
affadbef BM |
8359 | [Peter.Sylvester@EdelWeb.fr] |
8360 | ||
bbb8de09 BM |
8361 | *) Fix BN_mul_word for the case where the word is 0. (We have to use |
8362 | BN_zero, we may not return a BIGNUM with an array consisting of | |
8363 | words set to zero.) | |
8364 | [Bodo Moeller] | |
8365 | ||
8366 | *) Avoid calling abort() from within the library when problems are | |
8367 | detected, except if preprocessor symbols have been defined | |
8368 | (such as REF_CHECK, BN_DEBUG etc.). | |
8369 | [Bodo Moeller] | |
8370 | ||
bd08a2bd DSH |
8371 | *) New openssl application 'rsautl'. This utility can be |
8372 | used for low level RSA operations. DER public key | |
8373 | BIO/fp routines also added. | |
8374 | [Steve Henson] | |
8375 | ||
a545c6f6 BM |
8376 | *) New Configure entry and patches for compiling on QNX 4. |
8377 | [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>] | |
8378 | ||
7049ef5f BL |
8379 | *) A demo state-machine implementation was sponsored by |
8380 | Nuron (http://www.nuron.com/) and is now available in | |
8381 | demos/state_machine. | |
8382 | [Ben Laurie] | |
8383 | ||
7df1c720 DSH |
8384 | *) New options added to the 'dgst' utility for signature |
8385 | generation and verification. | |
8386 | [Steve Henson] | |
8387 | ||
d096b524 DSH |
8388 | *) Unrecognized PKCS#7 content types are now handled via a |
8389 | catch all ASN1_TYPE structure. This allows unsupported | |
8390 | types to be stored as a "blob" and an application can | |
8391 | encode and decode it manually. | |
8392 | [Steve Henson] | |
8393 | ||
7df1c720 | 8394 | *) Fix various signed/unsigned issues to make a_strex.c |
469938cb DSH |
8395 | compile under VC++. |
8396 | [Oscar Jacobsson <oscar.jacobsson@celocom.com>] | |
8397 | ||
8398 | *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct | |
8399 | length if passed a buffer. ASN1_INTEGER_to_BN failed | |
8400 | if passed a NULL BN and its argument was negative. | |
8401 | [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>] | |
8402 | ||
eaa28181 DSH |
8403 | *) Modification to PKCS#7 encoding routines to output definite |
8404 | length encoding. Since currently the whole structures are in | |
8405 | memory there's not real point in using indefinite length | |
8406 | constructed encoding. However if OpenSSL is compiled with | |
8407 | the flag PKCS7_INDEFINITE_ENCODING the old form is used. | |
8408 | [Steve Henson] | |
8409 | ||
e6629837 RL |
8410 | *) Added BIO_vprintf() and BIO_vsnprintf(). |
8411 | [Richard Levitte] | |
8412 | ||
6fd5a047 RL |
8413 | *) Added more prefixes to parse for in the the strings written |
8414 | through a logging bio, to cover all the levels that are available | |
8415 | through syslog. The prefixes are now: | |
8416 | ||
8417 | PANIC, EMERG, EMR => LOG_EMERG | |
8418 | ALERT, ALR => LOG_ALERT | |
8419 | CRIT, CRI => LOG_CRIT | |
8420 | ERROR, ERR => LOG_ERR | |
8421 | WARNING, WARN, WAR => LOG_WARNING | |
8422 | NOTICE, NOTE, NOT => LOG_NOTICE | |
8423 | INFO, INF => LOG_INFO | |
8424 | DEBUG, DBG => LOG_DEBUG | |
8425 | ||
8426 | and as before, if none of those prefixes are present at the | |
8427 | beginning of the string, LOG_ERR is chosen. | |
8428 | ||
8429 | On Win32, the LOG_* levels are mapped according to this: | |
8430 | ||
8431 | LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE | |
8432 | LOG_WARNING => EVENTLOG_WARNING_TYPE | |
8433 | LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE | |
8434 | ||
8435 | [Richard Levitte] | |
8436 | ||
368f8554 RL |
8437 | *) Made it possible to reconfigure with just the configuration |
8438 | argument "reconf" or "reconfigure". The command line arguments | |
8439 | are stored in Makefile.ssl in the variable CONFIGURE_ARGS, | |
8440 | and are retrieved from there when reconfiguring. | |
8441 | [Richard Levitte] | |
8442 | ||
3009458e | 8443 | *) MD4 implemented. |
bb531a0a | 8444 | [Assar Westerlund <assar@sics.se>, Richard Levitte] |
3009458e | 8445 | |
88364bc2 RL |
8446 | *) Add the arguments -CAfile and -CApath to the pkcs12 utility. |
8447 | [Richard Levitte] | |
8448 | ||
d4fbe318 DSH |
8449 | *) The obj_dat.pl script was messing up the sorting of object |
8450 | names. The reason was that it compared the quoted version | |
8451 | of strings as a result "OCSP" > "OCSP Signing" because | |
8452 | " > SPACE. Changed script to store unquoted versions of | |
8453 | names and add quotes on output. It was also omitting some | |
8454 | names from the lookup table if they were given a default | |
8455 | value (that is if SN is missing it is given the same | |
8456 | value as LN and vice versa), these are now added on the | |
8457 | grounds that if an object has a name we should be able to | |
8458 | look it up. Finally added warning output when duplicate | |
8459 | short or long names are found. | |
8460 | [Steve Henson] | |
8461 | ||
2d978cbd | 8462 | *) Changes needed for Tandem NSK. |
d49da3aa | 8463 | [Scott Uroff <scott@xypro.com>] |
2d978cbd | 8464 | |
aa826d88 BM |
8465 | *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in |
8466 | RSA_padding_check_SSLv23(), special padding was never detected | |
8467 | and thus the SSL 3.0/TLS 1.0 countermeasure against protocol | |
8468 | version rollback attacks was not effective. | |
8469 | ||
37569e64 BM |
8470 | In s23_clnt.c, don't use special rollback-attack detection padding |
8471 | (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the | |
8472 | client; similarly, in s23_srvr.c, don't do the rollback check if | |
8473 | SSL 2.0 is the only protocol enabled in the server. | |
8474 | [Bodo Moeller] | |
8475 | ||
ca1e465f RL |
8476 | *) Make it possible to get hexdumps of unprintable data with 'openssl |
8477 | asn1parse'. By implication, the functions ASN1_parse_dump() and | |
8478 | BIO_dump_indent() are added. | |
8479 | [Richard Levitte] | |
8480 | ||
a657546f DSH |
8481 | *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex() |
8482 | these print out strings and name structures based on various | |
8483 | flags including RFC2253 support and proper handling of | |
8484 | multibyte characters. Added options to the 'x509' utility | |
8485 | to allow the various flags to be set. | |
8486 | [Steve Henson] | |
8487 | ||
284ef5f3 DSH |
8488 | *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME. |
8489 | Also change the functions X509_cmp_current_time() and | |
8490 | X509_gmtime_adj() work with an ASN1_TIME structure, | |
8491 | this will enable certificates using GeneralizedTime in validity | |
8492 | dates to be checked. | |
8493 | [Steve Henson] | |
8494 | ||
8495 | *) Make the NEG_PUBKEY_BUG code (which tolerates invalid | |
8496 | negative public key encodings) on by default, | |
8497 | NO_NEG_PUBKEY_BUG can be set to disable it. | |
8498 | [Steve Henson] | |
8499 | ||
8500 | *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT | |
8501 | content octets. An i2c_ASN1_OBJECT is unnecessary because | |
8502 | the encoding can be trivially obtained from the structure. | |
8503 | [Steve Henson] | |
8504 | ||
fa729135 BM |
8505 | *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock), |
8506 | not read locks (CRYPTO_r_[un]lock). | |
8507 | [Bodo Moeller] | |
8508 | ||
b436a982 RL |
8509 | *) A first attempt at creating official support for shared |
8510 | libraries through configuration. I've kept it so the | |
8511 | default is static libraries only, and the OpenSSL programs | |
8512 | are always statically linked for now, but there are | |
8513 | preparations for dynamic linking in place. | |
6bc847e4 | 8514 | This has been tested on Linux and Tru64. |
b436a982 RL |
8515 | [Richard Levitte] |
8516 | ||
c0722725 UM |
8517 | *) Randomness polling function for Win9x, as described in: |
8518 | Peter Gutmann, Software Generation of Practically Strong | |
8519 | Random Numbers. | |
9f0b86c6 | 8520 | [Ulf Möller] |
c0722725 | 8521 | |
fd13f0ee DSH |
8522 | *) Fix so PRNG is seeded in req if using an already existing |
8523 | DSA key. | |
8524 | [Steve Henson] | |
8525 | ||
094fe66d DSH |
8526 | *) New options to smime application. -inform and -outform |
8527 | allow alternative formats for the S/MIME message including | |
8528 | PEM and DER. The -content option allows the content to be | |
8529 | specified separately. This should allow things like Netscape | |
8530 | form signing output easier to verify. | |
8531 | [Steve Henson] | |
8532 | ||
8533 | *) Fix the ASN1 encoding of tags using the 'long form'. | |
8534 | [Steve Henson] | |
8535 | ||
a338e21b DSH |
8536 | *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT |
8537 | STRING types. These convert content octets to and from the | |
8538 | underlying type. The actual tag and length octets are | |
8539 | already assumed to have been read in and checked. These | |
8540 | are needed because all other string types have virtually | |
8541 | identical handling apart from the tag. By having versions | |
8542 | of the ASN1 functions that just operate on content octets | |
8543 | IMPLICIT tagging can be handled properly. It also allows | |
8544 | the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED | |
8545 | and ASN1_INTEGER are identical apart from the tag. | |
8546 | [Steve Henson] | |
8547 | ||
d5870bbe RL |
8548 | *) Change the handling of OID objects as follows: |
8549 | ||
8550 | - New object identifiers are inserted in objects.txt, following | |
8551 | the syntax given in objects.README. | |
8552 | - objects.pl is used to process obj_mac.num and create a new | |
8553 | obj_mac.h. | |
8554 | - obj_dat.pl is used to create a new obj_dat.h, using the data in | |
8555 | obj_mac.h. | |
8556 | ||
8557 | This is currently kind of a hack, and the perl code in objects.pl | |
8558 | isn't very elegant, but it works as I intended. The simplest way | |
8559 | to check that it worked correctly is to look in obj_dat.h and | |
8560 | check the array nid_objs and make sure the objects haven't moved | |
8561 | around (this is important!). Additions are OK, as well as | |
8562 | consistent name changes. | |
8563 | [Richard Levitte] | |
8564 | ||
1f4643a2 BM |
8565 | *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1'). |
8566 | [Bodo Moeller] | |
8567 | ||
fb0b844a | 8568 | *) Addition of the command line parameter '-rand file' to 'openssl req'. |
1023b122 RL |
8569 | The given file adds to whatever has already been seeded into the |
8570 | random pool through the RANDFILE configuration file option or | |
8571 | environment variable, or the default random state file. | |
fb0b844a RL |
8572 | [Richard Levitte] |
8573 | ||
4dd45354 DSH |
8574 | *) mkstack.pl now sorts each macro group into lexical order. |
8575 | Previously the output order depended on the order the files | |
8576 | appeared in the directory, resulting in needless rewriting | |
8577 | of safestack.h . | |
8578 | [Steve Henson] | |
8579 | ||
13083215 DSH |
8580 | *) Patches to make OpenSSL compile under Win32 again. Mostly |
8581 | work arounds for the VC++ problem that it treats func() as | |
8582 | func(void). Also stripped out the parts of mkdef.pl that | |
8583 | added extra typesafe functions: these no longer exist. | |
8584 | [Steve Henson] | |
8585 | ||
3aceb94b DSH |
8586 | *) Reorganisation of the stack code. The macros are now all |
8587 | collected in safestack.h . Each macro is defined in terms of | |
8588 | a "stack macro" of the form SKM_<name>(type, a, b). The | |
8589 | DEBUG_SAFESTACK is now handled in terms of function casts, | |
8590 | this has the advantage of retaining type safety without the | |
8591 | use of additional functions. If DEBUG_SAFESTACK is not defined | |
8592 | then the non typesafe macros are used instead. Also modified the | |
8593 | mkstack.pl script to handle the new form. Needs testing to see | |
8594 | if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK | |
7ef82068 DSH |
8595 | the default if no major problems. Similar behaviour for ASN1_SET_OF |
8596 | and PKCS12_STACK_OF. | |
3aceb94b DSH |
8597 | [Steve Henson] |
8598 | ||
d3ed8ceb DSH |
8599 | *) When some versions of IIS use the 'NET' form of private key the |
8600 | key derivation algorithm is different. Normally MD5(password) is | |
8601 | used as a 128 bit RC4 key. In the modified case | |
8602 | MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some | |
8603 | new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same | |
8604 | as the old Netscape_RSA functions except they have an additional | |
8605 | 'sgckey' parameter which uses the modified algorithm. Also added | |
8606 | an -sgckey command line option to the rsa utility. Thanks to | |
8607 | Adrian Peck <bertie@ncipher.com> for posting details of the modified | |
8608 | algorithm to openssl-dev. | |
8609 | [Steve Henson] | |
8610 | ||
e366f2b8 DSH |
8611 | *) The evp_local.h macros were using 'c.##kname' which resulted in |
8612 | invalid expansion on some systems (SCO 5.0.5 for example). | |
8613 | Corrected to 'c.kname'. | |
8614 | [Phillip Porch <root@theporch.com>] | |
8615 | ||
a91dedca DSH |
8616 | *) New X509_get1_email() and X509_REQ_get1_email() functions that return |
8617 | a STACK of email addresses from a certificate or request, these look | |
8618 | in the subject name and the subject alternative name extensions and | |
8619 | omit any duplicate addresses. | |
8620 | [Steve Henson] | |
8621 | ||
dc434bbc BM |
8622 | *) Re-implement BN_mod_exp2_mont using independent (and larger) windows. |
8623 | This makes DSA verification about 2 % faster. | |
8624 | [Bodo Moeller] | |
8625 | ||
8626 | *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5 | |
8627 | (meaning that now 2^5 values will be precomputed, which is only 4 KB | |
8628 | plus overhead for 1024 bit moduli). | |
8629 | This makes exponentiations about 0.5 % faster for 1024 bit | |
8630 | exponents (as measured by "openssl speed rsa2048"). | |
8631 | [Bodo Moeller] | |
8632 | ||
947b3b8b BM |
8633 | *) Rename memory handling macros to avoid conflicts with other |
8634 | software: | |
8635 | Malloc => OPENSSL_malloc | |
8636 | Malloc_locked => OPENSSL_malloc_locked | |
8637 | Realloc => OPENSSL_realloc | |
8638 | Free => OPENSSL_free | |
8639 | [Richard Levitte] | |
8640 | ||
482a9d41 BM |
8641 | *) New function BN_mod_exp_mont_word for small bases (roughly 15% |
8642 | faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange). | |
6dad7bd6 BM |
8643 | [Bodo Moeller] |
8644 | ||
be5d92e0 UM |
8645 | *) CygWin32 support. |
8646 | [John Jarvie <jjarvie@newsguy.com>] | |
8647 | ||
e41c8d6a GT |
8648 | *) The type-safe stack code has been rejigged. It is now only compiled |
8649 | in when OpenSSL is configured with the DEBUG_SAFESTACK option and | |
8650 | by default all type-specific stack functions are "#define"d back to | |
8651 | standard stack functions. This results in more streamlined output | |
8652 | but retains the type-safety checking possibilities of the original | |
8653 | approach. | |
8654 | [Geoff Thorpe] | |
8655 | ||
ccd86b68 GT |
8656 | *) The STACK code has been cleaned up, and certain type declarations |
8657 | that didn't make a lot of sense have been brought in line. This has | |
8658 | also involved a cleanup of sorts in safestack.h to more correctly | |
8659 | map type-safe stack functions onto their plain stack counterparts. | |
8660 | This work has also resulted in a variety of "const"ifications of | |
8661 | lots of the code, especially "_cmp" operations which should normally | |
8662 | be prototyped with "const" parameters anyway. | |
8663 | [Geoff Thorpe] | |
8664 | ||
361ee973 BM |
8665 | *) When generating bytes for the first time in md_rand.c, 'stir the pool' |
8666 | by seeding with STATE_SIZE dummy bytes (with zero entropy count). | |
8667 | (The PRNG state consists of two parts, the large pool 'state' and 'md', | |
8668 | where all of 'md' is used each time the PRNG is used, but 'state' | |
8669 | is used only indexed by a cyclic counter. As entropy may not be | |
8670 | well distributed from the beginning, 'md' is important as a | |
8671 | chaining variable. However, the output function chains only half | |
8672 | of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains | |
8673 | all of 'md', and seeding with STATE_SIZE dummy bytes will result | |
8674 | in all of 'state' being rewritten, with the new values depending | |
8675 | on virtually all of 'md'. This overcomes the 80 bit limitation.) | |
8676 | [Bodo Moeller] | |
8677 | ||
49528751 DSH |
8678 | *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when |
8679 | the handshake is continued after ssl_verify_cert_chain(); | |
8680 | otherwise, if SSL_VERIFY_NONE is set, remaining error codes | |
8681 | can lead to 'unexplainable' connection aborts later. | |
8682 | [Bodo Moeller; problem tracked down by Lutz Jaenicke] | |
8683 | ||
8684 | *) Major EVP API cipher revision. | |
8685 | Add hooks for extra EVP features. This allows various cipher | |
8686 | parameters to be set in the EVP interface. Support added for variable | |
8687 | key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and | |
8688 | setting of RC2 and RC5 parameters. | |
8689 | ||
8690 | Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length | |
8691 | ciphers. | |
8692 | ||
8693 | Remove lots of duplicated code from the EVP library. For example *every* | |
360370d9 DSH |
8694 | cipher init() function handles the 'iv' in the same way according to the |
8695 | cipher mode. They also all do nothing if the 'key' parameter is NULL and | |
8696 | for CFB and OFB modes they zero ctx->num. | |
8697 | ||
49528751 DSH |
8698 | New functionality allows removal of S/MIME code RC2 hack. |
8699 | ||
57ae2e24 DSH |
8700 | Most of the routines have the same form and so can be declared in terms |
8701 | of macros. | |
8702 | ||
360370d9 DSH |
8703 | By shifting this to the top level EVP_CipherInit() it can be removed from |
8704 | all individual ciphers. If the cipher wants to handle IVs or keys | |
8705 | differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT | |
8706 | flags. | |
be06a934 DSH |
8707 | |
8708 | Change lots of functions like EVP_EncryptUpdate() to now return a | |
8709 | value: although software versions of the algorithms cannot fail | |
8710 | any installed hardware versions can. | |
7f060601 DSH |
8711 | [Steve Henson] |
8712 | ||
2c05c494 BM |
8713 | *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if |
8714 | this option is set, tolerate broken clients that send the negotiated | |
8715 | protocol version number instead of the requested protocol version | |
8716 | number. | |
8717 | [Bodo Moeller] | |
8718 | ||
8719 | *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag; | |
8720 | i.e. non-zero for export ciphersuites, zero otherwise. | |
8721 | Previous versions had this flag inverted, inconsistent with | |
8722 | rsa_tmp_cb (..._TMP_RSA_CB). | |
8723 | [Bodo Moeller; problem reported by Amit Chopra] | |
8724 | ||
b4b41f48 DSH |
8725 | *) Add missing DSA library text string. Work around for some IIS |
8726 | key files with invalid SEQUENCE encoding. | |
8727 | [Steve Henson] | |
8728 | ||
6d7cce48 RL |
8729 | *) Add a document (doc/standards.txt) that list all kinds of standards |
8730 | and so on that are implemented in OpenSSL. | |
8731 | [Richard Levitte] | |
8732 | ||
439df508 DSH |
8733 | *) Enhance c_rehash script. Old version would mishandle certificates |
8734 | with the same subject name hash and wouldn't handle CRLs at all. | |
8735 | Added -fingerprint option to crl utility, to support new c_rehash | |
8736 | features. | |
8737 | [Steve Henson] | |
8738 | ||
0e1c0612 | 8739 | *) Eliminate non-ANSI declarations in crypto.h and stack.h. |
9f0b86c6 | 8740 | [Ulf Möller] |
0e1c0612 | 8741 | |
0cb957a6 DSH |
8742 | *) Fix for SSL server purpose checking. Server checking was |
8743 | rejecting certificates which had extended key usage present | |
8744 | but no ssl client purpose. | |
8745 | [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>] | |
8746 | ||
a331a305 DSH |
8747 | *) Make PKCS#12 code work with no password. The PKCS#12 spec |
8748 | is a little unclear about how a blank password is handled. | |
8749 | Since the password in encoded as a BMPString with terminating | |
8750 | double NULL a zero length password would end up as just the | |
8751 | double NULL. However no password at all is different and is | |
8752 | handled differently in the PKCS#12 key generation code. NS | |
8753 | treats a blank password as zero length. MSIE treats it as no | |
8754 | password on export: but it will try both on import. We now do | |
8755 | the same: PKCS12_parse() tries zero length and no password if | |
8756 | the password is set to "" or NULL (NULL is now a valid password: | |
8757 | it wasn't before) as does the pkcs12 application. | |
8758 | [Steve Henson] | |
8759 | ||
316e6a66 BM |
8760 | *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use |
8761 | perror when PEM_read_bio_X509_REQ fails, the error message must | |
8762 | be obtained from the error queue. | |
8763 | [Bodo Moeller] | |
8764 | ||
dcba2534 BM |
8765 | *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing |
8766 | it in ERR_remove_state if appropriate, and change ERR_get_state | |
8767 | accordingly to avoid race conditions (this is necessary because | |
8768 | thread_hash is no longer constant once set). | |
8769 | [Bodo Moeller] | |
8770 | ||
3973628e | 8771 | *) Bugfix for linux-elf makefile.one. |
9f0b86c6 | 8772 | [Ulf Möller] |
3973628e | 8773 | |
deb4d50e GT |
8774 | *) RSA_get_default_method() will now cause a default |
8775 | RSA_METHOD to be chosen if one doesn't exist already. | |
8776 | Previously this was only set during a call to RSA_new() | |
8777 | or RSA_new_method(NULL) meaning it was possible for | |
8778 | RSA_get_default_method() to return NULL. | |
8779 | [Geoff Thorpe] | |
8780 | ||
b9e63915 GT |
8781 | *) Added native name translation to the existing DSO code |
8782 | that will convert (if the flag to do so is set) filenames | |
8783 | that are sufficiently small and have no path information | |
8784 | into a canonical native form. Eg. "blah" converted to | |
8785 | "libblah.so" or "blah.dll" etc. | |
8786 | [Geoff Thorpe] | |
8787 | ||
e5c84d51 BM |
8788 | *) New function ERR_error_string_n(e, buf, len) which is like |
8789 | ERR_error_string(e, buf), but writes at most 'len' bytes | |
8790 | including the 0 terminator. For ERR_error_string_n, 'buf' | |
8791 | may not be NULL. | |
8792 | [Damien Miller <djm@mindrot.org>, Bodo Moeller] | |
8793 | ||
a9831305 RL |
8794 | *) CONF library reworked to become more general. A new CONF |
8795 | configuration file reader "class" is implemented as well as a | |
8796 | new functions (NCONF_*, for "New CONF") to handle it. The now | |
8797 | old CONF_* functions are still there, but are reimplemented to | |
8798 | work in terms of the new functions. Also, a set of functions | |
8799 | to handle the internal storage of the configuration data is | |
8800 | provided to make it easier to write new configuration file | |
8801 | reader "classes" (I can definitely see something reading a | |
8802 | configuration file in XML format, for example), called _CONF_*, | |
8803 | or "the configuration storage API"... | |
8804 | ||
8805 | The new configuration file reading functions are: | |
8806 | ||
2c05c494 BM |
8807 | NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio, |
8808 | NCONF_get_section, NCONF_get_string, NCONF_get_numbre | |
a9831305 | 8809 | |
2c05c494 | 8810 | NCONF_default, NCONF_WIN32 |
a9831305 | 8811 | |
2c05c494 | 8812 | NCONF_dump_fp, NCONF_dump_bio |
a9831305 RL |
8813 | |
8814 | NCONF_default and NCONF_WIN32 are method (or "class") choosers, | |
8815 | NCONF_new creates a new CONF object. This works in the same way | |
8816 | as other interfaces in OpenSSL, like the BIO interface. | |
8817 | NCONF_dump_* dump the internal storage of the configuration file, | |
8818 | which is useful for debugging. All other functions take the same | |
8819 | arguments as the old CONF_* functions wth the exception of the | |
8820 | first that must be a `CONF *' instead of a `LHASH *'. | |
8821 | ||
8822 | To make it easer to use the new classes with the old CONF_* functions, | |
8823 | the function CONF_set_default_method is provided. | |
8824 | [Richard Levitte] | |
8825 | ||
1d90f280 BM |
8826 | *) Add '-tls1' option to 'openssl ciphers', which was already |
8827 | mentioned in the documentation but had not been implemented. | |
8828 | (This option is not yet really useful because even the additional | |
8829 | experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.) | |
8830 | [Bodo Moeller] | |
8831 | ||
6ef4d9d5 GT |
8832 | *) Initial DSO code added into libcrypto for letting OpenSSL (and |
8833 | OpenSSL-based applications) load shared libraries and bind to | |
8834 | them in a portable way. | |
8835 | [Geoff Thorpe, with contributions from Richard Levitte] | |
c90341a1 | 8836 | |
5e61580b RL |
8837 | Changes between 0.9.5 and 0.9.5a [1 Apr 2000] |
8838 | ||
8839 | *) Make sure _lrotl and _lrotr are only used with MSVC. | |
865874f2 | 8840 | |
cf194c1f BM |
8841 | *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status |
8842 | (the default implementation of RAND_status). | |
8843 | ||
3bc90f23 BM |
8844 | *) Rename openssl x509 option '-crlext', which was added in 0.9.5, |
8845 | to '-clrext' (= clear extensions), as intended and documented. | |
8846 | [Bodo Moeller; inconsistency pointed out by Michael Attili | |
8847 | <attili@amaxo.com>] | |
8848 | ||
b475baff DSH |
8849 | *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length |
8850 | was larger than the MD block size. | |
8851 | [Steve Henson, pointed out by Yost William <YostW@tce.com>] | |
8852 | ||
e77066ea DSH |
8853 | *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument |
8854 | fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set() | |
8855 | using the passed key: if the passed key was a private key the result | |
8856 | of X509_print(), for example, would be to print out all the private key | |
8857 | components. | |
8858 | [Steve Henson] | |
8859 | ||
7af4816f | 8860 | *) des_quad_cksum() byte order bug fix. |
9f0b86c6 | 8861 | [Ulf Möller, using the problem description in krb4-0.9.7, where |
7af4816f UM |
8862 | the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>] |
8863 | ||
80870566 DSH |
8864 | *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly |
8865 | discouraged. | |
8866 | [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>] | |
8867 | ||
7694ddcb BM |
8868 | *) For easily testing in shell scripts whether some command |
8869 | 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX' | |
df1ff3f1 | 8870 | returns with exit code 0 iff no command of the given name is available. |
7694ddcb | 8871 | 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases, |
46c4647e BM |
8872 | the output goes to stdout and nothing is printed to stderr. |
8873 | Additional arguments are always ignored. | |
8874 | ||
8875 | Since for each cipher there is a command of the same name, | |
8876 | the 'no-cipher' compilation switches can be tested this way. | |
7694ddcb BM |
8877 | |
8878 | ('openssl no-XXX' is not able to detect pseudo-commands such | |
8879 | as 'quit', 'list-XXX-commands', or 'no-XXX' itself.) | |
46c4647e BM |
8880 | [Bodo Moeller] |
8881 | ||
65b002f3 BM |
8882 | *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration. |
8883 | [Bodo Moeller] | |
8884 | ||
e11f0de6 BM |
8885 | *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE |
8886 | is set; it will be thrown away anyway because each handshake creates | |
8887 | its own key. | |
8888 | ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition | |
8889 | to parameters -- in previous versions (since OpenSSL 0.9.3) the | |
8890 | 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining | |
8891 | you effectivly got SSL_OP_SINGLE_DH_USE when using this macro. | |
8892 | [Bodo Moeller] | |
8893 | ||
2d5e449a BM |
8894 | *) New s_client option -ign_eof: EOF at stdin is ignored, and |
8895 | 'Q' and 'R' lose their special meanings (quit/renegotiate). | |
8896 | This is part of what -quiet does; unlike -quiet, -ign_eof | |
8897 | does not suppress any output. | |
8898 | [Richard Levitte] | |
8899 | ||
daf4e53e | 8900 | *) Add compatibility options to the purpose and trust code. The |
068fdce8 DSH |
8901 | purpose X509_PURPOSE_ANY is "any purpose" which automatically |
8902 | accepts a certificate or CA, this was the previous behaviour, | |
8903 | with all the associated security issues. | |
8904 | ||
8905 | X509_TRUST_COMPAT is the old trust behaviour: only and | |
8906 | automatically trust self signed roots in certificate store. A | |
8907 | new trust setting X509_TRUST_DEFAULT is used to specify that | |
8908 | a purpose has no associated trust setting and it should instead | |
8909 | use the value in the default purpose. | |
8910 | [Steve Henson] | |
8911 | ||
48fe0eec DSH |
8912 | *) Fix the PKCS#8 DSA private key code so it decodes keys again |
8913 | and fix a memory leak. | |
8914 | [Steve Henson] | |
8915 | ||
59fc2b0f BM |
8916 | *) In util/mkerr.pl (which implements 'make errors'), preserve |
8917 | reason strings from the previous version of the .c file, as | |
8918 | the default to have only downcase letters (and digits) in | |
8919 | automatically generated reasons codes is not always appropriate. | |
8920 | [Bodo Moeller] | |
8921 | ||
0a150c5c BM |
8922 | *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table |
8923 | using strerror. Previously, ERR_reason_error_string() returned | |
8924 | library names as reason strings for SYSerr; but SYSerr is a special | |
8925 | case where small numbers are errno values, not library numbers. | |
8926 | [Bodo Moeller] | |
8927 | ||
41918458 BM |
8928 | *) Add '-dsaparam' option to 'openssl dhparam' application. This |
8929 | converts DSA parameters into DH parameters. (When creating parameters, | |
8930 | DSA_generate_parameters is used.) | |
8931 | [Bodo Moeller] | |
8932 | ||
8933 | *) Include 'length' (recommended exponent length) in C code generated | |
8934 | by 'openssl dhparam -C'. | |
8935 | [Bodo Moeller] | |
8936 | ||
d9c88a39 DSH |
8937 | *) The second argument to set_label in perlasm was already being used |
8938 | so couldn't be used as a "file scope" flag. Moved to third argument | |
8939 | which was free. | |
8940 | [Steve Henson] | |
8941 | ||
84d14408 BM |
8942 | *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes |
8943 | instead of RAND_bytes for encryption IVs and salts. | |
8944 | [Bodo Moeller] | |
8945 | ||
5eb8ca4d BM |
8946 | *) Include RAND_status() into RAND_METHOD instead of implementing |
8947 | it only for md_rand.c Otherwise replacing the PRNG by calling | |
8948 | RAND_set_rand_method would be impossible. | |
8949 | [Bodo Moeller] | |
8950 | ||
7a2dfc2a UM |
8951 | *) Don't let DSA_generate_key() enter an infinite loop if the random |
8952 | number generation fails. | |
8953 | [Bodo Moeller] | |
8954 | ||
55f7d65d BM |
8955 | *) New 'rand' application for creating pseudo-random output. |
8956 | [Bodo Moeller] | |
8957 | ||
010712ff RE |
8958 | *) Added configuration support for Linux/IA64 |
8959 | [Rolf Haberrecker <rolf@suse.de>] | |
8960 | ||
2da0c119 | 8961 | *) Assembler module support for Mingw32. |
9f0b86c6 | 8962 | [Ulf Möller] |
2da0c119 | 8963 | |
a4709b3d UM |
8964 | *) Shared library support for HPUX (in shlib/). |
8965 | [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous] | |
8966 | ||
8967 | *) Shared library support for Solaris gcc. | |
8968 | [Lutz Behnke <behnke@trustcenter.de>] | |
865874f2 | 8969 | |
74cdf6f7 | 8970 | Changes between 0.9.4 and 0.9.5 [28 Feb 2000] |
38e33cef | 8971 | |
82b93186 DSH |
8972 | *) PKCS7_encrypt() was adding text MIME headers twice because they |
8973 | were added manually and by SMIME_crlf_copy(). | |
8974 | [Steve Henson] | |
8975 | ||
587bb0e0 DSH |
8976 | *) In bntest.c don't call BN_rand with zero bits argument. |
8977 | [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>] | |
8978 | ||
688938fb | 8979 | *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n] |
a5770be6 | 8980 | case was implemented. This caused BN_div_recp() to fail occasionally. |
9f0b86c6 | 8981 | [Ulf Möller] |
688938fb | 8982 | |
94de0419 DSH |
8983 | *) Add an optional second argument to the set_label() in the perl |
8984 | assembly language builder. If this argument exists and is set | |
8985 | to 1 it signals that the assembler should use a symbol whose | |
8986 | scope is the entire file, not just the current function. This | |
8987 | is needed with MASM which uses the format label:: for this scope. | |
8988 | [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>] | |
8989 | ||
0202197d DSH |
8990 | *) Change the ASN1 types so they are typedefs by default. Before |
8991 | almost all types were #define'd to ASN1_STRING which was causing | |
8992 | STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING) | |
8993 | for example. | |
8994 | [Steve Henson] | |
8995 | ||
6d0d5431 BM |
8996 | *) Change names of new functions to the new get1/get0 naming |
8997 | convention: After 'get1', the caller owns a reference count | |
8998 | and has to call ..._free; 'get0' returns a pointer to some | |
8999 | data structure without incrementing reference counters. | |
9000 | (Some of the existing 'get' functions increment a reference | |
9001 | counter, some don't.) | |
9002 | Similarly, 'set1' and 'add1' functions increase reference | |
9003 | counters or duplicate objects. | |
c7cb16a8 DSH |
9004 | [Steve Henson] |
9005 | ||
fbb41ae0 DSH |
9006 | *) Allow for the possibility of temp RSA key generation failure: |
9007 | the code used to assume it always worked and crashed on failure. | |
9008 | [Steve Henson] | |
9009 | ||
505b5a0e | 9010 | *) Fix potential buffer overrun problem in BIO_printf(). |
9f0b86c6 | 9011 | [Ulf Möller, using public domain code by Patrick Powell; problem |
505b5a0e UM |
9012 | pointed out by David Sacerdote <das33@cornell.edu>] |
9013 | ||
4ec2d4d2 UM |
9014 | *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions |
9015 | RAND_egd() and RAND_status(). In the command line application, | |
9016 | the EGD socket can be specified like a seed file using RANDFILE | |
9017 | or -rand. | |
9f0b86c6 | 9018 | [Ulf Möller] |
4ec2d4d2 | 9019 | |
3142c86d DSH |
9020 | *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures. |
9021 | Some CAs (e.g. Verisign) distribute certificates in this form. | |
9022 | [Steve Henson] | |
9023 | ||
9024 | *) Remove the SSL_ALLOW_ADH compile option and set the default cipher | |
9025 | list to exclude them. This means that no special compilation option | |
9026 | is needed to use anonymous DH: it just needs to be included in the | |
9027 | cipher list. | |
9028 | [Steve Henson] | |
9029 | ||
72b60351 DSH |
9030 | *) Change the EVP_MD_CTX_type macro so its meaning consistent with |
9031 | EVP_MD_type. The old functionality is available in a new macro called | |
9032 | EVP_MD_md(). Change code that uses it and update docs. | |
9033 | [Steve Henson] | |
9034 | ||
745c70e5 BM |
9035 | *) ..._ctrl functions now have corresponding ..._callback_ctrl functions |
9036 | where the 'void *' argument is replaced by a function pointer argument. | |
9037 | Previously 'void *' was abused to point to functions, which works on | |
9038 | many platforms, but is not correct. As these functions are usually | |
9039 | called by macros defined in OpenSSL header files, most source code | |
9040 | should work without changes. | |
cdf20e08 | 9041 | [Richard Levitte] |
745c70e5 BM |
9042 | |
9043 | *) <openssl/opensslconf.h> (which is created by Configure) now contains | |
9044 | sections with information on -D... compiler switches used for | |
9045 | compiling the library so that applications can see them. To enable | |
9046 | one of these sections, a pre-processor symbol OPENSSL_..._DEFINES | |
9047 | must be defined. E.g., | |
9048 | #define OPENSSL_ALGORITHM_DEFINES | |
9049 | #include <openssl/opensslconf.h> | |
9050 | defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc. | |
9f0b86c6 | 9051 | [Richard Levitte, Ulf and Bodo Möller] |
745c70e5 | 9052 | |
b35e9050 BM |
9053 | *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS |
9054 | record layer. | |
9055 | [Bodo Moeller] | |
9056 | ||
d754b385 DSH |
9057 | *) Change the 'other' type in certificate aux info to a STACK_OF |
9058 | X509_ALGOR. Although not an AlgorithmIdentifier as such it has | |
9059 | the required ASN1 format: arbitrary types determined by an OID. | |
9060 | [Steve Henson] | |
9061 | ||
8a208cba DSH |
9062 | *) Add some PEM_write_X509_REQ_NEW() functions and a command line |
9063 | argument to 'req'. This is not because the function is newer or | |
9064 | better than others it just uses the work 'NEW' in the certificate | |
9065 | request header lines. Some software needs this. | |
9066 | [Steve Henson] | |
9067 | ||
a3fe382e DSH |
9068 | *) Reorganise password command line arguments: now passwords can be |
9069 | obtained from various sources. Delete the PEM_cb function and make | |
9070 | it the default behaviour: i.e. if the callback is NULL and the | |
9071 | usrdata argument is not NULL interpret it as a null terminated pass | |
9072 | phrase. If usrdata and the callback are NULL then the pass phrase | |
9073 | is prompted for as usual. | |
9074 | [Steve Henson] | |
9075 | ||
bd03b99b BL |
9076 | *) Add support for the Compaq Atalla crypto accelerator. If it is installed, |
9077 | the support is automatically enabled. The resulting binaries will | |
9078 | autodetect the card and use it if present. | |
9079 | [Ben Laurie and Compaq Inc.] | |
9080 | ||
de469ef2 DSH |
9081 | *) Work around for Netscape hang bug. This sends certificate request |
9082 | and server done in one record. Since this is perfectly legal in the | |
9083 | SSL/TLS protocol it isn't a "bug" option and is on by default. See | |
9084 | the bugs/SSLv3 entry for more info. | |
9085 | [Steve Henson] | |
9086 | ||
bcba6cc6 AP |
9087 | *) HP-UX tune-up: new unified configs, HP C compiler bug workaround. |
9088 | [Andy Polyakov] | |
9089 | ||
d13e4eb0 DSH |
9090 | *) Add -rand argument to smime and pkcs12 applications and read/write |
9091 | of seed file. | |
9092 | [Steve Henson] | |
9093 | ||
3ebf0be1 | 9094 | *) New 'passwd' tool for crypt(3) and apr1 password hashes. |
bb325c7d BM |
9095 | [Bodo Moeller] |
9096 | ||
f07fb9b2 DSH |
9097 | *) Add command line password options to the remaining applications. |
9098 | [Steve Henson] | |
9099 | ||
cae55bfc UM |
9100 | *) Bug fix for BN_div_recp() for numerators with an even number of |
9101 | bits. | |
9f0b86c6 | 9102 | [Ulf Möller] |
cae55bfc UM |
9103 | |
9104 | *) More tests in bntest.c, and changed test_bn output. | |
9f0b86c6 | 9105 | [Ulf Möller] |
cae55bfc | 9106 | |
0fad6cb7 AP |
9107 | *) ./config recognizes MacOS X now. |
9108 | [Andy Polyakov] | |
9109 | ||
4a6222d7 UM |
9110 | *) Bug fix for BN_div() when the first words of num and divsor are |
9111 | equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0). | |
9f0b86c6 | 9112 | [Ulf Möller] |
4a6222d7 | 9113 | |
66430207 DSH |
9114 | *) Add support for various broken PKCS#8 formats, and command line |
9115 | options to produce them. | |
9116 | [Steve Henson] | |
9117 | ||
9b141126 UM |
9118 | *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to |
9119 | get temporary BIGNUMs from a BN_CTX. | |
9f0b86c6 | 9120 | [Ulf Möller] |
9b141126 UM |
9121 | |
9122 | *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont() | |
9123 | for p == 0. | |
9f0b86c6 | 9124 | [Ulf Möller] |
9b141126 | 9125 | |
af57d843 DSH |
9126 | *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and |
9127 | include a #define from the old name to the new. The original intent | |
9128 | was that statically linked binaries could for example just call | |
9129 | SSLeay_add_all_ciphers() to just add ciphers to the table and not | |
9130 | link with digests. This never worked becayse SSLeay_add_all_digests() | |
9131 | and SSLeay_add_all_ciphers() were in the same source file so calling | |
9132 | one would link with the other. They are now in separate source files. | |
9133 | [Steve Henson] | |
9134 | ||
82fc1d9c DSH |
9135 | *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'. |
9136 | [Steve Henson] | |
9137 | ||
e74231ed BM |
9138 | *) Use a less unusual form of the Miller-Rabin primality test (it used |
9139 | a binary algorithm for exponentiation integrated into the Miller-Rabin | |
9140 | loop, our standard modexp algorithms are faster). | |
9141 | [Bodo Moeller] | |
9142 | ||
2c5fe5b1 | 9143 | *) Support for the EBCDIC character set completed. |
8efb6014 UM |
9144 | [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>] |
9145 | ||
98d0b2e3 UM |
9146 | *) Source code cleanups: use const where appropriate, eliminate casts, |
9147 | use void * instead of char * in lhash. | |
9f0b86c6 | 9148 | [Ulf Möller] |
98d0b2e3 | 9149 | |
a87030a1 BM |
9150 | *) Bugfix: ssl3_send_server_key_exchange was not restartable |
9151 | (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of | |
9152 | this the server could overwrite ephemeral keys that the client | |
9153 | has already seen). | |
9154 | [Bodo Moeller] | |
9155 | ||
9156 | *) Turn DSA_is_prime into a macro that calls BN_is_prime, | |
9157 | using 50 iterations of the Rabin-Miller test. | |
9158 | ||
9159 | DSA_generate_parameters now uses BN_is_prime_fasttest (with 50 | |
9160 | iterations of the Rabin-Miller test as required by the appendix | |
9161 | to FIPS PUB 186[-1]) instead of DSA_is_prime. | |
9162 | As BN_is_prime_fasttest includes trial division, DSA parameter | |
9163 | generation becomes much faster. | |
9164 | ||
9165 | This implies a change for the callback functions in DSA_is_prime | |
cdd43b5b BM |
9166 | and DSA_generate_parameters: The callback function is called once |
9167 | for each positive witness in the Rabin-Miller test, not just | |
9168 | occasionally in the inner loop; and the parameters to the | |
9169 | callback function now provide an iteration count for the outer | |
9170 | loop rather than for the current invocation of the inner loop. | |
9171 | DSA_generate_parameters additionally can call the callback | |
9172 | function with an 'iteration count' of -1, meaning that a | |
9173 | candidate has passed the trial division test (when q is generated | |
9174 | from an application-provided seed, trial division is skipped). | |
a87030a1 BM |
9175 | [Bodo Moeller] |
9176 | ||
7865b871 | 9177 | *) New function BN_is_prime_fasttest that optionally does trial |
a87030a1 BM |
9178 | division before starting the Rabin-Miller test and has |
9179 | an additional BN_CTX * argument (whereas BN_is_prime always | |
9180 | has to allocate at least one BN_CTX). | |
1baa9490 BM |
9181 | 'callback(1, -1, cb_arg)' is called when a number has passed the |
9182 | trial division stage. | |
9183 | [Bodo Moeller] | |
a87030a1 | 9184 | |
e1314b57 DSH |
9185 | *) Fix for bug in CRL encoding. The validity dates weren't being handled |
9186 | as ASN1_TIME. | |
9187 | [Steve Henson] | |
9188 | ||
90644dd7 DSH |
9189 | *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file. |
9190 | [Steve Henson] | |
9191 | ||
38e33cef | 9192 | *) New function BN_pseudo_rand(). |
9f0b86c6 | 9193 | [Ulf Möller] |
d91e201e | 9194 | |
e93f9a32 UM |
9195 | *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable) |
9196 | bignum version of BN_from_montgomery() with the working code from | |
9197 | SSLeay 0.9.0 (the word based version is faster anyway), and clean up | |
9198 | the comments. | |
9f0b86c6 | 9199 | [Ulf Möller] |
e93f9a32 | 9200 | |
2557eaea BM |
9201 | *) Avoid a race condition in s2_clnt.c (function get_server_hello) that |
9202 | made it impossible to use the same SSL_SESSION data structure in | |
9203 | SSL2 clients in multiple threads. | |
9204 | [Bodo Moeller] | |
9205 | ||
a46faa2b BM |
9206 | *) The return value of RAND_load_file() no longer counts bytes obtained |
9207 | by stat(). RAND_load_file(..., -1) is new and uses the complete file | |
9208 | to seed the PRNG (previously an explicit byte count was required). | |
9f0b86c6 | 9209 | [Ulf Möller, Bodo Möller] |
aabbb745 | 9210 | |
dd9d233e DSH |
9211 | *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes |
9212 | used (char *) instead of (void *) and had casts all over the place. | |
9213 | [Steve Henson] | |
9214 | ||
4486d0cd | 9215 | *) Make BN_generate_prime() return NULL on error if ret!=NULL. |
9f0b86c6 | 9216 | [Ulf Möller] |
4486d0cd | 9217 | |
a87030a1 BM |
9218 | *) Retain source code compatibility for BN_prime_checks macro: |
9219 | BN_is_prime(..., BN_prime_checks, ...) now uses | |
9220 | BN_prime_checks_for_size to determine the appropriate number of | |
9221 | Rabin-Miller iterations. | |
9f0b86c6 | 9222 | [Ulf Möller] |
4486d0cd UM |
9223 | |
9224 | *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to | |
9225 | DH_CHECK_P_NOT_SAFE_PRIME. | |
9226 | (Check if this is true? OpenPGP calls them "strong".) | |
9f0b86c6 | 9227 | [Ulf Möller] |
4486d0cd | 9228 | |
09483c58 DSH |
9229 | *) Merge the functionality of "dh" and "gendh" programs into a new program |
9230 | "dhparam". The old programs are retained for now but will handle DH keys | |
9231 | (instead of parameters) in future. | |
9232 | [Steve Henson] | |
9233 | ||
fabce041 DSH |
9234 | *) Make the ciphers, s_server and s_client programs check the return values |
9235 | when a new cipher list is set. | |
9236 | [Steve Henson] | |
9237 | ||
9238 | *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit | |
9239 | ciphers. Before when the 56bit ciphers were enabled the sorting was | |
9240 | wrong. | |
9241 | ||
9242 | The syntax for the cipher sorting has been extended to support sorting by | |
9243 | cipher-strength (using the strength_bits hard coded in the tables). | |
9244 | The new command is "@STRENGTH" (see also doc/apps/ciphers.pod). | |
9245 | ||
9246 | Fix a bug in the cipher-command parser: when supplying a cipher command | |
9247 | string with an "undefined" symbol (neither command nor alphanumeric | |
9248 | [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now | |
9249 | an error is flagged. | |
9250 | ||
9251 | Due to the strength-sorting extension, the code of the | |
9252 | ssl_create_cipher_list() function was completely rearranged. I hope that | |
9253 | the readability was also increased :-) | |
9254 | [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>] | |
09483c58 | 9255 | |
8100490a DSH |
9256 | *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1 |
9257 | for the first serial number and places 2 in the serial number file. This | |
9258 | avoids problems when the root CA is created with serial number zero and | |
9259 | the first user certificate has the same issuer name and serial number | |
9260 | as the root CA. | |
9261 | [Steve Henson] | |
9262 | ||
6e6bc352 DSH |
9263 | *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses |
9264 | the new code. Add documentation for this stuff. | |
9265 | [Steve Henson] | |
9266 | ||
77b47b90 DSH |
9267 | *) Changes to X509_ATTRIBUTE utilities. These have been renamed from |
9268 | X509_*() to X509at_*() on the grounds that they don't handle X509 | |
9269 | structures and behave in an analagous way to the X509v3 functions: | |
9270 | they shouldn't be called directly but wrapper functions should be used | |
9271 | instead. | |
9272 | ||
9273 | So we also now have some wrapper functions that call the X509at functions | |
9274 | when passed certificate requests. (TO DO: similar things can be done with | |
9275 | PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other | |
9276 | things. Some of these need some d2i or i2d and print functionality | |
6e6bc352 | 9277 | because they handle more complex structures.) |
77b47b90 DSH |
9278 | [Steve Henson] |
9279 | ||
aa82db4f UM |
9280 | *) Add missing #ifndefs that caused missing symbols when building libssl |
9281 | as a shared library without RSA. Use #ifndef NO_SSL2 instead of | |
9282 | NO_RSA in ssl/s2*.c. | |
9f0b86c6 | 9283 | [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller] |
aa82db4f | 9284 | |
eb952088 | 9285 | *) Precautions against using the PRNG uninitialized: RAND_bytes() now |
0983760d BM |
9286 | has a return value which indicates the quality of the random data |
9287 | (1 = ok, 0 = not seeded). Also an error is recorded on the thread's | |
373b575f | 9288 | error queue. New function RAND_pseudo_bytes() generates output that is |
853f757e BM |
9289 | guaranteed to be unique but not unpredictable. RAND_add is like |
9290 | RAND_seed, but takes an extra argument for an entropy estimate | |
9291 | (RAND_seed always assumes full entropy). | |
9f0b86c6 | 9292 | [Ulf Möller] |
eb952088 | 9293 | |
76aa0ddc BM |
9294 | *) Do more iterations of Rabin-Miller probable prime test (specifically, |
9295 | 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes | |
a87030a1 | 9296 | instead of only 2 for all lengths; see BN_prime_checks_for_size definition |
4486d0cd | 9297 | in crypto/bn/bn_prime.c for the complete table). This guarantees a |
a87030a1 | 9298 | false-positive rate of at most 2^-80 for random input. |
76aa0ddc BM |
9299 | [Bodo Moeller] |
9300 | ||
3cc6cdea | 9301 | *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs. |
c51ae173 BM |
9302 | [Bodo Moeller] |
9303 | ||
6d0d5431 BM |
9304 | *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain |
9305 | in the 0.9.5 release), this returns the chain | |
25f923dd DSH |
9306 | from an X509_CTX structure with a dup of the stack and all |
9307 | the X509 reference counts upped: so the stack will exist | |
9308 | after X509_CTX_cleanup() has been called. Modify pkcs12.c | |
9309 | to use this. | |
9310 | ||
9311 | Also make SSL_SESSION_print() print out the verify return | |
9312 | code. | |
9313 | [Steve Henson] | |
9314 | ||
dad666fb DSH |
9315 | *) Add manpage for the pkcs12 command. Also change the default |
9316 | behaviour so MAC iteration counts are used unless the new | |
9317 | -nomaciter option is used. This improves file security and | |
9318 | only older versions of MSIE (4.0 for example) need it. | |
9319 | [Steve Henson] | |
9320 | ||
0f583f69 | 9321 | *) Honor the no-xxx Configure options when creating .DEF files. |
9f0b86c6 | 9322 | [Ulf Möller] |
0f583f69 | 9323 | |
35f4850a DSH |
9324 | *) Add PKCS#10 attributes to field table: challengePassword, |
9325 | unstructuredName and unstructuredAddress. These are taken from | |
9326 | draft PKCS#9 v2.0 but are compatible with v1.2 provided no | |
9327 | international characters are used. | |
9328 | ||
9329 | More changes to X509_ATTRIBUTE code: allow the setting of types | |
9330 | based on strings. Remove the 'loc' parameter when adding | |
9331 | attributes because these will be a SET OF encoding which is sorted | |
9332 | in ASN1 order. | |
9333 | [Steve Henson] | |
9334 | ||
b38f9f66 DSH |
9335 | *) Initial changes to the 'req' utility to allow request generation |
9336 | automation. This will allow an application to just generate a template | |
9337 | file containing all the field values and have req construct the | |
9338 | request. | |
9339 | ||
9340 | Initial support for X509_ATTRIBUTE handling. Stacks of these are | |
9341 | used all over the place including certificate requests and PKCS#7 | |
9342 | structures. They are currently handled manually where necessary with | |
9343 | some primitive wrappers for PKCS#7. The new functions behave in a | |
0f583f69 | 9344 | manner analogous to the X509 extension functions: they allow |
b38f9f66 DSH |
9345 | attributes to be looked up by NID and added. |
9346 | ||
9347 | Later something similar to the X509V3 code would be desirable to | |
9348 | automatically handle the encoding, decoding and printing of the | |
9349 | more complex types. The string types like challengePassword can | |
0f583f69 | 9350 | be handled by the string table functions. |
b38f9f66 DSH |
9351 | |
9352 | Also modified the multi byte string table handling. Now there is | |
9353 | a 'global mask' which masks out certain types. The table itself | |
9354 | can use the flag STABLE_NO_MASK to ignore the mask setting: this | |
9355 | is useful when for example there is only one permissible type | |
9356 | (as in countryName) and using the mask might result in no valid | |
9357 | types at all. | |
9358 | [Steve Henson] | |
9359 | ||
ca03109c BM |
9360 | *) Clean up 'Finished' handling, and add functions SSL_get_finished and |
9361 | SSL_get_peer_finished to allow applications to obtain the latest | |
9362 | Finished messages sent to the peer or expected from the peer, | |
9363 | respectively. (SSL_get_peer_finished is usually the Finished message | |
9364 | actually received from the peer, otherwise the protocol will be aborted.) | |
9365 | ||
9366 | As the Finished message are message digests of the complete handshake | |
9367 | (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can | |
9368 | be used for external authentication procedures when the authentication | |
9369 | provided by SSL/TLS is not desired or is not enough. | |
f2d9a32c BM |
9370 | [Bodo Moeller] |
9371 | ||
bdf5e183 AP |
9372 | *) Enhanced support for Alpha Linux is added. Now ./config checks if |
9373 | the host supports BWX extension and if Compaq C is present on the | |
0f583f69 | 9374 | $PATH. Just exploiting of the BWX extension results in 20-30% |
bdf5e183 AP |
9375 | performance kick for some algorithms, e.g. DES and RC4 to mention |
9376 | a couple. Compaq C in turn generates ~20% faster code for MD5 and | |
9377 | SHA1. | |
9378 | [Andy Polyakov] | |
9379 | ||
3d14b9d0 DSH |
9380 | *) Add support for MS "fast SGC". This is arguably a violation of the |
9381 | SSL3/TLS protocol. Netscape SGC does two handshakes: the first with | |
9382 | weak crypto and after checking the certificate is SGC a second one | |
9383 | with strong crypto. MS SGC stops the first handshake after receiving | |
9384 | the server certificate message and sends a second client hello. Since | |
9385 | a server will typically do all the time consuming operations before | |
9386 | expecting any further messages from the client (server key exchange | |
9387 | is the most expensive) there is little difference between the two. | |
9388 | ||
9389 | To get OpenSSL to support MS SGC we have to permit a second client | |
9390 | hello message after we have sent server done. In addition we have to | |
745c70e5 | 9391 | reset the MAC if we do get this second client hello. |
3d14b9d0 DSH |
9392 | [Steve Henson] |
9393 | ||
20432eae DSH |
9394 | *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide |
9395 | if a DER encoded private key is RSA or DSA traditional format. Changed | |
9396 | d2i_PrivateKey_bio() to use it. This is only needed for the "traditional" | |
9397 | format DER encoded private key. Newer code should use PKCS#8 format which | |
9398 | has the key type encoded in the ASN1 structure. Added DER private key | |
9399 | support to pkcs8 application. | |
9400 | [Steve Henson] | |
9401 | ||
47134b78 BM |
9402 | *) SSL 3/TLS 1 servers now don't request certificates when an anonymous |
9403 | ciphersuites has been selected (as required by the SSL 3/TLS 1 | |
9404 | specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT | |
9405 | is set, we interpret this as a request to violate the specification | |
9406 | (the worst that can happen is a handshake failure, and 'correct' | |
9407 | behaviour would result in a handshake failure anyway). | |
9408 | [Bodo Moeller] | |
9409 | ||
45fd4dbb BM |
9410 | *) In SSL_CTX_add_session, take into account that there might be multiple |
9411 | SSL_SESSION structures with the same session ID (e.g. when two threads | |
9412 | concurrently obtain them from an external cache). | |
9413 | The internal cache can handle only one SSL_SESSION with a given ID, | |
9414 | so if there's a conflict, we now throw out the old one to achieve | |
9415 | consistency. | |
9416 | [Bodo Moeller] | |
9417 | ||
f45f40ff DSH |
9418 | *) Add OIDs for idea and blowfish in CBC mode. This will allow both |
9419 | to be used in PKCS#5 v2.0 and S/MIME. Also add checking to | |
9420 | some routines that use cipher OIDs: some ciphers do not have OIDs | |
9421 | defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for | |
9422 | example. | |
9423 | [Steve Henson] | |
9424 | ||
6447cce3 DSH |
9425 | *) Simplify the trust setting structure and code. Now we just have |
9426 | two sequences of OIDs for trusted and rejected settings. These will | |
9427 | typically have values the same as the extended key usage extension | |
9428 | and any application specific purposes. | |
9429 | ||
9430 | The trust checking code now has a default behaviour: it will just | |
9431 | check for an object with the same NID as the passed id. Functions can | |
9432 | be provided to override either the default behaviour or the behaviour | |
9433 | for a given id. SSL client, server and email already have functions | |
20432eae | 9434 | in place for compatibility: they check the NID and also return "trusted" |
6447cce3 DSH |
9435 | if the certificate is self signed. |
9436 | [Steve Henson] | |
9437 | ||
e6f3c585 DSH |
9438 | *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the |
9439 | traditional format into an EVP_PKEY structure. | |
9440 | [Steve Henson] | |
9441 | ||
36217a94 DSH |
9442 | *) Add a password callback function PEM_cb() which either prompts for |
9443 | a password if usr_data is NULL or otherwise assumes it is a null | |
e6f3c585 | 9444 | terminated password. Allow passwords to be passed on command line |
36217a94 DSH |
9445 | environment or config files in a few more utilities. |
9446 | [Steve Henson] | |
9447 | ||
525f51f6 DSH |
9448 | *) Add a bunch of DER and PEM functions to handle PKCS#8 format private |
9449 | keys. Add some short names for PKCS#8 PBE algorithms and allow them | |
9450 | to be specified on the command line for the pkcs8 and pkcs12 utilities. | |
9451 | Update documentation. | |
9452 | [Steve Henson] | |
9453 | ||
e76f935e DSH |
9454 | *) Support for ASN1 "NULL" type. This could be handled before by using |
9455 | ASN1_TYPE but there wasn't any function that would try to read a NULL | |
78baa17a | 9456 | and produce an error if it couldn't. For compatibility we also have |
e76f935e DSH |
9457 | ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and |
9458 | don't allocate anything because they don't need to. | |
9459 | [Steve Henson] | |
9460 | ||
099f1b32 AP |
9461 | *) Initial support for MacOS is now provided. Examine INSTALL.MacOS |
9462 | for details. | |
9463 | [Andy Polyakov, Roy Woods <roy@centicsystems.ca>] | |
9464 | ||
9ac42ed8 RL |
9465 | *) Rebuild of the memory allocation routines used by OpenSSL code and |
9466 | possibly others as well. The purpose is to make an interface that | |
9467 | provide hooks so anyone can build a separate set of allocation and | |
cbfa4c32 RL |
9468 | deallocation routines to be used by OpenSSL, for example memory |
9469 | pool implementations, or something else, which was previously hard | |
9470 | since Malloc(), Realloc() and Free() were defined as macros having | |
de73e397 RL |
9471 | the values malloc, realloc and free, respectively (except for Win32 |
9472 | compilations). The same is provided for memory debugging code. | |
9473 | OpenSSL already comes with functionality to find memory leaks, but | |
9474 | this gives people a chance to debug other memory problems. | |
d8df48a9 | 9475 | |
f3a2a044 RL |
9476 | With these changes, a new set of functions and macros have appeared: |
9477 | ||
2c05c494 BM |
9478 | CRYPTO_set_mem_debug_functions() [F] |
9479 | CRYPTO_get_mem_debug_functions() [F] | |
9480 | CRYPTO_dbg_set_options() [F] | |
9481 | CRYPTO_dbg_get_options() [F] | |
9482 | CRYPTO_malloc_debug_init() [M] | |
f3a2a044 RL |
9483 | |
9484 | The memory debug functions are NULL by default, unless the library | |
9485 | is compiled with CRYPTO_MDEBUG or friends is defined. If someone | |
cbfa4c32 RL |
9486 | wants to debug memory anyway, CRYPTO_malloc_debug_init() (which |
9487 | gives the standard debugging functions that come with OpenSSL) or | |
9488 | CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions | |
9489 | provided by the library user) must be used. When the standard | |
9490 | debugging functions are used, CRYPTO_dbg_set_options can be used to | |
9491 | request additional information: | |
9492 | CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting | |
9493 | the CRYPTO_MDEBUG_xxx macro when compiling the library. | |
f3a2a044 RL |
9494 | |
9495 | Also, things like CRYPTO_set_mem_functions will always give the | |
9496 | expected result (the new set of functions is used for allocation | |
9497 | and deallocation) at all times, regardless of platform and compiler | |
9498 | options. | |
9499 | ||
9500 | To finish it up, some functions that were never use in any other | |
9501 | way than through macros have a new API and new semantic: | |
9502 | ||
9503 | CRYPTO_dbg_malloc() | |
9504 | CRYPTO_dbg_realloc() | |
9505 | CRYPTO_dbg_free() | |
9506 | ||
9507 | All macros of value have retained their old syntax. | |
cbfa4c32 | 9508 | [Richard Levitte and Bodo Moeller] |
9ac42ed8 | 9509 | |
b216664f DSH |
9510 | *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the |
9511 | ordering of SMIMECapabilities wasn't in "strength order" and there | |
9512 | was a missing NULL in the AlgorithmIdentifier for the SHA1 signature | |
9513 | algorithm. | |
9514 | [Steve Henson] | |
9515 | ||
d8223efd DSH |
9516 | *) Some ASN1 types with illegal zero length encoding (INTEGER, |
9517 | ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines. | |
9518 | [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson] | |
9519 | ||
5a9a4b29 DSH |
9520 | *) Merge in my S/MIME library for OpenSSL. This provides a simple |
9521 | S/MIME API on top of the PKCS#7 code, a MIME parser (with enough | |
9522 | functionality to handle multipart/signed properly) and a utility | |
9523 | called 'smime' to call all this stuff. This is based on code I | |
9524 | originally wrote for Celo who have kindly allowed it to be | |
9525 | included in OpenSSL. | |
9526 | [Steve Henson] | |
9527 | ||
cddfe788 BM |
9528 | *) Add variants des_set_key_checked and des_set_key_unchecked of |
9529 | des_set_key (aka des_key_sched). Global variable des_check_key | |
9530 | decides which of these is called by des_set_key; this way | |
9531 | des_check_key behaves as it always did, but applications and | |
9532 | the library itself, which was buggy for des_check_key == 1, | |
9533 | have a cleaner way to pick the version they need. | |
9534 | [Bodo Moeller] | |
9535 | ||
21131f00 DSH |
9536 | *) New function PKCS12_newpass() which changes the password of a |
9537 | PKCS12 structure. | |
9538 | [Steve Henson] | |
9539 | ||
dd413410 DSH |
9540 | *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and |
9541 | dynamic mix. In both cases the ids can be used as an index into the | |
9542 | table. Also modified the X509_TRUST_add() and X509_PURPOSE_add() | |
9543 | functions so they accept a list of the field values and the | |
9544 | application doesn't need to directly manipulate the X509_TRUST | |
9545 | structure. | |
9546 | [Steve Henson] | |
9547 | ||
9548 | *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't | |
9549 | need initialising. | |
9550 | [Steve Henson] | |
9551 | ||
08cba610 DSH |
9552 | *) Modify the way the V3 extension code looks up extensions. This now |
9553 | works in a similar way to the object code: we have some "standard" | |
9554 | extensions in a static table which is searched with OBJ_bsearch() | |
9555 | and the application can add dynamic ones if needed. The file | |
9556 | crypto/x509v3/ext_dat.h now has the info: this file needs to be | |
9557 | updated whenever a new extension is added to the core code and kept | |
9558 | in ext_nid order. There is a simple program 'tabtest.c' which checks | |
9559 | this. New extensions are not added too often so this file can readily | |
9560 | be maintained manually. | |
9561 | ||
9562 | There are two big advantages in doing things this way. The extensions | |
9563 | can be looked up immediately and no longer need to be "added" using | |
9564 | X509V3_add_standard_extensions(): this function now does nothing. | |
9565 | [Side note: I get *lots* of email saying the extension code doesn't | |
9566 | work because people forget to call this function] | |
9567 | Also no dynamic allocation is done unless new extensions are added: | |
9568 | so if we don't add custom extensions there is no need to call | |
9569 | X509V3_EXT_cleanup(). | |
9570 | [Steve Henson] | |
9571 | ||
fea9afbf BL |
9572 | *) Modify enc utility's salting as follows: make salting the default. Add a |
9573 | magic header, so unsalted files fail gracefully instead of just decrypting | |
9574 | to garbage. This is because not salting is a big security hole, so people | |
9575 | should be discouraged from doing it. | |
9576 | [Ben Laurie] | |
9577 | ||
9868232a DSH |
9578 | *) Fixes and enhancements to the 'x509' utility. It allowed a message |
9579 | digest to be passed on the command line but it only used this | |
9580 | parameter when signing a certificate. Modified so all relevant | |
9581 | operations are affected by the digest parameter including the | |
9582 | -fingerprint and -x509toreq options. Also -x509toreq choked if a | |
9583 | DSA key was used because it didn't fix the digest. | |
9584 | [Steve Henson] | |
9585 | ||
51630a37 DSH |
9586 | *) Initial certificate chain verify code. Currently tests the untrusted |
9587 | certificates for consistency with the verify purpose (which is set | |
9588 | when the X509_STORE_CTX structure is set up) and checks the pathlength. | |
9589 | ||
9590 | There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour: | |
bb7cd4e3 DSH |
9591 | this is because it will reject chains with invalid extensions whereas |
9592 | every previous version of OpenSSL and SSLeay made no checks at all. | |
51630a37 DSH |
9593 | |
9594 | Trust code: checks the root CA for the relevant trust settings. Trust | |
9595 | settings have an initial value consistent with the verify purpose: e.g. | |
9596 | if the verify purpose is for SSL client use it expects the CA to be | |
9597 | trusted for SSL client use. However the default value can be changed to | |
9598 | permit custom trust settings: one example of this would be to only trust | |
9599 | certificates from a specific "secure" set of CAs. | |
11262391 DSH |
9600 | |
9601 | Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions | |
9602 | which should be used for version portability: especially since the | |
9603 | verify structure is likely to change more often now. | |
d4cec6a1 | 9604 | |
bb7cd4e3 DSH |
9605 | SSL integration. Add purpose and trust to SSL_CTX and SSL and functions |
9606 | to set them. If not set then assume SSL clients will verify SSL servers | |
9607 | and vice versa. | |
9608 | ||
d4cec6a1 DSH |
9609 | Two new options to the verify program: -untrusted allows a set of |
9610 | untrusted certificates to be passed in and -purpose which sets the | |
9611 | intended purpose of the certificate. If a purpose is set then the | |
9612 | new chain verify code is used to check extension consistency. | |
11262391 DSH |
9613 | [Steve Henson] |
9614 | ||
9615 | *) Support for the authority information access extension. | |
6d3724d3 DSH |
9616 | [Steve Henson] |
9617 | ||
52664f50 DSH |
9618 | *) Modify RSA and DSA PEM read routines to transparently handle |
9619 | PKCS#8 format private keys. New *_PUBKEY_* functions that handle | |
9620 | public keys in a format compatible with certificate | |
9621 | SubjectPublicKeyInfo structures. Unfortunately there were already | |
9622 | functions called *_PublicKey_* which used various odd formats so | |
78baa17a | 9623 | these are retained for compatibility: however the DSA variants were |
52664f50 DSH |
9624 | never in a public release so they have been deleted. Changed dsa/rsa |
9625 | utilities to handle the new format: note no releases ever handled public | |
9626 | keys so we should be OK. | |
9627 | ||
9628 | The primary motivation for this change is to avoid the same fiasco | |
9629 | that dogs private keys: there are several incompatible private key | |
9630 | formats some of which are standard and some OpenSSL specific and | |
9631 | require various evil hacks to allow partial transparent handling and | |
9632 | even then it doesn't work with DER formats. Given the option anything | |
9633 | other than PKCS#8 should be dumped: but the other formats have to | |
78baa17a | 9634 | stay in the name of compatibility. |
52664f50 DSH |
9635 | |
9636 | With public keys and the benefit of hindsight one standard format | |
9637 | is used which works with EVP_PKEY, RSA or DSA structures: though | |
9638 | it clearly returns an error if you try to read the wrong kind of key. | |
9639 | ||
9640 | Added a -pubkey option to the 'x509' utility to output the public key. | |
6d0d5431 BM |
9641 | Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*() |
9642 | (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add | |
9643 | EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*()) | |
9644 | that do the same as the EVP_PKEY_assign_*() except they up the | |
9645 | reference count of the added key (they don't "swallow" the | |
9646 | supplied key). | |
52664f50 DSH |
9647 | [Steve Henson] |
9648 | ||
9649 | *) Fixes to crypto/x509/by_file.c the code to read in certificates and | |
9650 | CRLs would fail if the file contained no certificates or no CRLs: | |
9651 | added a new function to read in both types and return the number | |
9652 | read: this means that if none are read it will be an error. The | |
9653 | DER versions of the certificate and CRL reader would always fail | |
9654 | because it isn't possible to mix certificates and CRLs in DER format | |
9655 | without choking one or the other routine. Changed this to just read | |
9656 | a certificate: this is the best we can do. Also modified the code | |
9657 | in apps/verify.c to take notice of return codes: it was previously | |
9658 | attempting to read in certificates from NULL pointers and ignoring | |
9659 | any errors: this is one reason why the cert and CRL reader seemed | |
9660 | to work. It doesn't check return codes from the default certificate | |
9661 | routines: these may well fail if the certificates aren't installed. | |
9662 | [Steve Henson] | |
9663 | ||
a716d727 DSH |
9664 | *) Code to support otherName option in GeneralName. |
9665 | [Steve Henson] | |
9666 | ||
f76d8c47 DSH |
9667 | *) First update to verify code. Change the verify utility |
9668 | so it warns if it is passed a self signed certificate: | |
9669 | for consistency with the normal behaviour. X509_verify | |
9670 | has been modified to it will now verify a self signed | |
9671 | certificate if *exactly* the same certificate appears | |
9672 | in the store: it was previously impossible to trust a | |
9673 | single self signed certificate. This means that: | |
9674 | openssl verify ss.pem | |
9675 | now gives a warning about a self signed certificate but | |
9676 | openssl verify -CAfile ss.pem ss.pem | |
9677 | is OK. | |
9678 | [Steve Henson] | |
9679 | ||
b1fe6ca1 BM |
9680 | *) For servers, store verify_result in SSL_SESSION data structure |
9681 | (and add it to external session representation). | |
9682 | This is needed when client certificate verifications fails, | |
9683 | but an application-provided verification callback (set by | |
9684 | SSL_CTX_set_cert_verify_callback) allows accepting the session | |
9685 | anyway (i.e. leaves x509_store_ctx->error != X509_V_OK | |
9686 | but returns 1): When the session is reused, we have to set | |
9687 | ssl->verify_result to the appropriate error code to avoid | |
9688 | security holes. | |
9689 | [Bodo Moeller, problem pointed out by Lutz Jaenicke] | |
9690 | ||
91895a59 DSH |
9691 | *) Fix a bug in the new PKCS#7 code: it didn't consider the |
9692 | case in PKCS7_dataInit() where the signed PKCS7 structure | |
9693 | didn't contain any existing data because it was being created. | |
f76d8c47 | 9694 | [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson] |
91895a59 | 9695 | |
fd699ac5 DSH |
9696 | *) Add a salt to the key derivation routines in enc.c. This |
9697 | forms the first 8 bytes of the encrypted file. Also add a | |
9698 | -S option to allow a salt to be input on the command line. | |
9699 | [Steve Henson] | |
9700 | ||
e947f396 DSH |
9701 | *) New function X509_cmp(). Oddly enough there wasn't a function |
9702 | to compare two certificates. We do this by working out the SHA1 | |
9703 | hash and comparing that. X509_cmp() will be needed by the trust | |
9704 | code. | |
9705 | [Steve Henson] | |
9706 | ||
07e6dbde BM |
9707 | *) SSL_get1_session() is like SSL_get_session(), but increments |
9708 | the reference count in the SSL_SESSION returned. | |
b7cfcfb7 MC |
9709 | [Geoff Thorpe <geoff@eu.c2.net>] |
9710 | ||
06556a17 DSH |
9711 | *) Fix for 'req': it was adding a null to request attributes. |
9712 | Also change the X509_LOOKUP and X509_INFO code to handle | |
9713 | certificate auxiliary information. | |
9714 | [Steve Henson] | |
9715 | ||
a0e9f529 DSH |
9716 | *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document |
9717 | the 'enc' command. | |
9718 | [Steve Henson] | |
9719 | ||
71d7526b RL |
9720 | *) Add the possibility to add extra information to the memory leak |
9721 | detecting output, to form tracebacks, showing from where each | |
a873356c BM |
9722 | allocation was originated: CRYPTO_push_info("constant string") adds |
9723 | the string plus current file name and line number to a per-thread | |
9724 | stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info() | |
9725 | is like calling CYRPTO_pop_info() until the stack is empty. | |
9726 | Also updated memory leak detection code to be multi-thread-safe. | |
71d7526b RL |
9727 | [Richard Levitte] |
9728 | ||
a0e9f529 | 9729 | *) Add options -text and -noout to pkcs7 utility and delete the |
954ef7ef DSH |
9730 | encryption options which never did anything. Update docs. |
9731 | [Steve Henson] | |
9732 | ||
af29811e DSH |
9733 | *) Add options to some of the utilities to allow the pass phrase |
9734 | to be included on either the command line (not recommended on | |
9735 | OSes like Unix) or read from the environment. Update the | |
9736 | manpages and fix a few bugs. | |
9737 | [Steve Henson] | |
9738 | ||
aba3e65f DSH |
9739 | *) Add a few manpages for some of the openssl commands. |
9740 | [Steve Henson] | |
9741 | ||
a0ad17bb DSH |
9742 | *) Fix the -revoke option in ca. It was freeing up memory twice, |
9743 | leaking and not finding already revoked certificates. | |
9744 | [Steve Henson] | |
9745 | ||
ce1b4fe1 DSH |
9746 | *) Extensive changes to support certificate auxiliary information. |
9747 | This involves the use of X509_CERT_AUX structure and X509_AUX | |
9748 | functions. An X509_AUX function such as PEM_read_X509_AUX() | |
9749 | can still read in a certificate file in the usual way but it | |
9750 | will also read in any additional "auxiliary information". By | |
78baa17a | 9751 | doing things this way a fair degree of compatibility can be |
ce1b4fe1 DSH |
9752 | retained: existing certificates can have this information added |
9753 | using the new 'x509' options. | |
9754 | ||
9755 | Current auxiliary information includes an "alias" and some trust | |
9756 | settings. The trust settings will ultimately be used in enhanced | |
9757 | certificate chain verification routines: currently a certificate | |
9758 | can only be trusted if it is self signed and then it is trusted | |
9759 | for all purposes. | |
9760 | [Steve Henson] | |
9761 | ||
a873356c BM |
9762 | *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD). |
9763 | The problem was that one of the replacement routines had not been working | |
9764 | since SSLeay releases. For now the offending routine has been replaced | |
9765 | with non-optimised assembler. Even so, this now gives around 95% | |
9766 | performance improvement for 1024 bit RSA signs. | |
ce2c95b2 MC |
9767 | [Mark Cox] |
9768 | ||
9716a8f9 DSH |
9769 | *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2 |
9770 | handling. Most clients have the effective key size in bits equal to | |
9771 | the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key. | |
9772 | A few however don't do this and instead use the size of the decrypted key | |
9773 | to determine the RC2 key length and the AlgorithmIdentifier to determine | |
0f583f69 | 9774 | the effective key length. In this case the effective key length can still |
9716a8f9 DSH |
9775 | be 40 bits but the key length can be 168 bits for example. This is fixed |
9776 | by manually forcing an RC2 key into the EVP_PKEY structure because the | |
9777 | EVP code can't currently handle unusual RC2 key sizes: it always assumes | |
9778 | the key length and effective key length are equal. | |
9779 | [Steve Henson] | |
9780 | ||
74400f73 DSH |
9781 | *) Add a bunch of functions that should simplify the creation of |
9782 | X509_NAME structures. Now you should be able to do: | |
9783 | X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0); | |
9784 | and have it automatically work out the correct field type and fill in | |
9785 | the structures. The more adventurous can try: | |
9786 | X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0); | |
9787 | and it will (hopefully) work out the correct multibyte encoding. | |
9788 | [Steve Henson] | |
9789 | ||
9790 | *) Change the 'req' utility to use the new field handling and multibyte | |
9791 | copy routines. Before the DN field creation was handled in an ad hoc | |
9792 | way in req, ca, and x509 which was rather broken and didn't support | |
9793 | BMPStrings or UTF8Strings. Since some software doesn't implement | |
9794 | BMPStrings or UTF8Strings yet, they can be enabled using the config file | |
9795 | using the dirstring_type option. See the new comment in the default | |
9796 | openssl.cnf for more info. | |
9797 | [Steve Henson] | |
9798 | ||
c1e744b9 | 9799 | *) Make crypto/rand/md_rand.c more robust: |
62ac2938 | 9800 | - Assure unique random numbers after fork(). |
c1e744b9 BM |
9801 | - Make sure that concurrent threads access the global counter and |
9802 | md serializably so that we never lose entropy in them | |
9803 | or use exactly the same state in multiple threads. | |
9804 | Access to the large state is not always serializable because | |
9805 | the additional locking could be a performance killer, and | |
9806 | md should be large enough anyway. | |
9807 | [Bodo Moeller] | |
9808 | ||
a31011e8 BM |
9809 | *) New file apps/app_rand.c with commonly needed functionality |
9810 | for handling the random seed file. | |
9811 | ||
9812 | Use the random seed file in some applications that previously did not: | |
9813 | ca, | |
78baa17a | 9814 | dsaparam -genkey (which also ignored its '-rand' option), |
a31011e8 BM |
9815 | s_client, |
9816 | s_server, | |
9817 | x509 (when signing). | |
9818 | Except on systems with /dev/urandom, it is crucial to have a random | |
9819 | seed file at least for key creation, DSA signing, and for DH exchanges; | |
99e87569 | 9820 | for RSA signatures we could do without one. |
a31011e8 BM |
9821 | |
9822 | gendh and gendsa (unlike genrsa) used to read only the first byte | |
78baa17a | 9823 | of each file listed in the '-rand' option. The function as previously |
a31011e8 | 9824 | found in genrsa is now in app_rand.c and is used by all programs |
78baa17a | 9825 | that support '-rand'. |
a31011e8 BM |
9826 | [Bodo Moeller] |
9827 | ||
9828 | *) In RAND_write_file, use mode 0600 for creating files; | |
9829 | don't just chmod when it may be too late. | |
9830 | [Bodo Moeller] | |
9831 | ||
9832 | *) Report an error from X509_STORE_load_locations | |
9833 | when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed. | |
9834 | [Bill Perry] | |
9835 | ||
462f79ec DSH |
9836 | *) New function ASN1_mbstring_copy() this copies a string in either |
9837 | ASCII, Unicode, Universal (4 bytes per character) or UTF8 format | |
9838 | into an ASN1_STRING type. A mask of permissible types is passed | |
9839 | and it chooses the "minimal" type to use or an error if not type | |
9840 | is suitable. | |
9841 | [Steve Henson] | |
9842 | ||
08e9c1af DSH |
9843 | *) Add function equivalents to the various macros in asn1.h. The old |
9844 | macros are retained with an M_ prefix. Code inside the library can | |
9845 | use the M_ macros. External code (including the openssl utility) | |
9846 | should *NOT* in order to be "shared library friendly". | |
9847 | [Steve Henson] | |
9848 | ||
673b102c DSH |
9849 | *) Add various functions that can check a certificate's extensions |
9850 | to see if it usable for various purposes such as SSL client, | |
9851 | server or S/MIME and CAs of these types. This is currently | |
9852 | VERY EXPERIMENTAL but will ultimately be used for certificate chain | |
9853 | verification. Also added a -purpose flag to x509 utility to | |
9854 | print out all the purposes. | |
9855 | [Steve Henson] | |
9856 | ||
56a3fec1 DSH |
9857 | *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated |
9858 | functions. | |
9859 | [Steve Henson] | |
9860 | ||
4654ef98 DSH |
9861 | *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search |
9862 | for, obtain and decode and extension and obtain its critical flag. | |
9863 | This allows all the necessary extension code to be handled in a | |
9864 | single function call. | |
9865 | [Steve Henson] | |
9866 | ||
7e102e28 AP |
9867 | *) RC4 tune-up featuring 30-40% performance improvement on most RISC |
9868 | platforms. See crypto/rc4/rc4_enc.c for further details. | |
9869 | [Andy Polyakov] | |
9870 | ||
d71c6bc5 DSH |
9871 | *) New -noout option to asn1parse. This causes no output to be produced |
9872 | its main use is when combined with -strparse and -out to extract data | |
9873 | from a file (which may not be in ASN.1 format). | |
9874 | [Steve Henson] | |
9875 | ||
2d681b77 DSH |
9876 | *) Fix for pkcs12 program. It was hashing an invalid certificate pointer |
9877 | when producing the local key id. | |
9878 | [Richard Levitte <levitte@stacken.kth.se>] | |
9879 | ||
3908cdf4 DSH |
9880 | *) New option -dhparam in s_server. This allows a DH parameter file to be |
9881 | stated explicitly. If it is not stated then it tries the first server | |
9882 | certificate file. The previous behaviour hard coded the filename | |
9883 | "server.pem". | |
9884 | [Steve Henson] | |
9885 | ||
3ea23631 DSH |
9886 | *) Add -pubin and -pubout options to the rsa and dsa commands. These allow |
9887 | a public key to be input or output. For example: | |
9888 | openssl rsa -in key.pem -pubout -out pubkey.pem | |
9889 | Also added necessary DSA public key functions to handle this. | |
9890 | [Steve Henson] | |
9891 | ||
393f2c65 DSH |
9892 | *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained |
9893 | in the message. This was handled by allowing | |
9894 | X509_find_by_issuer_and_serial() to tolerate a NULL passed to it. | |
9895 | [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>] | |
9896 | ||
9897 | *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null | |
9898 | to the end of the strings whereas this didn't. This would cause problems | |
9899 | if strings read with d2i_ASN1_bytes() were later modified. | |
9900 | [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>] | |
9901 | ||
4579dd5d DSH |
9902 | *) Fix for base64 decode bug. When a base64 bio reads only one line of |
9903 | data and it contains EOF it will end up returning an error. This is | |
9904 | caused by input 46 bytes long. The cause is due to the way base64 | |
9905 | BIOs find the start of base64 encoded data. They do this by trying a | |
9906 | trial decode on each line until they find one that works. When they | |
9907 | do a flag is set and it starts again knowing it can pass all the | |
9908 | data directly through the decoder. Unfortunately it doesn't reset | |
9909 | the context it uses. This means that if EOF is reached an attempt | |
9910 | is made to pass two EOFs through the context and this causes the | |
9911 | resulting error. This can also cause other problems as well. As is | |
9912 | usual with these problems it takes *ages* to find and the fix is | |
9913 | trivial: move one line. | |
9914 | [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ] | |
9915 | ||
06f4536a DSH |
9916 | *) Ugly workaround to get s_client and s_server working under Windows. The |
9917 | old code wouldn't work because it needed to select() on sockets and the | |
9918 | tty (for keypresses and to see if data could be written). Win32 only | |
9919 | supports select() on sockets so we select() with a 1s timeout on the | |
9920 | sockets and then see if any characters are waiting to be read, if none | |
9921 | are present then we retry, we also assume we can always write data to | |
9922 | the tty. This isn't nice because the code then blocks until we've | |
9923 | received a complete line of data and it is effectively polling the | |
9924 | keyboard at 1s intervals: however it's quite a bit better than not | |
9925 | working at all :-) A dedicated Windows application might handle this | |
9926 | with an event loop for example. | |
9927 | [Steve Henson] | |
9928 | ||
1c80019a DSH |
9929 | *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign |
9930 | and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions | |
9931 | will be called when RSA_sign() and RSA_verify() are used. This is useful | |
9932 | if rsa_pub_dec() and rsa_priv_enc() equivalents are not available. | |
9933 | For this to work properly RSA_public_decrypt() and RSA_private_encrypt() | |
9934 | should *not* be used: RSA_sign() and RSA_verify() must be used instead. | |
9935 | This necessitated the support of an extra signature type NID_md5_sha1 | |
9936 | for SSL signatures and modifications to the SSL library to use it instead | |
9937 | of calling RSA_public_decrypt() and RSA_private_encrypt(). | |
9938 | [Steve Henson] | |
9939 | ||
090d848e DSH |
9940 | *) Add new -verify -CAfile and -CApath options to the crl program, these |
9941 | will lookup a CRL issuers certificate and verify the signature in a | |
9942 | similar way to the verify program. Tidy up the crl program so it | |
0f583f69 | 9943 | no longer accesses structures directly. Make the ASN1 CRL parsing a bit |
090d848e DSH |
9944 | less strict. It will now permit CRL extensions even if it is not |
9945 | a V2 CRL: this will allow it to tolerate some broken CRLs. | |
9946 | [Steve Henson] | |
9947 | ||
396f6314 BM |
9948 | *) Initialize all non-automatic variables each time one of the openssl |
9949 | sub-programs is started (this is necessary as they may be started | |
9950 | multiple times from the "OpenSSL>" prompt). | |
9951 | [Lennart Bang, Bodo Moeller] | |
9952 | ||
4a61a64f DSH |
9953 | *) Preliminary compilation option RSA_NULL which disables RSA crypto without |
9954 | removing all other RSA functionality (this is what NO_RSA does). This | |
9955 | is so (for example) those in the US can disable those operations covered | |
9956 | by the RSA patent while allowing storage and parsing of RSA keys and RSA | |
9957 | key generation. | |
9958 | [Steve Henson] | |
9959 | ||
c1082a90 | 9960 | *) Non-copying interface to BIO pairs. |
6f7af152 | 9961 | (still largely untested) |
c1082a90 BM |
9962 | [Bodo Moeller] |
9963 | ||
a785abc3 DSH |
9964 | *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive |
9965 | ASCII string. This was handled independently in various places before. | |
9966 | [Steve Henson] | |
9967 | ||
aef838fc DSH |
9968 | *) New functions UTF8_getc() and UTF8_putc() that parse and generate |
9969 | UTF8 strings a character at a time. | |
9970 | [Steve Henson] | |
9971 | ||
074309b7 BM |
9972 | *) Use client_version from client hello to select the protocol |
9973 | (s23_srvr.c) and for RSA client key exchange verification | |
9974 | (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications. | |
9975 | [Bodo Moeller] | |
9976 | ||
8ce97163 DSH |
9977 | *) Add various utility functions to handle SPKACs, these were previously |
9978 | handled by poking round in the structure internals. Added new function | |
9979 | NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to | |
9980 | print, verify and generate SPKACs. Based on an original idea from | |
9981 | Massimiliano Pala <madwolf@comune.modena.it> but extensively modified. | |
9982 | [Steve Henson] | |
9983 | ||
2d4287da AP |
9984 | *) RIPEMD160 is operational on all platforms and is back in 'make test'. |
9985 | [Andy Polyakov] | |
9986 | ||
87a25f90 DSH |
9987 | *) Allow the config file extension section to be overwritten on the |
9988 | command line. Based on an original idea from Massimiliano Pala | |
9989 | <madwolf@comune.modena.it>. The new option is called -extensions | |
9990 | and can be applied to ca, req and x509. Also -reqexts to override | |
9991 | the request extensions in req and -crlexts to override the crl extensions | |
9992 | in ca. | |
9993 | [Steve Henson] | |
9994 | ||
f9150e54 DSH |
9995 | *) Add new feature to the SPKAC handling in ca. Now you can include |
9996 | the same field multiple times by preceding it by "XXXX." for example: | |
9997 | 1.OU="Unit name 1" | |
9998 | 2.OU="Unit name 2" | |
9999 | this is the same syntax as used in the req config file. | |
10000 | [Steve Henson] | |
10001 | ||
c79b16e1 DSH |
10002 | *) Allow certificate extensions to be added to certificate requests. These |
10003 | are specified in a 'req_extensions' option of the req section of the | |
10004 | config file. They can be printed out with the -text option to req but | |
10005 | are otherwise ignored at present. | |
10006 | [Steve Henson] | |
10007 | ||
96c2201b | 10008 | *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first |
0f7e6fe1 | 10009 | data read consists of only the final block it would not decrypted because |
7b65c329 DSH |
10010 | EVP_CipherUpdate() would correctly report zero bytes had been decrypted. |
10011 | A misplaced 'break' also meant the decrypted final block might not be | |
10012 | copied until the next read. | |
10013 | [Steve Henson] | |
10014 | ||
13066cee DSH |
10015 | *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added |
10016 | a few extra parameters to the DH structure: these will be useful if | |
10017 | for example we want the value of 'q' or implement X9.42 DH. | |
10018 | [Steve Henson] | |
10019 | ||
c0711f7f DSH |
10020 | *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and |
10021 | provides hooks that allow the default DSA functions or functions on a | |
10022 | "per key" basis to be replaced. This allows hardware acceleration and | |
10023 | hardware key storage to be handled without major modification to the | |
10024 | library. Also added low level modexp hooks and CRYPTO_EX structure and | |
10025 | associated functions. | |
10026 | [Steve Henson] | |
10027 | ||
8484721a DSH |
10028 | *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO |
10029 | as "read only": it can't be written to and the buffer it points to will | |
10030 | not be freed. Reading from a read only BIO is much more efficient than | |
10031 | a normal memory BIO. This was added because there are several times when | |
10032 | an area of memory needs to be read from a BIO. The previous method was | |
10033 | to create a memory BIO and write the data to it, this results in two | |
10034 | copies of the data and an O(n^2) reading algorithm. There is a new | |
10035 | function BIO_new_mem_buf() which creates a read only memory BIO from | |
10036 | an area of memory. Also modified the PKCS#7 routines to use read only | |
0f583f69 | 10037 | memory BIOs. |
8484721a DSH |
10038 | [Steve Henson] |
10039 | ||
de1915e4 BM |
10040 | *) Bugfix: ssl23_get_client_hello did not work properly when called in |
10041 | state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of | |
10042 | a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read, | |
10043 | but a retry condition occured while trying to read the rest. | |
10044 | [Bodo Moeller] | |
10045 | ||
c6c34506 DSH |
10046 | *) The PKCS7_ENC_CONTENT_new() function was setting the content type as |
10047 | NID_pkcs7_encrypted by default: this was wrong since this should almost | |
10048 | always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle | |
10049 | the encrypted data type: this is a more sensible place to put it and it | |
10050 | allows the PKCS#12 code to be tidied up that duplicated this | |
10051 | functionality. | |
10052 | [Steve Henson] | |
10053 | ||
fd520577 DSH |
10054 | *) Changed obj_dat.pl script so it takes its input and output files on |
10055 | the command line. This should avoid shell escape redirection problems | |
10056 | under Win32. | |
10057 | [Steve Henson] | |
10058 | ||
87c49f62 | 10059 | *) Initial support for certificate extension requests, these are included |
fd520577 DSH |
10060 | in things like Xenroll certificate requests. Included functions to allow |
10061 | extensions to be obtained and added. | |
87c49f62 DSH |
10062 | [Steve Henson] |
10063 | ||
1b1a6e78 BM |
10064 | *) -crlf option to s_client and s_server for sending newlines as |
10065 | CRLF (as required by many protocols). | |
10066 | [Bodo Moeller] | |
10067 | ||
9a577e29 | 10068 | Changes between 0.9.3a and 0.9.4 [09 Aug 1999] |
dfbaf956 | 10069 | |
9a577e29 | 10070 | *) Install libRSAglue.a when OpenSSL is built with RSAref. |
dfbaf956 | 10071 | [Ralf S. Engelschall] |
74678cc2 | 10072 | |
96395158 RE |
10073 | *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency. |
10074 | [Andrija Antonijevic <TheAntony2@bigfoot.com>] | |
10075 | ||
ed7f60fb DSH |
10076 | *) Fix -startdate and -enddate (which was missing) arguments to 'ca' |
10077 | program. | |
10078 | [Steve Henson] | |
10079 | ||
48c843c3 BM |
10080 | *) New function DSA_dup_DH, which duplicates DSA parameters/keys as |
10081 | DH parameters/keys (q is lost during that conversion, but the resulting | |
10082 | DH parameters contain its length). | |
10083 | ||
10084 | For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is | |
10085 | much faster than DH_generate_parameters (which creates parameters | |
10086 | where p = 2*q + 1), and also the smaller q makes DH computations | |
10087 | much more efficient (160-bit exponentiation instead of 1024-bit | |
10088 | exponentiation); so this provides a convenient way to support DHE | |
10089 | ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of | |
10090 | utter importance to use | |
10091 | SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); | |
10092 | or | |
10093 | SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE); | |
10094 | when such DH parameters are used, because otherwise small subgroup | |
10095 | attacks may become possible! | |
10096 | [Bodo Moeller] | |
10097 | ||
10098 | *) Avoid memory leak in i2d_DHparams. | |
10099 | [Bodo Moeller] | |
10100 | ||
922180d7 DSH |
10101 | *) Allow the -k option to be used more than once in the enc program: |
10102 | this allows the same encrypted message to be read by multiple recipients. | |
10103 | [Steve Henson] | |
10104 | ||
3e3d2ea2 DSH |
10105 | *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts |
10106 | an ASN1_OBJECT to a text string. If the "no_name" parameter is set then | |
10107 | it will always use the numerical form of the OID, even if it has a short | |
10108 | or long name. | |
10109 | [Steve Henson] | |
10110 | ||
770d19b8 DSH |
10111 | *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp |
10112 | method only got called if p,q,dmp1,dmq1,iqmp components were present, | |
10113 | otherwise bn_mod_exp was called. In the case of hardware keys for example | |
10114 | no private key components need be present and it might store extra data | |
96c2201b BM |
10115 | in the RSA structure, which cannot be accessed from bn_mod_exp. |
10116 | By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for | |
10117 | private key operations. | |
770d19b8 DSH |
10118 | [Steve Henson] |
10119 | ||
a0618e3e AP |
10120 | *) Added support for SPARC Linux. |
10121 | [Andy Polyakov] | |
10122 | ||
74678cc2 BM |
10123 | *) pem_password_cb function type incompatibly changed from |
10124 | typedef int pem_password_cb(char *buf, int size, int rwflag); | |
10125 | to | |
10126 | ....(char *buf, int size, int rwflag, void *userdata); | |
10127 | so that applications can pass data to their callbacks: | |
10128 | The PEM[_ASN1]_{read,write}... functions and macros now take an | |
10129 | additional void * argument, which is just handed through whenever | |
10130 | the password callback is called. | |
96c2201b | 10131 | [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller] |
74678cc2 BM |
10132 | |
10133 | New function SSL_CTX_set_default_passwd_cb_userdata. | |
10134 | ||
10135 | Compatibility note: As many C implementations push function arguments | |
10136 | onto the stack in reverse order, the new library version is likely to | |
10137 | interoperate with programs that have been compiled with the old | |
10138 | pem_password_cb definition (PEM_whatever takes some data that | |
10139 | happens to be on the stack as its last argument, and the callback | |
10140 | just ignores this garbage); but there is no guarantee whatsoever that | |
10141 | this will work. | |
0cceb1c7 | 10142 | |
664b9985 BM |
10143 | *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=... |
10144 | (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused | |
10145 | problems not only on Windows, but also on some Unix platforms. | |
2e0fc875 | 10146 | To avoid problematic command lines, these definitions are now in an |
57119943 BM |
10147 | auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl |
10148 | for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds). | |
664b9985 BM |
10149 | [Bodo Moeller] |
10150 | ||
7363455f AP |
10151 | *) MIPS III/IV assembler module is reimplemented. |
10152 | [Andy Polyakov] | |
10153 | ||
6434450c UM |
10154 | *) More DES library cleanups: remove references to srand/rand and |
10155 | delete an unused file. | |
9f0b86c6 | 10156 | [Ulf Möller] |
6434450c | 10157 | |
b617a5be DSH |
10158 | *) Add support for the the free Netwide assembler (NASM) under Win32, |
10159 | since not many people have MASM (ml) and it can be hard to obtain. | |
10160 | This is currently experimental but it seems to work OK and pass all | |
10161 | the tests. Check out INSTALL.W32 for info. | |
10162 | [Steve Henson] | |
10163 | ||
50596582 BM |
10164 | *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections |
10165 | without temporary keys kept an extra copy of the server key, | |
10166 | and connections with temporary keys did not free everything in case | |
10167 | of an error. | |
10168 | [Bodo Moeller] | |
10169 | ||
03cd4944 BM |
10170 | *) New function RSA_check_key and new openssl rsa option -check |
10171 | for verifying the consistency of RSA keys. | |
10172 | [Ulf Moeller, Bodo Moeller] | |
10173 | ||
f598cd13 DSH |
10174 | *) Various changes to make Win32 compile work: |
10175 | 1. Casts to avoid "loss of data" warnings in p5_crpt2.c | |
10176 | 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned | |
10177 | comparison" warnings. | |
10178 | 3. Add sk_<TYPE>_sort to DEF file generator and do make update. | |
b617a5be | 10179 | [Steve Henson] |
f598cd13 | 10180 | |
f513939e DSH |
10181 | *) Add a debugging option to PKCS#5 v2 key generation function: when |
10182 | you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and | |
10183 | derived keys are printed to stderr. | |
10184 | [Steve Henson] | |
10185 | ||
0ab8beb4 DSH |
10186 | *) Copy the flags in ASN1_STRING_dup(). |
10187 | [Roman E. Pavlov <pre@mo.msk.ru>] | |
10188 | ||
f7daafa4 DSH |
10189 | *) The x509 application mishandled signing requests containing DSA |
10190 | keys when the signing key was also DSA and the parameters didn't match. | |
10191 | ||
10192 | It was supposed to omit the parameters when they matched the signing key: | |
10193 | the verifying software was then supposed to automatically use the CA's | |
10194 | parameters if they were absent from the end user certificate. | |
10195 | ||
10196 | Omitting parameters is no longer recommended. The test was also | |
10197 | the wrong way round! This was probably due to unusual behaviour in | |
10198 | EVP_cmp_parameters() which returns 1 if the parameters match. | |
10199 | This meant that parameters were omitted when they *didn't* match and | |
10200 | the certificate was useless. Certificates signed with 'ca' didn't have | |
10201 | this bug. | |
10202 | [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>] | |
10203 | ||
458cddc1 BM |
10204 | *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems. |
10205 | The interface is as follows: | |
777ab7e6 BM |
10206 | Applications can use |
10207 | CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(), | |
10208 | CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop(); | |
10209 | "off" is now the default. | |
10210 | The library internally uses | |
10211 | CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(), | |
10212 | CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on() | |
10213 | to disable memory-checking temporarily. | |
10214 | ||
10215 | Some inconsistent states that previously were possible (and were | |
10216 | even the default) are now avoided. | |
458cddc1 BM |
10217 | |
10218 | -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time | |
10219 | with each memory chunk allocated; this is occasionally more helpful | |
10220 | than just having a counter. | |
e391116a BM |
10221 | |
10222 | -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID. | |
10223 | ||
10224 | -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future | |
10225 | extensions. | |
777ab7e6 BM |
10226 | [Bodo Moeller] |
10227 | ||
e1056435 BM |
10228 | *) Introduce "mode" for SSL structures (with defaults in SSL_CTX), |
10229 | which largely parallels "options", but is for changing API behaviour, | |
10230 | whereas "options" are about protocol behaviour. | |
9c962484 | 10231 | Initial "mode" flags are: |
e1056435 BM |
10232 | |
10233 | SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when | |
10234 | a single record has been written. | |
10235 | SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write | |
10236 | retries use the same buffer location. | |
10237 | (But all of the contents must be | |
10238 | copied!) | |
10239 | [Bodo Moeller] | |
10240 | ||
4b49bf6a | 10241 | *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options |
e1056435 BM |
10242 | worked. |
10243 | ||
5271ebd9 | 10244 | *) Fix problems with no-hmac etc. |
9f0b86c6 | 10245 | [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>] |
5271ebd9 | 10246 | |
ce8b2574 DSH |
10247 | *) New functions RSA_get_default_method(), RSA_set_method() and |
10248 | RSA_get_method(). These allows replacement of RSA_METHODs without having | |
10249 | to mess around with the internals of an RSA structure. | |
10250 | [Steve Henson] | |
10251 | ||
9c729e0a BM |
10252 | *) Fix memory leaks in DSA_do_sign and DSA_is_prime. |
10253 | Also really enable memory leak checks in openssl.c and in some | |
10254 | test programs. | |
10255 | [Chad C. Mulligan, Bodo Moeller] | |
10256 | ||
034292ad DSH |
10257 | *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess |
10258 | up the length of negative integers. This has now been simplified to just | |
10259 | store the length when it is first determined and use it later, rather | |
10260 | than trying to keep track of where data is copied and updating it to | |
10261 | point to the end. | |
10262 | [Steve Henson, reported by Brien Wheeler | |
10263 | <bwheeler@authentica-security.com>] | |
10264 | ||
170afce5 DSH |
10265 | *) Add a new function PKCS7_signatureVerify. This allows the verification |
10266 | of a PKCS#7 signature but with the signing certificate passed to the | |
10267 | function itself. This contrasts with PKCS7_dataVerify which assumes the | |
10268 | certificate is present in the PKCS#7 structure. This isn't always the | |
10269 | case: certificates can be omitted from a PKCS#7 structure and be | |
10270 | distributed by "out of band" means (such as a certificate database). | |
10271 | [Steve Henson] | |
10272 | ||
dbd665c2 DSH |
10273 | *) Complete the PEM_* macros with DECLARE_PEM versions to replace the |
10274 | function prototypes in pem.h, also change util/mkdef.pl to add the | |
10275 | necessary function names. | |
10276 | [Steve Henson] | |
10277 | ||
f76a8084 | 10278 | *) mk1mf.pl (used by Windows builds) did not properly read the |
6888f2b3 | 10279 | options set by Configure in the top level Makefile, and Configure |
975d3dc2 | 10280 | was not even able to write more than one option correctly. |
6888f2b3 | 10281 | Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended. |
f76a8084 BM |
10282 | [Bodo Moeller] |
10283 | ||
8623f693 DSH |
10284 | *) New functions CONF_load_bio() and CONF_load_fp() to allow a config |
10285 | file to be loaded from a BIO or FILE pointer. The BIO version will | |
10286 | for example allow memory BIOs to contain config info. | |
10287 | [Steve Henson] | |
10288 | ||
a111306b BM |
10289 | *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS. |
10290 | Whoever hopes to achieve shared-library compatibility across versions | |
10291 | must use this, not the compile-time macro. | |
11af1a27 BM |
10292 | (Exercise 0.9.4: Which is the minimum library version required by |
10293 | such programs?) | |
10294 | Note: All this applies only to multi-threaded programs, others don't | |
10295 | need locks. | |
a111306b BM |
10296 | [Bodo Moeller] |
10297 | ||
95d29597 BM |
10298 | *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests |
10299 | through a BIO pair triggered the default case, i.e. | |
10300 | SSLerr(...,SSL_R_UNKNOWN_STATE). | |
10301 | [Bodo Moeller] | |
10302 | ||
10303 | *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications | |
10304 | can use the SSL library even if none of the specific BIOs is | |
10305 | appropriate. | |
10306 | [Bodo Moeller] | |
10307 | ||
9bce3070 DSH |
10308 | *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value |
10309 | for the encoded length. | |
10310 | [Jeon KyoungHo <khjeon@sds.samsung.co.kr>] | |
10311 | ||
565d1065 DSH |
10312 | *) Add initial documentation of the X509V3 functions. |
10313 | [Steve Henson] | |
10314 | ||
b7d135b3 DSH |
10315 | *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and |
10316 | PEM_write_bio_PKCS8PrivateKey() that are equivalent to | |
10317 | PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more | |
10318 | secure PKCS#8 private key format with a high iteration count. | |
10319 | [Steve Henson] | |
10320 | ||
9d9b559e RE |
10321 | *) Fix determination of Perl interpreter: A perl or perl5 |
10322 | _directory_ in $PATH was also accepted as the interpreter. | |
10323 | [Ralf S. Engelschall] | |
10324 | ||
5f6d0ea2 DSH |
10325 | *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking |
10326 | wrong with it but it was very old and did things like calling | |
10327 | PEM_ASN1_read() directly and used MD5 for the hash not to mention some | |
10328 | unusual formatting. | |
10329 | [Steve Henson] | |
10330 | ||
f62676b9 DSH |
10331 | *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed |
10332 | to use the new extension code. | |
10333 | [Steve Henson] | |
10334 | ||
10335 | *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c | |
10336 | with macros. This should make it easier to change their form, add extra | |
10337 | arguments etc. Fix a few PEM prototypes which didn't have cipher as a | |
10338 | constant. | |
10339 | [Steve Henson] | |
10340 | ||
8151f52a BM |
10341 | *) Add to configuration table a new entry that can specify an alternative |
10342 | name for unistd.h (for pre-POSIX systems); we need this for NeXTstep, | |
10343 | according to Mark Crispin <MRC@Panda.COM>. | |
10344 | [Bodo Moeller] | |
10345 | ||
c77f47ab | 10346 | #if 0 |
05861c77 BL |
10347 | *) DES CBC did not update the IV. Weird. |
10348 | [Ben Laurie] | |
c77f47ab | 10349 | #else |
a7bd0396 BM |
10350 | des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does. |
10351 | Changing the behaviour of the former might break existing programs -- | |
10352 | where IV updating is needed, des_ncbc_encrypt can be used. | |
c77f47ab | 10353 | #endif |
05861c77 | 10354 | |
233bf734 BL |
10355 | *) When bntest is run from "make test" it drives bc to check its |
10356 | calculations, as well as internally checking them. If an internal check | |
10357 | fails, it needs to cause bc to give a non-zero result or make test carries | |
10358 | on without noticing the failure. Fixed. | |
10359 | [Ben Laurie] | |
10360 | ||
908eb7b8 | 10361 | *) DES library cleanups. |
9f0b86c6 | 10362 | [Ulf Möller] |
908eb7b8 | 10363 | |
8eb57af5 DSH |
10364 | *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be |
10365 | used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit | |
10366 | ciphers. NOTE: although the key derivation function has been verified | |
10367 | against some published test vectors it has not been extensively tested | |
10368 | yet. Added a -v2 "cipher" option to pkcs8 application to allow the use | |
10369 | of v2.0. | |
10370 | [Steve Henson] | |
10371 | ||
d4443edc BM |
10372 | *) Instead of "mkdir -p", which is not fully portable, use new |
10373 | Perl script "util/mkdir-p.pl". | |
8151f52a | 10374 | [Bodo Moeller] |
d4443edc | 10375 | |
69cbf468 DSH |
10376 | *) Rewrite the way password based encryption (PBE) is handled. It used to |
10377 | assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter | |
10378 | structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms | |
10379 | but doesn't apply to PKCS#5 v2.0 where it can be something else. Now | |
10380 | the 'parameter' field of the AlgorithmIdentifier is passed to the | |
10381 | underlying key generation function so it must do its own ASN1 parsing. | |
10382 | This has also changed the EVP_PBE_CipherInit() function which now has a | |
10383 | 'parameter' argument instead of literal salt and iteration count values | |
10384 | and the function EVP_PBE_ALGOR_CipherInit() has been deleted. | |
10385 | [Steve Henson] | |
10386 | ||
ef8335d9 | 10387 | *) Support for PKCS#5 v1.5 compatible password based encryption algorithms |
e7871ffa DSH |
10388 | and PKCS#8 functionality. New 'pkcs8' application linked to openssl. |
10389 | Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE | |
10390 | KEY" because this clashed with PKCS#8 unencrypted string. Since this | |
10391 | value was just used as a "magic string" and not used directly its | |
10392 | value doesn't matter. | |
ef8335d9 DSH |
10393 | [Steve Henson] |
10394 | ||
84c15db5 BL |
10395 | *) Introduce some semblance of const correctness to BN. Shame C doesn't |
10396 | support mutable. | |
10397 | [Ben Laurie] | |
10398 | ||
272c9333 | 10399 | *) "linux-sparc64" configuration (ultrapenguin). |
885982dc | 10400 | [Ray Miller <ray.miller@oucs.ox.ac.uk>] |
272c9333 BM |
10401 | "linux-sparc" configuration. |
10402 | [Christian Forster <fo@hawo.stw.uni-erlangen.de>] | |
885982dc | 10403 | |
a53955d8 | 10404 | *) config now generates no-xxx options for missing ciphers. |
9f0b86c6 | 10405 | [Ulf Möller] |
a53955d8 UM |
10406 | |
10407 | *) Support the EBCDIC character set (work in progress). | |
10408 | File ebcdic.c not yet included because it has a different license. | |
10409 | [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] | |
10410 | ||
10411 | *) Support BS2000/OSD-POSIX. | |
10412 | [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>] | |
10413 | ||
b4f76582 BL |
10414 | *) Make callbacks for key generation use void * instead of char *. |
10415 | [Ben Laurie] | |
10416 | ||
213a75db BL |
10417 | *) Make S/MIME samples compile (not yet tested). |
10418 | [Ben Laurie] | |
10419 | ||
748365ee BM |
10420 | *) Additional typesafe stacks. |
10421 | [Ben Laurie] | |
10422 | ||
885982dc | 10423 | *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x). |
0cceb1c7 BM |
10424 | [Bodo Moeller] |
10425 | ||
748365ee | 10426 | |
31fab3e8 | 10427 | Changes between 0.9.3 and 0.9.3a [29 May 1999] |
472bde40 | 10428 | |
2e36cc41 BM |
10429 | *) New configuration variant "sco5-gcc". |
10430 | ||
71f08093 | 10431 | *) Updated some demos. |
054009a6 | 10432 | [Sean O Riordain, Wade Scholine] |
71f08093 | 10433 | |
e95f6268 BM |
10434 | *) Add missing BIO_free at exit of pkcs12 application. |
10435 | [Wu Zhigang] | |
10436 | ||
10437 | *) Fix memory leak in conf.c. | |
10438 | [Steve Henson] | |
10439 | ||
472bde40 BM |
10440 | *) Updates for Win32 to assembler version of MD5. |
10441 | [Steve Henson] | |
10442 | ||
10443 | *) Set #! path to perl in apps/der_chop to where we found it | |
10444 | instead of using a fixed path. | |
10445 | [Bodo Moeller] | |
10446 | ||
10447 | *) SHA library changes for irix64-mips4-cc. | |
10448 | [Andy Polyakov] | |
10449 | ||
10450 | *) Improvements for VMS support. | |
10451 | [Richard Levitte] | |
10452 | ||
748365ee | 10453 | |
557068c0 | 10454 | Changes between 0.9.2b and 0.9.3 [24 May 1999] |
7d7d2cbc | 10455 | |
e14d4443 UM |
10456 | *) Bignum library bug fix. IRIX 6 passes "make test" now! |
10457 | This also avoids the problems with SC4.2 and unpatched SC5. | |
10458 | [Andy Polyakov <appro@fy.chalmers.se>] | |
10459 | ||
e84240d4 DSH |
10460 | *) New functions sk_num, sk_value and sk_set to replace the previous macros. |
10461 | These are required because of the typesafe stack would otherwise break | |
10462 | existing code. If old code used a structure member which used to be STACK | |
10463 | and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with | |
10464 | sk_num or sk_value it would produce an error because the num, data members | |
10465 | are not present in STACK_OF. Now it just produces a warning. sk_set | |
10466 | replaces the old method of assigning a value to sk_value | |
10467 | (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code | |
10468 | that does this will no longer work (and should use sk_set instead) but | |
10469 | this could be regarded as a "questionable" behaviour anyway. | |
10470 | [Steve Henson] | |
10471 | ||
1b266dab DSH |
10472 | *) Fix most of the other PKCS#7 bugs. The "experimental" code can now |
10473 | correctly handle encrypted S/MIME data. | |
10474 | [Steve Henson] | |
10475 | ||
55519bbb | 10476 | *) Change type of various DES function arguments from des_cblock |
f43c8149 | 10477 | (which means, in function argument declarations, pointer to char) |
55519bbb BM |
10478 | to des_cblock * (meaning pointer to array with 8 char elements), |
10479 | which allows the compiler to do more typechecking; it was like | |
10480 | that back in SSLeay, but with lots of ugly casts. | |
10481 | ||
10482 | Introduce new type const_des_cblock. | |
10483 | [Bodo Moeller] | |
10484 | ||
84fa704c DSH |
10485 | *) Reorganise the PKCS#7 library and get rid of some of the more obvious |
10486 | problems: find RecipientInfo structure that matches recipient certificate | |
10487 | and initialise the ASN1 structures properly based on passed cipher. | |
10488 | [Steve Henson] | |
10489 | ||
62bad771 BL |
10490 | *) Belatedly make the BN tests actually check the results. |
10491 | [Ben Laurie] | |
10492 | ||
1ad2ecb6 DSH |
10493 | *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion |
10494 | to and from BNs: it was completely broken. New compilation option | |
10495 | NEG_PUBKEY_BUG to allow for some broken certificates that encode public | |
10496 | key elements as negative integers. | |
10497 | [Steve Henson] | |
10498 | ||
bd3576d2 UM |
10499 | *) Reorganize and speed up MD5. |
10500 | [Andy Polyakov <appro@fy.chalmers.se>] | |
10501 | ||
7d7d2cbc UM |
10502 | *) VMS support. |
10503 | [Richard Levitte <richard@levitte.org>] | |
1b276f30 | 10504 | |
f5eac85e DSH |
10505 | *) New option -out to asn1parse to allow the parsed structure to be |
10506 | output to a file. This is most useful when combined with the -strparse | |
10507 | option to examine the output of things like OCTET STRINGS. | |
10508 | [Steve Henson] | |
10509 | ||
b31b04d9 BM |
10510 | *) Make SSL library a little more fool-proof by not requiring any longer |
10511 | that SSL_set_{accept,connect}_state be called before | |
10512 | SSL_{accept,connect} may be used (SSL_set_..._state is omitted | |
10513 | in many applications because usually everything *appeared* to work as | |
10514 | intended anyway -- now it really works as intended). | |
10515 | [Bodo Moeller] | |
10516 | ||
d5a2ea4b | 10517 | *) Move openssl.cnf out of lib/. |
9f0b86c6 | 10518 | [Ulf Möller] |
d5a2ea4b | 10519 | |
397f7038 RE |
10520 | *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall |
10521 | -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes | |
10522 | -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ | |
10523 | [Ralf S. Engelschall] | |
10524 | ||
884e8ec6 DSH |
10525 | *) Various fixes to the EVP and PKCS#7 code. It may now be able to |
10526 | handle PKCS#7 enveloped data properly. | |
10527 | [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve] | |
10528 | ||
ca8e5b9b BM |
10529 | *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of |
10530 | copying pointers. The cert_st handling is changed by this in | |
10531 | various ways (and thus what used to be known as ctx->default_cert | |
10532 | is now called ctx->cert, since we don't resort to s->ctx->[default_]cert | |
10533 | any longer when s->cert does not give us what we need). | |
10534 | ssl_cert_instantiate becomes obsolete by this change. | |
10535 | As soon as we've got the new code right (possibly it already is?), | |
10536 | we have solved a couple of bugs of the earlier code where s->cert | |
10537 | was used as if it could not have been shared with other SSL structures. | |
10538 | ||
10539 | Note that using the SSL API in certain dirty ways now will result | |
10540 | in different behaviour than observed with earlier library versions: | |
10541 | Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx) | |
10542 | does not influence s as it used to. | |
10543 | ||
ca8e5b9b | 10544 | In order to clean up things more thoroughly, inside SSL_SESSION |
b56bce4f BM |
10545 | we don't use CERT any longer, but a new structure SESS_CERT |
10546 | that holds per-session data (if available); currently, this is | |
10547 | the peer's certificate chain and, for clients, the server's certificate | |
10548 | and temporary key. CERT holds only those values that can have | |
10549 | meaningful defaults in an SSL_CTX. | |
ca8e5b9b BM |
10550 | [Bodo Moeller] |
10551 | ||
c8b41850 DSH |
10552 | *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure |
10553 | from the internal representation. Various PKCS#7 fixes: remove some | |
10554 | evil casts and set the enc_dig_alg field properly based on the signing | |
10555 | key type. | |
10556 | [Steve Henson] | |
10557 | ||
e40b7abe DSH |
10558 | *) Allow PKCS#12 password to be set from the command line or the |
10559 | environment. Let 'ca' get its config file name from the environment | |
10560 | variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req' | |
10561 | and 'x509'). | |
10562 | [Steve Henson] | |
10563 | ||
10564 | *) Allow certificate policies extension to use an IA5STRING for the | |
10565 | organization field. This is contrary to the PKIX definition but | |
10566 | VeriSign uses it and IE5 only recognises this form. Document 'x509' | |
10567 | extension option. | |
10568 | [Steve Henson] | |
10569 | ||
5b640028 BL |
10570 | *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic, |
10571 | without disallowing inline assembler and the like for non-pedantic builds. | |
10572 | [Ben Laurie] | |
10573 | ||
31a674d8 | 10574 | *) Support Borland C++ builder. |
9f0b86c6 | 10575 | [Janez Jere <jj@void.si>, modified by Ulf Möller] |
31a674d8 UM |
10576 | |
10577 | *) Support Mingw32. | |
9f0b86c6 | 10578 | [Ulf Möller] |
31a674d8 | 10579 | |
8e7f966b UM |
10580 | *) SHA-1 cleanups and performance enhancements. |
10581 | [Andy Polyakov <appro@fy.chalmers.se>] | |
10582 | ||
4f5fac80 | 10583 | *) Sparc v8plus assembler for the bignum library. |
8e7f966b | 10584 | [Andy Polyakov <appro@fy.chalmers.se>] |
4f5fac80 | 10585 | |
afd1f9e8 | 10586 | *) Accept any -xxx and +xxx compiler options in Configure. |
9f0b86c6 | 10587 | [Ulf Möller] |
afd1f9e8 UM |
10588 | |
10589 | *) Update HPUX configuration. | |
10590 | [Anonymous] | |
10591 | ||
dee75ecf RE |
10592 | *) Add missing sk_<type>_unshift() function to safestack.h |
10593 | [Ralf S. Engelschall] | |
10594 | ||
b3ca645f BM |
10595 | *) New function SSL_CTX_use_certificate_chain_file that sets the |
10596 | "extra_cert"s in addition to the certificate. (This makes sense | |
10597 | only for "PEM" format files, as chains as a whole are not | |
10598 | DER-encoded.) | |
10599 | [Bodo Moeller] | |
10600 | ||
7f89714e BM |
10601 | *) Support verify_depth from the SSL API. |
10602 | x509_vfy.c had what can be considered an off-by-one-error: | |
10603 | Its depth (which was not part of the external interface) | |
10604 | was actually counting the number of certificates in a chain; | |
10605 | now it really counts the depth. | |
10606 | [Bodo Moeller] | |
10607 | ||
dc1f607a BM |
10608 | *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used |
10609 | instead of X509err, which often resulted in confusing error | |
10610 | messages since the error codes are not globally unique | |
10611 | (e.g. an alleged error in ssl3_accept when a certificate | |
10612 | didn't match the private key). | |
10613 | ||
4eb77b26 | 10614 | *) New function SSL_CTX_set_session_id_context that allows to set a default |
dd1462fd BM |
10615 | value (so that you don't need SSL_set_session_id_context for each |
10616 | connection using the SSL_CTX). | |
4eb77b26 BM |
10617 | [Bodo Moeller] |
10618 | ||
c6652749 | 10619 | *) OAEP decoding bug fix. |
9f0b86c6 | 10620 | [Ulf Möller] |
c6652749 | 10621 | |
e5f3045f BM |
10622 | *) Support INSTALL_PREFIX for package builders, as proposed by |
10623 | David Harris. | |
10624 | [Bodo Moeller] | |
10625 | ||
87bc2c00 BM |
10626 | *) New Configure options "threads" and "no-threads". For systems |
10627 | where the proper compiler options are known (currently Solaris | |
10628 | and Linux), "threads" is the default. | |
10629 | [Bodo Moeller] | |
10630 | ||
6e6acfd4 BM |
10631 | *) New script util/mklink.pl as a faster substitute for util/mklink.sh. |
10632 | [Bodo Moeller] | |
10633 | ||
ddeee82c BM |
10634 | *) Install various scripts to $(OPENSSLDIR)/misc, not to |
10635 | $(INSTALLTOP)/bin -- they shouldn't clutter directories | |
10636 | such as /usr/local/bin. | |
10637 | [Bodo Moeller] | |
10638 | ||
0973910f | 10639 | *) "make linux-shared" to build shared libraries. |
ddeee82c | 10640 | [Niels Poppe <niels@netbox.org>] |
0973910f | 10641 | |
f5d7a031 | 10642 | *) New Configure option no-<cipher> (rsa, idea, rc5, ...). |
9f0b86c6 | 10643 | [Ulf Möller] |
f5d7a031 | 10644 | |
b64f8256 DSH |
10645 | *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for |
10646 | extension adding in x509 utility. | |
10647 | [Steve Henson] | |
10648 | ||
a9be3af5 | 10649 | *) Remove NOPROTO sections and error code comments. |
9f0b86c6 | 10650 | [Ulf Möller] |
a9be3af5 | 10651 | |
47339f61 DSH |
10652 | *) Partial rewrite of the DEF file generator to now parse the ANSI |
10653 | prototypes. | |
10654 | [Steve Henson] | |
10655 | ||
b0b7b1c5 | 10656 | *) New Configure options --prefix=DIR and --openssldir=DIR. |
9f0b86c6 | 10657 | [Ulf Möller] |
b0b7b1c5 | 10658 | |
6d311938 DSH |
10659 | *) Complete rewrite of the error code script(s). It is all now handled |
10660 | by one script at the top level which handles error code gathering, | |
10661 | header rewriting and C source file generation. It should be much better | |
10662 | than the old method: it now uses a modified version of Ulf's parser to | |
10663 | read the ANSI prototypes in all header files (thus the old K&R definitions | |
10664 | aren't needed for error creation any more) and do a better job of | |
10665 | translating function codes into names. The old 'ASN1 error code imbedded | |
10666 | in a comment' is no longer necessary and it doesn't use .err files which | |
6e781e8e DSH |
10667 | have now been deleted. Also the error code call doesn't have to appear all |
10668 | on one line (which resulted in some large lines...). | |
6d311938 DSH |
10669 | [Steve Henson] |
10670 | ||
018b4ee9 | 10671 | *) Change #include filenames from <foo.h> to <openssl/foo.h>. |
92df9607 BM |
10672 | [Bodo Moeller] |
10673 | ||
85f48f7e BM |
10674 | *) Change behaviour of ssl2_read when facing length-0 packets: Don't return |
10675 | 0 (which usually indicates a closed connection), but continue reading. | |
10676 | [Bodo Moeller] | |
10677 | ||
90b8bbb8 BM |
10678 | *) Fix some race conditions. |
10679 | [Bodo Moeller] | |
10680 | ||
d943e372 DSH |
10681 | *) Add support for CRL distribution points extension. Add Certificate |
10682 | Policies and CRL distribution points documentation. | |
10683 | [Steve Henson] | |
10684 | ||
8e10f2b3 | 10685 | *) Move the autogenerated header file parts to crypto/opensslconf.h. |
9f0b86c6 | 10686 | [Ulf Möller] |
8e10f2b3 | 10687 | |
4997138a BL |
10688 | *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of |
10689 | 8 of keying material. Merlin has also confirmed interop with this fix | |
10690 | between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0. | |
10691 | [Merlin Hughes <merlin@baltimore.ie>] | |
10692 | ||
95dc05bc UM |
10693 | *) Fix lots of warnings. |
10694 | [Richard Levitte <levitte@stacken.kth.se>] | |
10695 | ||
10696 | *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if | |
10697 | the directory spec didn't end with a LIST_SEPARATOR_CHAR. | |
4997138a | 10698 | [Richard Levitte <levitte@stacken.kth.se>] |
95dc05bc | 10699 | |
8fb04b98 UM |
10700 | *) Fix problems with sizeof(long) == 8. |
10701 | [Andy Polyakov <appro@fy.chalmers.se>] | |
10702 | ||
6b691a5c | 10703 | *) Change functions to ANSI C. |
9f0b86c6 | 10704 | [Ulf Möller] |
6b691a5c | 10705 | |
df82f5c8 | 10706 | *) Fix typos in error codes. |
9f0b86c6 | 10707 | [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller] |
df82f5c8 | 10708 | |
22a4f969 | 10709 | *) Remove defunct assembler files from Configure. |
9f0b86c6 | 10710 | [Ulf Möller] |
22a4f969 | 10711 | |
5e85b6ab UM |
10712 | *) SPARC v8 assembler BIGNUM implementation. |
10713 | [Andy Polyakov <appro@fy.chalmers.se>] | |
10714 | ||
3edd7ed1 | 10715 | *) Support for Certificate Policies extension: both print and set. |
d943e372 | 10716 | Various additions to support the r2i method this uses. |
41b731f2 DSH |
10717 | [Steve Henson] |
10718 | ||
e778802f BL |
10719 | *) A lot of constification, and fix a bug in X509_NAME_oneline() that could |
10720 | return a const string when you are expecting an allocated buffer. | |
10721 | [Ben Laurie] | |
10722 | ||
c83e523d DSH |
10723 | *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE |
10724 | types DirectoryString and DisplayText. | |
d77b3054 DSH |
10725 | [Steve Henson] |
10726 | ||
1d48dd00 DSH |
10727 | *) Add code to allow r2i extensions to access the configuration database, |
10728 | add an LHASH database driver and add several ctx helper functions. | |
10729 | [Steve Henson] | |
10730 | ||
953937bd DSH |
10731 | *) Fix an evil bug in bn_expand2() which caused various BN functions to |
10732 | fail when they extended the size of a BIGNUM. | |
10733 | [Steve Henson] | |
10734 | ||
28a98809 DSH |
10735 | *) Various utility functions to handle SXNet extension. Modify mkdef.pl to |
10736 | support typesafe stack. | |
10737 | [Steve Henson] | |
10738 | ||
8f7de4f0 BL |
10739 | *) Fix typo in SSL_[gs]et_options(). |
10740 | [Nils Frostberg <nils@medcom.se>] | |
10741 | ||
0490a86d DSH |
10742 | *) Delete various functions and files that belonged to the (now obsolete) |
10743 | old X509V3 handling code. | |
10744 | [Steve Henson] | |
10745 | ||
5fbe91d8 | 10746 | *) New Configure option "rsaref". |
9f0b86c6 | 10747 | [Ulf Möller] |
5fbe91d8 | 10748 | |
5fd4e2b1 BM |
10749 | *) Don't auto-generate pem.h. |
10750 | [Bodo Moeller] | |
10751 | ||
f73e07cf BL |
10752 | *) Introduce type-safe ASN.1 SETs. |
10753 | [Ben Laurie] | |
10754 | ||
9263e882 | 10755 | *) Convert various additional casted stacks to type-safe STACK_OF() variants. |
135a1dca | 10756 | [Ben Laurie, Ralf S. Engelschall, Steve Henson] |
9263e882 | 10757 | |
f73e07cf BL |
10758 | *) Introduce type-safe STACKs. This will almost certainly break lots of code |
10759 | that links with OpenSSL (well at least cause lots of warnings), but fear | |
10760 | not: the conversion is trivial, and it eliminates loads of evil casts. A | |
10761 | few STACKed things have been converted already. Feel free to convert more. | |
10762 | In the fullness of time, I'll do away with the STACK type altogether. | |
10763 | [Ben Laurie] | |
10764 | ||
f9a25931 RE |
10765 | *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate |
10766 | specified in <certfile> by updating the entry in the index.txt file. | |
10767 | This way one no longer has to edit the index.txt file manually for | |
10768 | revoking a certificate. The -revoke option does the gory details now. | |
10769 | [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall] | |
10770 | ||
2f0cd195 RE |
10771 | *) Fix `openssl crl -noout -text' combination where `-noout' killed the |
10772 | `-text' option at all and this way the `-noout -text' combination was | |
10773 | inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'. | |
10774 | [Ralf S. Engelschall] | |
10775 | ||
268c2102 RE |
10776 | *) Make sure a corresponding plain text error message exists for the |
10777 | X509_V_ERR_CERT_REVOKED/23 error number which can occur when a | |
10778 | verify callback function determined that a certificate was revoked. | |
10779 | [Ralf S. Engelschall] | |
10780 | ||
fc8ee06b BM |
10781 | *) Bugfix: In test/testenc, don't test "openssl <cipher>" for |
10782 | ciphers that were excluded, e.g. by -DNO_IDEA. Also, test | |
10783 | all available cipers including rc5, which was forgotten until now. | |
10784 | In order to let the testing shell script know which algorithms | |
10785 | are available, a new (up to now undocumented) command | |
10786 | "openssl list-cipher-commands" is used. | |
10787 | [Bodo Moeller] | |
10788 | ||
c7ac31e2 BM |
10789 | *) Bugfix: s_client occasionally would sleep in select() when |
10790 | it should have checked SSL_pending() first. | |
10791 | [Bodo Moeller] | |
10792 | ||
9d892e28 UM |
10793 | *) New functions DSA_do_sign and DSA_do_verify to provide access to |
10794 | the raw DSA values prior to ASN.1 encoding. | |
9f0b86c6 | 10795 | [Ulf Möller] |
9d892e28 UM |
10796 | |
10797 | *) Tweaks to Configure | |
748365ee | 10798 | [Niels Poppe <niels@netbox.org>] |
9d892e28 | 10799 | |
d2e26dcc DSH |
10800 | *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support, |
10801 | yet... | |
10802 | [Steve Henson] | |
10803 | ||
99aab161 | 10804 | *) New variables $(RANLIB) and $(PERL) in the Makefiles. |
9f0b86c6 | 10805 | [Ulf Möller] |
99aab161 | 10806 | |
2613c1fa UM |
10807 | *) New config option to avoid instructions that are illegal on the 80386. |
10808 | The default code is faster, but requires at least a 486. | |
9f0b86c6 | 10809 | [Ulf Möller] |
2613c1fa | 10810 | |
6d02d8e4 BM |
10811 | *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and |
10812 | SSL2_SERVER_VERSION (not used at all) macros, which are now the | |
10813 | same as SSL2_VERSION anyway. | |
10814 | [Bodo Moeller] | |
10815 | ||
10816 | *) New "-showcerts" option for s_client. | |
10817 | [Bodo Moeller] | |
10818 | ||
ee0508d4 DSH |
10819 | *) Still more PKCS#12 integration. Add pkcs12 application to openssl |
10820 | application. Various cleanups and fixes. | |
10821 | [Steve Henson] | |
10822 | ||
8d8c7266 DSH |
10823 | *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and |
10824 | modify error routines to work internally. Add error codes and PBE init | |
10825 | to library startup routines. | |
10826 | [Steve Henson] | |
10827 | ||
cfcefcbe DSH |
10828 | *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and |
10829 | packing functions to asn1 and evp. Changed function names and error | |
10830 | codes along the way. | |
10831 | [Steve Henson] | |
10832 | ||
4b518c26 DSH |
10833 | *) PKCS12 integration: and so it begins... First of several patches to |
10834 | slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12 | |
cfcefcbe | 10835 | objects to objects.h |
4b518c26 DSH |
10836 | [Steve Henson] |
10837 | ||
785cdf20 DSH |
10838 | *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1 |
10839 | and display support for Thawte strong extranet extension. | |
10840 | [Steve Henson] | |
10841 | ||
ba423add BL |
10842 | *) Add LinuxPPC support. |
10843 | [Jeff Dubrule <igor@pobox.org>] | |
10844 | ||
67da3df7 BL |
10845 | *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to |
10846 | bn_div_words in alpha.s. | |
10847 | [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie] | |
10848 | ||
0e9fc711 RE |
10849 | *) Make sure the RSA OAEP test is skipped under -DRSAref because |
10850 | OAEP isn't supported when OpenSSL is built with RSAref. | |
10851 | [Ulf Moeller <ulf@fitug.de>] | |
10852 | ||
1b276f30 RE |
10853 | *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h |
10854 | so they no longer are missing under -DNOPROTO. | |
10855 | [Soren S. Jorvang <soren@t.dk>] | |
10856 | ||
1b24cca9 BM |
10857 | |
10858 | Changes between 0.9.1c and 0.9.2b [22 Mar 1999] | |
4f43d0e7 | 10859 | |
b4cadc6e BL |
10860 | *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still |
10861 | doesn't work when the session is reused. Coming soon! | |
10862 | [Ben Laurie] | |
10863 | ||
10864 | *) Fix a security hole, that allows sessions to be reused in the wrong | |
10865 | context thus bypassing client cert protection! All software that uses | |
10866 | client certs and session caches in multiple contexts NEEDS PATCHING to | |
10867 | allow session reuse! A fuller solution is in the works. | |
10868 | [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)] | |
10869 | ||
afb23063 RE |
10870 | *) Some more source tree cleanups (removed obsolete files |
10871 | crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed | |
10872 | permission on "config" script to be executable) and a fix for the INSTALL | |
10873 | document. | |
10874 | [Ulf Moeller <ulf@fitug.de>] | |
10875 | ||
199d59e5 DSH |
10876 | *) Remove some legacy and erroneous uses of malloc, free instead of |
10877 | Malloc, Free. | |
10878 | [Lennart Bang <lob@netstream.se>, with minor changes by Steve] | |
10879 | ||
b4899bb1 BL |
10880 | *) Make rsa_oaep_test return non-zero on error. |
10881 | [Ulf Moeller <ulf@fitug.de>] | |
10882 | ||
29c0fccb BL |
10883 | *) Add support for native Solaris shared libraries. Configure |
10884 | solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice | |
10885 | if someone would make that last step automatic. | |
10886 | [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>] | |
10887 | ||
cadf126b BL |
10888 | *) ctx_size was not built with the right compiler during "make links". Fixed. |
10889 | [Ben Laurie] | |
10890 | ||
bc420ac5 DSH |
10891 | *) Change the meaning of 'ALL' in the cipher list. It now means "everything |
10892 | except NULL ciphers". This means the default cipher list will no longer | |
10893 | enable NULL ciphers. They need to be specifically enabled e.g. with | |
10894 | the string "DEFAULT:eNULL". | |
10895 | [Steve Henson] | |
10896 | ||
abd4c915 DSH |
10897 | *) Fix to RSA private encryption routines: if p < q then it would |
10898 | occasionally produce an invalid result. This will only happen with | |
10899 | externally generated keys because OpenSSL (and SSLeay) ensure p > q. | |
10900 | [Steve Henson] | |
10901 | ||
7e37e72a RE |
10902 | *) Be less restrictive and allow also `perl util/perlpath.pl |
10903 | /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin', | |
10904 | because this way one can also use an interpreter named `perl5' (which is | |
10905 | usually the name of Perl 5.xxx on platforms where an Perl 4.x is still | |
10906 | installed as `perl'). | |
10907 | [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] | |
10908 | ||
637691e6 RE |
10909 | *) Let util/clean-depend.pl work also with older Perl 5.00x versions. |
10910 | [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] | |
10911 | ||
83ec54b4 DSH |
10912 | *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add |
10913 | advapi32.lib to Win32 build and change the pem test comparision | |
10914 | to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the | |
38138020 DSH |
10915 | suggestion). Fix misplaced ASNI prototypes and declarations in evp.h |
10916 | and crypto/des/ede_cbcm_enc.c. | |
10917 | [Steve Henson] | |
83ec54b4 | 10918 | |
b241fefd BL |
10919 | *) DES quad checksum was broken on big-endian architectures. Fixed. |
10920 | [Ben Laurie] | |
10921 | ||
d4d2f98c DSH |
10922 | *) Comment out two functions in bio.h that aren't implemented. Fix up the |
10923 | Win32 test batch file so it (might) work again. The Win32 test batch file | |
10924 | is horrible: I feel ill.... | |
10925 | [Steve Henson] | |
10926 | ||
0cc39579 DSH |
10927 | *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected |
10928 | in e_os.h. Audit of header files to check ANSI and non ANSI | |
10929 | sections: 10 functions were absent from non ANSI section and not exported | |
10930 | from Windows DLLs. Fixed up libeay.num for new functions. | |
d4d2f98c | 10931 | [Steve Henson] |
0cc39579 | 10932 | |
d10f052b RE |
10933 | *) Make `openssl version' output lines consistent. |
10934 | [Ralf S. Engelschall] | |
10935 | ||
c0e538e1 RE |
10936 | *) Fix Win32 symbol export lists for BIO functions: Added |
10937 | BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data | |
10938 | to ms/libeay{16,32}.def. | |
10939 | [Ralf S. Engelschall] | |
10940 | ||
84107e6c RE |
10941 | *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled |
10942 | fine under Unix and passes some trivial tests I've now added. But the | |
10943 | whole stuff is horribly incomplete, so a README.1ST with a disclaimer was | |
10944 | added to make sure no one expects that this stuff really works in the | |
10945 | OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources | |
10946 | up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and | |
10947 | openssl_bio.xs. | |
10948 | [Ralf S. Engelschall] | |
10949 | ||
26a0846f BL |
10950 | *) Fix the generation of two part addresses in perl. |
10951 | [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie] | |
10952 | ||
7d3ce7ba BL |
10953 | *) Add config entry for Linux on MIPS. |
10954 | [John Tobey <jtobey@channel1.com>] | |
10955 | ||
efadf60f | 10956 | *) Make links whenever Configure is run, unless we are on Windoze. |
cba5068d BL |
10957 | [Ben Laurie] |
10958 | ||
1756d405 DSH |
10959 | *) Permit extensions to be added to CRLs using crl_section in openssl.cnf. |
10960 | Currently only issuerAltName and AuthorityKeyIdentifier make any sense | |
10961 | in CRLs. | |
d4d2f98c | 10962 | [Steve Henson] |
1756d405 | 10963 | |
116e3153 RE |
10964 | *) Add a useful kludge to allow package maintainers to specify compiler and |
10965 | other platforms details on the command line without having to patch the | |
10966 | Configure script everytime: One now can use ``perl Configure | |
10967 | <id>:<details>'', i.e. platform ids are allowed to have details appended | |
10968 | to them (seperated by colons). This is treated as there would be a static | |
10969 | pre-configured entry in Configure's %table under key <id> with value | |
10970 | <details> and ``perl Configure <id>'' is called. So, when you want to | |
10971 | perform a quick test-compile under FreeBSD 3.1 with pgcc and without | |
10972 | assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"'' | |
10973 | now, which overrides the FreeBSD-elf entry on-the-fly. | |
10974 | [Ralf S. Engelschall] | |
10975 | ||
bc348244 BL |
10976 | *) Disable new TLS1 ciphersuites by default: they aren't official yet. |
10977 | [Ben Laurie] | |
10978 | ||
3eb0ed6d RE |
10979 | *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified |
10980 | on the `perl Configure ...' command line. This way one can compile | |
10981 | OpenSSL libraries with Position Independent Code (PIC) which is needed | |
10982 | for linking it into DSOs. | |
10983 | [Ralf S. Engelschall] | |
10984 | ||
f415fa32 BL |
10985 | *) Remarkably, export ciphers were totally broken and no-one had noticed! |
10986 | Fixed. | |
10987 | [Ben Laurie] | |
10988 | ||
0b903ec0 RE |
10989 | *) Cleaned up the LICENSE document: The official contact for any license |
10990 | questions now is the OpenSSL core team under openssl-core@openssl.org. | |
10991 | And add a paragraph about the dual-license situation to make sure people | |
10992 | recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply | |
10993 | to the OpenSSL toolkit. | |
10994 | [Ralf S. Engelschall] | |
10995 | ||
bb8f3c58 RE |
10996 | *) General source tree makefile cleanups: Made `making xxx in yyy...' |
10997 | display consistent in the source tree and replaced `/bin/rm' by `rm'. | |
10998 | Additonally cleaned up the `make links' target: Remove unnecessary | |
10999 | semicolons, subsequent redundant removes, inline point.sh into mklink.sh | |
11000 | to speed processing and no longer clutter the display with confusing | |
11001 | stuff. Instead only the actually done links are displayed. | |
11002 | [Ralf S. Engelschall] | |
11003 | ||
988788f6 BL |
11004 | *) Permit null encryption ciphersuites, used for authentication only. It used |
11005 | to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this. | |
11006 | It is now necessary to set SSL_FORBID_ENULL to prevent the use of null | |
11007 | encryption. | |
11008 | [Ben Laurie] | |
11009 | ||
924acc54 DSH |
11010 | *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder |
11011 | signed attributes when verifying signatures (this would break them), | |
11012 | the detached data encoding was wrong and public keys obtained using | |
11013 | X509_get_pubkey() weren't freed. | |
11014 | [Steve Henson] | |
11015 | ||
d00b7aad DSH |
11016 | *) Add text documentation for the BUFFER functions. Also added a work around |
11017 | to a Win95 console bug. This was triggered by the password read stuff: the | |
11018 | last character typed gets carried over to the next fread(). If you were | |
11019 | generating a new cert request using 'req' for example then the last | |
11020 | character of the passphrase would be CR which would then enter the first | |
11021 | field as blank. | |
9985bed3 DSH |
11022 | [Steve Henson] |
11023 | ||
789285aa RE |
11024 | *) Added the new `Includes OpenSSL Cryptography Software' button as |
11025 | doc/openssl_button.{gif,html} which is similar in style to the old SSLeay | |
11026 | button and can be used by applications based on OpenSSL to show the | |
11027 | relationship to the OpenSSL project. | |
11028 | [Ralf S. Engelschall] | |
11029 | ||
a06c602e RE |
11030 | *) Remove confusing variables in function signatures in files |
11031 | ssl/ssl_lib.c and ssl/ssl.h. | |
11032 | [Lennart Bong <lob@kulthea.stacken.kth.se>] | |
11033 | ||
8d697db1 RE |
11034 | *) Don't install bss_file.c under PREFIX/include/ |
11035 | [Lennart Bong <lob@kulthea.stacken.kth.se>] | |
11036 | ||
06c68491 DSH |
11037 | *) Get the Win32 compile working again. Modify mkdef.pl so it can handle |
11038 | functions that return function pointers and has support for NT specific | |
11039 | stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various | |
11040 | #ifdef WIN32 and WINNTs sprinkled about the place and some changes from | |
11041 | unsigned to signed types: this was killing the Win32 compile. | |
11042 | [Steve Henson] | |
11043 | ||
72e442a3 RE |
11044 | *) Add new certificate file to stack functions, |
11045 | SSL_add_dir_cert_subjects_to_stack() and | |
11046 | SSL_add_file_cert_subjects_to_stack(). These largely supplant | |
11047 | SSL_load_client_CA_file(), and can be used to add multiple certs easily | |
11048 | to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()). | |
eb90a483 BL |
11049 | This means that Apache-SSL and similar packages don't have to mess around |
11050 | to add as many CAs as they want to the preferred list. | |
11051 | [Ben Laurie] | |
11052 | ||
4f43d0e7 BL |
11053 | *) Experiment with doxygen documentation. Currently only partially applied to |
11054 | ssl/ssl_lib.c. | |
11055 | See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with | |
11056 | openssl.doxy as the configuration file. | |
11057 | [Ben Laurie] | |
74d7abc2 RE |
11058 | |
11059 | *) Get rid of remaining C++-style comments which strict C compilers hate. | |
11060 | [Ralf S. Engelschall, pointed out by Carlos Amengual] | |
0172f988 | 11061 | |
7283ecea DSH |
11062 | *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not |
11063 | compiled in by default: it has problems with large keys. | |
11064 | [Steve Henson] | |
11065 | ||
15d21c2d RE |
11066 | *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and |
11067 | DH private keys and/or callback functions which directly correspond to | |
11068 | their SSL_CTX_xxx() counterparts but work on a per-connection basis. This | |
11069 | is needed for applications which have to configure certificates on a | |
11070 | per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis | |
11071 | (e.g. s_server). | |
11072 | For the RSA certificate situation is makes no difference, but | |
11073 | for the DSA certificate situation this fixes the "no shared cipher" | |
11074 | problem where the OpenSSL cipher selection procedure failed because the | |
11075 | temporary keys were not overtaken from the context and the API provided | |
11076 | no way to reconfigure them. | |
11077 | The new functions now let applications reconfigure the stuff and they | |
11078 | are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh, | |
11079 | SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new | |
11080 | non-public-API function ssl_cert_instantiate() is used as a helper | |
11081 | function and also to reduce code redundancy inside ssl_rsa.c. | |
11082 | [Ralf S. Engelschall] | |
11083 | ||
ea14a91f RE |
11084 | *) Move s_server -dcert and -dkey options out of the undocumented feature |
11085 | area because they are useful for the DSA situation and should be | |
11086 | recognized by the users. | |
11087 | [Ralf S. Engelschall] | |
11088 | ||
90a52cec RE |
11089 | *) Fix the cipher decision scheme for export ciphers: the export bits are |
11090 | *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within | |
11091 | SSL_EXP_MASK. So, the original variable has to be used instead of the | |
11092 | already masked variable. | |
11093 | [Richard Levitte <levitte@stacken.kth.se>] | |
11094 | ||
def9f431 RE |
11095 | *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c |
11096 | [Richard Levitte <levitte@stacken.kth.se>] | |
11097 | ||
8aef252b RE |
11098 | *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal() |
11099 | from `int' to `unsigned int' because it's a length and initialized by | |
11100 | EVP_DigestFinal() which expects an `unsigned int *'. | |
11101 | [Richard Levitte <levitte@stacken.kth.se>] | |
11102 | ||
a4ed5532 RE |
11103 | *) Don't hard-code path to Perl interpreter on shebang line of Configure |
11104 | script. Instead use the usual Shell->Perl transition trick. | |
11105 | [Ralf S. Engelschall] | |
11106 | ||
7be304ac RE |
11107 | *) Make `openssl x509 -noout -modulus' functional also for DSA certificates |
11108 | (in addition to RSA certificates) to match the behaviour of `openssl dsa | |
11109 | -noout -modulus' as it's already the case for `openssl rsa -noout | |
11110 | -modulus'. For RSA the -modulus is the real "modulus" while for DSA | |
11111 | currently the public key is printed (a decision which was already done by | |
11112 | `openssl dsa -modulus' in the past) which serves a similar purpose. | |
11113 | Additionally the NO_RSA no longer completely removes the whole -modulus | |
11114 | option; it now only avoids using the RSA stuff. Same applies to NO_DSA | |
11115 | now, too. | |
11116 | [Ralf S. Engelschall] | |
11117 | ||
55ab3bf7 BL |
11118 | *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested |
11119 | BIO. See the source (crypto/evp/bio_ok.c) for more info. | |
11120 | [Arne Ansper <arne@ats.cyber.ee>] | |
11121 | ||
a43aa73e DSH |
11122 | *) Dump the old yucky req code that tried (and failed) to allow raw OIDs |
11123 | to be added. Now both 'req' and 'ca' can use new objects defined in the | |
11124 | config file. | |
11125 | [Steve Henson] | |
11126 | ||
0849d138 BL |
11127 | *) Add cool BIO that does syslog (or event log on NT). |
11128 | [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie] | |
11129 | ||
06ab81f9 BL |
11130 | *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5, |
11131 | TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and | |
11132 | TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher | |
11133 | Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt. | |
11134 | [Ben Laurie] | |
11135 | ||
deff75b6 DSH |
11136 | *) Add preliminary config info for new extension code. |
11137 | [Steve Henson] | |
11138 | ||
0c8a1281 DSH |
11139 | *) Make RSA_NO_PADDING really use no padding. |
11140 | [Ulf Moeller <ulf@fitug.de>] | |
11141 | ||
4004dbb7 BL |
11142 | *) Generate errors when private/public key check is done. |
11143 | [Ben Laurie] | |
11144 | ||
0ca5f8b1 DSH |
11145 | *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support |
11146 | for some CRL extensions and new objects added. | |
11147 | [Steve Henson] | |
11148 | ||
3d8accc3 DSH |
11149 | *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private |
11150 | key usage extension and fuller support for authority key id. | |
11151 | [Steve Henson] | |
11152 | ||
a4949896 BL |
11153 | *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved |
11154 | padding method for RSA, which is recommended for new applications in PKCS | |
11155 | #1 v2.0 (RFC 2437, October 1998). | |
11156 | OAEP (Optimal Asymmetric Encryption Padding) has better theoretical | |
11157 | foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure | |
11158 | against Bleichbacher's attack on RSA. | |
11159 | [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by | |
11160 | Ben Laurie] | |
11161 | ||
413c4f45 MC |
11162 | *) Updates to the new SSL compression code |
11163 | [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] | |
11164 | ||
11165 | *) Fix so that the version number in the master secret, when passed | |
11166 | via RSA, checks that if TLS was proposed, but we roll back to SSLv3 | |
11167 | (because the server will not accept higher), that the version number | |
11168 | is 0x03,0x01, not 0x03,0x00 | |
11169 | [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] | |
11170 | ||
a8236c8c DSH |
11171 | *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory |
11172 | leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes | |
3d8accc3 | 11173 | in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c |
a8236c8c DSH |
11174 | [Steve Henson] |
11175 | ||
388ff0b0 DSH |
11176 | *) Support for RAW extensions where an arbitrary extension can be |
11177 | created by including its DER encoding. See apps/openssl.cnf for | |
11178 | an example. | |
a8236c8c | 11179 | [Steve Henson] |
388ff0b0 | 11180 | |
6013fa83 RE |
11181 | *) Make sure latest Perl versions don't interpret some generated C array |
11182 | code as Perl array code in the crypto/err/err_genc.pl script. | |
11183 | [Lars Weber <3weber@informatik.uni-hamburg.de>] | |
11184 | ||
5c00879e DSH |
11185 | *) Modify ms/do_ms.bat to not generate assembly language makefiles since |
11186 | not many people have the assembler. Various Win32 compilation fixes and | |
11187 | update to the INSTALL.W32 file with (hopefully) more accurate Win32 | |
11188 | build instructions. | |
11189 | [Steve Henson] | |
11190 | ||
9becf666 DSH |
11191 | *) Modify configure script 'Configure' to automatically create crypto/date.h |
11192 | file under Win32 and also build pem.h from pem.org. New script | |
11193 | util/mkfiles.pl to create the MINFO file on environments that can't do a | |
11194 | 'make files': perl util/mkfiles.pl >MINFO should work. | |
11195 | [Steve Henson] | |
11196 | ||
4e31df2c BL |
11197 | *) Major rework of DES function declarations, in the pursuit of correctness |
11198 | and purity. As a result, many evil casts evaporated, and some weirdness, | |
11199 | too. You may find this causes warnings in your code. Zapping your evil | |
11200 | casts will probably fix them. Mostly. | |
11201 | [Ben Laurie] | |
11202 | ||
e4119b93 DSH |
11203 | *) Fix for a typo in asn1.h. Bug fix to object creation script |
11204 | obj_dat.pl. It considered a zero in an object definition to mean | |
11205 | "end of object": none of the objects in objects.h have any zeros | |
11206 | so it wasn't spotted. | |
11207 | [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>] | |
11208 | ||
4a71b90d BL |
11209 | *) Add support for Triple DES Cipher Block Chaining with Output Feedback |
11210 | Masking (CBCM). In the absence of test vectors, the best I have been able | |
11211 | to do is check that the decrypt undoes the encrypt, so far. Send me test | |
11212 | vectors if you have them. | |
11213 | [Ben Laurie] | |
11214 | ||
2c6ccde1 | 11215 | *) Correct calculation of key length for export ciphers (too much space was |
436d318c BL |
11216 | allocated for null ciphers). This has not been tested! |
11217 | [Ben Laurie] | |
11218 | ||
55a9cc6e DSH |
11219 | *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage |
11220 | message is now correct (it understands "crypto" and "ssl" on its | |
11221 | command line). There is also now an "update" option. This will update | |
11222 | the util/ssleay.num and util/libeay.num files with any new functions. | |
11223 | If you do a: | |
11224 | perl util/mkdef.pl crypto ssl update | |
11225 | it will update them. | |
e4119b93 | 11226 | [Steve Henson] |
55a9cc6e | 11227 | |
8073036d RE |
11228 | *) Overhauled the Perl interface (perl/*): |
11229 | - ported BN stuff to OpenSSL's different BN library | |
11230 | - made the perl/ source tree CVS-aware | |
11231 | - renamed the package from SSLeay to OpenSSL (the files still contain | |
11232 | their history because I've copied them in the repository) | |
11233 | - removed obsolete files (the test scripts will be replaced | |
11234 | by better Test::Harness variants in the future) | |
11235 | [Ralf S. Engelschall] | |
11236 | ||
483fdf18 RE |
11237 | *) First cut for a very conservative source tree cleanup: |
11238 | 1. merge various obsolete readme texts into doc/ssleay.txt | |
11239 | where we collect the old documents and readme texts. | |
11240 | 2. remove the first part of files where I'm already sure that we no | |
11241 | longer need them because of three reasons: either they are just temporary | |
11242 | files which were left by Eric or they are preserved original files where | |
11243 | I've verified that the diff is also available in the CVS via "cvs diff | |
11244 | -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for | |
11245 | the crypto/md/ stuff). | |
11246 | [Ralf S. Engelschall] | |
11247 | ||
175b0942 DSH |
11248 | *) More extension code. Incomplete support for subject and issuer alt |
11249 | name, issuer and authority key id. Change the i2v function parameters | |
11250 | and add an extra 'crl' parameter in the X509V3_CTX structure: guess | |
11251 | what that's for :-) Fix to ASN1 macro which messed up | |
11252 | IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED. | |
11253 | [Steve Henson] | |
11254 | ||
bceacf93 DSH |
11255 | *) Preliminary support for ENUMERATED type. This is largely copied from the |
11256 | INTEGER code. | |
11257 | [Steve Henson] | |
11258 | ||
351d8998 MC |
11259 | *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy. |
11260 | [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] | |
11261 | ||
b621d772 RE |
11262 | *) Make sure `make rehash' target really finds the `openssl' program. |
11263 | [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>] | |
11264 | ||
a96e7810 BL |
11265 | *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd |
11266 | like to hear about it if this slows down other processors. | |
11267 | [Ben Laurie] | |
11268 | ||
e04a6c2b RE |
11269 | *) Add CygWin32 platform information to Configure script. |
11270 | [Alan Batie <batie@aahz.jf.intel.com>] | |
11271 | ||
0172f988 RE |
11272 | *) Fixed ms/32all.bat script: `no_asm' -> `no-asm' |
11273 | [Rainer W. Gerling <gerling@mpg-gv.mpg.de>] | |
79dfa975 DSH |
11274 | |
11275 | *) New program nseq to manipulate netscape certificate sequences | |
11276 | [Steve Henson] | |
320a14cb | 11277 | |
9fe84296 DSH |
11278 | *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a |
11279 | few typos. | |
11280 | [Steve Henson] | |
11281 | ||
a0a54079 MC |
11282 | *) Fixes to BN code. Previously the default was to define BN_RECURSION |
11283 | but the BN code had some problems that would cause failures when | |
11284 | doing certificate verification and some other functions. | |
11285 | [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)] | |
11286 | ||
92c046ca DSH |
11287 | *) Add ASN1 and PEM code to support netscape certificate sequences. |
11288 | [Steve Henson] | |
11289 | ||
79dfa975 DSH |
11290 | *) Add ASN1 and PEM code to support netscape certificate sequences. |
11291 | [Steve Henson] | |
11292 | ||
a27598bf DSH |
11293 | *) Add several PKIX and private extended key usage OIDs. |
11294 | [Steve Henson] | |
11295 | ||
b2347661 DSH |
11296 | *) Modify the 'ca' program to handle the new extension code. Modify |
11297 | openssl.cnf for new extension format, add comments. | |
11298 | [Steve Henson] | |
11299 | ||
f317aa4c DSH |
11300 | *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req' |
11301 | and add a sample to openssl.cnf so req -x509 now adds appropriate | |
11302 | CA extensions. | |
11303 | [Steve Henson] | |
11304 | ||
834eeef9 DSH |
11305 | *) Continued X509 V3 changes. Add to other makefiles, integrate with the |
11306 | error code, add initial support to X509_print() and x509 application. | |
f317aa4c | 11307 | [Steve Henson] |
834eeef9 | 11308 | |
9aeaf1b4 DSH |
11309 | *) Takes a deep breath and start addding X509 V3 extension support code. Add |
11310 | files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this | |
11311 | stuff is currently isolated and isn't even compiled yet. | |
11312 | [Steve Henson] | |
11313 | ||
9b5cc156 DSH |
11314 | *) Continuing patches for GeneralizedTime. Fix up certificate and CRL |
11315 | ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print. | |
11316 | Removed the versions check from X509 routines when loading extensions: | |
11317 | this allows certain broken certificates that don't set the version | |
11318 | properly to be processed. | |
11319 | [Steve Henson] | |
11320 | ||
8039257d BL |
11321 | *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another |
11322 | Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which | |
11323 | can still be regenerated with "make depend". | |
11324 | [Ben Laurie] | |
11325 | ||
b13a1554 BL |
11326 | *) Spelling mistake in C version of CAST-128. |
11327 | [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>] | |
11328 | ||
6c8abdd7 DSH |
11329 | *) Changes to the error generation code. The perl script err-code.pl |
11330 | now reads in the old error codes and retains the old numbers, only | |
11331 | adding new ones if necessary. It also only changes the .err files if new | |
11332 | codes are added. The makefiles have been modified to only insert errors | |
11333 | when needed (to avoid needlessly modifying header files). This is done | |
11334 | by only inserting errors if the .err file is newer than the auto generated | |
11335 | C file. To rebuild all the error codes from scratch (the old behaviour) | |
11336 | either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl | |
11337 | or delete all the .err files. | |
9b5cc156 | 11338 | [Steve Henson] |
6c8abdd7 | 11339 | |
649cdb7b BL |
11340 | *) CAST-128 was incorrectly implemented for short keys. The C version has |
11341 | been fixed, but is untested. The assembler versions are also fixed, but | |
11342 | new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing | |
11343 | to regenerate it if needed. | |
11344 | [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun | |
11345 | Hagino <itojun@kame.net>] | |
11346 | ||
11347 | *) File was opened incorrectly in randfile.c. | |
9f0b86c6 | 11348 | [Ulf Möller <ulf@fitug.de>] |
649cdb7b | 11349 | |
fdd3b642 DSH |
11350 | *) Beginning of support for GeneralizedTime. d2i, i2d, check and print |
11351 | functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or | |
11352 | GeneralizedTime. ASN1_TIME is the proper type used in certificates et | |
11353 | al: it's just almost always a UTCTime. Note this patch adds new error | |
11354 | codes so do a "make errors" if there are problems. | |
11355 | [Steve Henson] | |
11356 | ||
dabba110 | 11357 | *) Correct Linux 1 recognition in config. |
9f0b86c6 | 11358 | [Ulf Möller <ulf@fitug.de>] |
dabba110 | 11359 | |
512d2228 BL |
11360 | *) Remove pointless MD5 hash when using DSA keys in ca. |
11361 | [Anonymous <nobody@replay.com>] | |
11362 | ||
2c1ef383 BL |
11363 | *) Generate an error if given an empty string as a cert directory. Also |
11364 | generate an error if handed NULL (previously returned 0 to indicate an | |
11365 | error, but didn't set one). | |
11366 | [Ben Laurie, reported by Anonymous <nobody@replay.com>] | |
11367 | ||
c3ae9a48 BL |
11368 | *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last. |
11369 | [Ben Laurie] | |
11370 | ||
ee13f9b1 DSH |
11371 | *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct |
11372 | parameters. This was causing a warning which killed off the Win32 compile. | |
11373 | [Steve Henson] | |
11374 | ||
27eb622b DSH |
11375 | *) Remove C++ style comments from crypto/bn/bn_local.h. |
11376 | [Neil Costigan <neil.costigan@celocom.com>] | |
11377 | ||
2d723902 DSH |
11378 | *) The function OBJ_txt2nid was broken. It was supposed to return a nid |
11379 | based on a text string, looking up short and long names and finally | |
11380 | "dot" format. The "dot" format stuff didn't work. Added new function | |
11381 | OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote | |
11382 | OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the | |
11383 | OID is not part of the table. | |
11384 | [Steve Henson] | |
11385 | ||
a6801a91 BL |
11386 | *) Add prototypes to X509 lookup/verify methods, fixing a bug in |
11387 | X509_LOOKUP_by_alias(). | |
11388 | [Ben Laurie] | |
11389 | ||
50acf46b BL |
11390 | *) Sort openssl functions by name. |
11391 | [Ben Laurie] | |
11392 | ||
7f9b7b07 DSH |
11393 | *) Get the gendsa program working (hopefully) and add it to app list. Remove |
11394 | encryption from sample DSA keys (in case anyone is interested the password | |
11395 | was "1234"). | |
11396 | [Steve Henson] | |
11397 | ||
e03ddfae BL |
11398 | *) Make _all_ *_free functions accept a NULL pointer. |
11399 | [Frans Heymans <fheymans@isaserver.be>] | |
11400 | ||
6fa89f94 BL |
11401 | *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use |
11402 | NULL pointers. | |
11403 | [Anonymous <nobody@replay.com>] | |
11404 | ||
c13d4799 BL |
11405 | *) s_server should send the CAfile as acceptable CAs, not its own cert. |
11406 | [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] | |
11407 | ||
bc4deee0 BL |
11408 | *) Don't blow it for numeric -newkey arguments to apps/req. |
11409 | [Bodo Moeller <3moeller@informatik.uni-hamburg.de>] | |
11410 | ||
5b00115a BL |
11411 | *) Temp key "for export" tests were wrong in s3_srvr.c. |
11412 | [Anonymous <nobody@replay.com>] | |
11413 | ||
f8c3c05d BL |
11414 | *) Add prototype for temp key callback functions |
11415 | SSL_CTX_set_tmp_{rsa,dh}_callback(). | |
11416 | [Ben Laurie] | |
11417 | ||
ad65ce75 DSH |
11418 | *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and |
11419 | DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey(). | |
384c479c | 11420 | [Steve Henson] |
ad65ce75 | 11421 | |
e416ad97 BL |
11422 | *) X509_name_add_entry() freed the wrong thing after an error. |
11423 | [Arne Ansper <arne@ats.cyber.ee>] | |
11424 | ||
4a18cddd BL |
11425 | *) rsa_eay.c would attempt to free a NULL context. |
11426 | [Arne Ansper <arne@ats.cyber.ee>] | |
11427 | ||
bb65e20b BL |
11428 | *) BIO_s_socket() had a broken should_retry() on Windoze. |
11429 | [Arne Ansper <arne@ats.cyber.ee>] | |
11430 | ||
b5e406f7 BL |
11431 | *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH. |
11432 | [Arne Ansper <arne@ats.cyber.ee>] | |
11433 | ||
cb0f35d7 RE |
11434 | *) Make sure the already existing X509_STORE->depth variable is initialized |
11435 | in X509_STORE_new(), but document the fact that this variable is still | |
11436 | unused in the certificate verification process. | |
11437 | [Ralf S. Engelschall] | |
11438 | ||
cfcf6453 | 11439 | *) Fix the various library and apps files to free up pkeys obtained from |
ad65ce75 | 11440 | X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions. |
cfcf6453 DSH |
11441 | [Steve Henson] |
11442 | ||
cdbb8c2f BL |
11443 | *) Fix reference counting in X509_PUBKEY_get(). This makes |
11444 | demos/maurice/example2.c work, amongst others, probably. | |
11445 | [Steve Henson and Ben Laurie] | |
11446 | ||
06d5b162 RE |
11447 | *) First cut of a cleanup for apps/. First the `ssleay' program is now named |
11448 | `openssl' and second, the shortcut symlinks for the `openssl <command>' | |
11449 | are no longer created. This way we have a single and consistent command | |
11450 | line interface `openssl <command>', similar to `cvs <command>'. | |
cdbb8c2f | 11451 | [Ralf S. Engelschall, Paul Sutton and Ben Laurie] |
06d5b162 | 11452 | |
c35f549e DSH |
11453 | *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey |
11454 | BIT STRING wrapper always have zero unused bits. | |
11455 | [Steve Henson] | |
11456 | ||
ebc828ca DSH |
11457 | *) Add CA.pl, perl version of CA.sh, add extended key usage OID. |
11458 | [Steve Henson] | |
11459 | ||
79e259e3 PS |
11460 | *) Make the top-level INSTALL documentation easier to understand. |
11461 | [Paul Sutton] | |
11462 | ||
56ee3117 PS |
11463 | *) Makefiles updated to exit if an error occurs in a sub-directory |
11464 | make (including if user presses ^C) [Paul Sutton] | |
11465 | ||
6063b27b BL |
11466 | *) Make Montgomery context stuff explicit in RSA data structure. |
11467 | [Ben Laurie] | |
11468 | ||
11469 | *) Fix build order of pem and err to allow for generated pem.h. | |
11470 | [Ben Laurie] | |
11471 | ||
11472 | *) Fix renumbering bug in X509_NAME_delete_entry(). | |
11473 | [Ben Laurie] | |
11474 | ||
792a9002 | 11475 | *) Enhanced the err-ins.pl script so it makes the error library number |
11476 | global and can add a library name. This is needed for external ASN1 and | |
11477 | other error libraries. | |
11478 | [Steve Henson] | |
11479 | ||
11480 | *) Fixed sk_insert which never worked properly. | |
11481 | [Steve Henson] | |
11482 | ||
11483 | *) Fix ASN1 macros so they can handle indefinite length construted | |
11484 | EXPLICIT tags. Some non standard certificates use these: they can now | |
11485 | be read in. | |
11486 | [Steve Henson] | |
11487 | ||
ce72df1c RE |
11488 | *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc) |
11489 | into a single doc/ssleay.txt bundle. This way the information is still | |
11490 | preserved but no longer messes up this directory. Now it's new room for | |
11491 | the new set of documenation files. | |
11492 | [Ralf S. Engelschall] | |
11493 | ||
4098e89c BL |
11494 | *) SETs were incorrectly DER encoded. This was a major pain, because they |
11495 | shared code with SEQUENCEs, which aren't coded the same. This means that | |
11496 | almost everything to do with SETs or SEQUENCEs has either changed name or | |
11497 | number of arguments. | |
11498 | [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>] | |
11499 | ||
11500 | *) Fix test data to work with the above. | |
11501 | [Ben Laurie] | |
11502 | ||
03f8b042 BL |
11503 | *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but |
11504 | was already fixed by Eric for 0.9.1 it seems. | |
9f0b86c6 | 11505 | [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>] |
03f8b042 | 11506 | |
5dcdcd47 BL |
11507 | *) Autodetect FreeBSD3. |
11508 | [Ben Laurie] | |
11509 | ||
1641cb60 BL |
11510 | *) Fix various bugs in Configure. This affects the following platforms: |
11511 | nextstep | |
11512 | ncr-scde | |
11513 | unixware-2.0 | |
11514 | unixware-2.0-pentium | |
11515 | sco5-cc. | |
11516 | [Ben Laurie] | |
ae82b46f | 11517 | |
8d7ed6ff BL |
11518 | *) Eliminate generated files from CVS. Reorder tests to regenerate files |
11519 | before they are needed. | |
11520 | [Ben Laurie] | |
11521 | ||
11522 | *) Generate Makefile.ssl from Makefile.org (to keep CVS happy). | |
11523 | [Ben Laurie] | |
11524 | ||
1b24cca9 BM |
11525 | |
11526 | Changes between 0.9.1b and 0.9.1c [23-Dec-1998] | |
9ce5db45 | 11527 | |
f10a5c2a RE |
11528 | *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and |
11529 | changed SSLeay to OpenSSL in version strings. | |
9ce5db45 | 11530 | [Ralf S. Engelschall] |
9acc2aa6 RE |
11531 | |
11532 | *) Some fixups to the top-level documents. | |
11533 | [Paul Sutton] | |
651d0aff | 11534 | |
13e91dd3 RE |
11535 | *) Fixed the nasty bug where rsaref.h was not found under compile-time |
11536 | because the symlink to include/ was missing. | |
11537 | [Ralf S. Engelschall] | |
11538 | ||
11539 | *) Incorporated the popular no-RSA/DSA-only patches | |
11540 | which allow to compile a RSA-free SSLeay. | |
320a14cb | 11541 | [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall] |
13e91dd3 RE |
11542 | |
11543 | *) Fixed nasty rehash problem under `make -f Makefile.ssl links' | |
11544 | when "ssleay" is still not found. | |
11545 | [Ralf S. Engelschall] | |
11546 | ||
11547 | *) Added more platforms to Configure: Cray T3E, HPUX 11, | |
11548 | [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>] | |
11549 | ||
651d0aff RE |
11550 | *) Updated the README file. |
11551 | [Ralf S. Engelschall] | |
11552 | ||
11553 | *) Added various .cvsignore files in the CVS repository subdirs | |
11554 | to make a "cvs update" really silent. | |
11555 | [Ralf S. Engelschall] | |
11556 | ||
11557 | *) Recompiled the error-definition header files and added | |
11558 | missing symbols to the Win32 linker tables. | |
11559 | [Ralf S. Engelschall] | |
11560 | ||
11561 | *) Cleaned up the top-level documents; | |
11562 | o new files: CHANGES and LICENSE | |
11563 | o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay | |
11564 | o merged COPYRIGHT into LICENSE | |
11565 | o removed obsolete TODO file | |
11566 | o renamed MICROSOFT to INSTALL.W32 | |
11567 | [Ralf S. Engelschall] | |
11568 | ||
11569 | *) Removed dummy files from the 0.9.1b source tree: | |
11570 | crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi | |
11571 | crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f | |
11572 | crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f | |
11573 | crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f | |
11574 | util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f | |
11575 | [Ralf S. Engelschall] | |
11576 | ||
13e91dd3 | 11577 | *) Added various platform portability fixes. |
9acc2aa6 | 11578 | [Mark J. Cox] |
651d0aff | 11579 | |
f1c236f8 | 11580 | *) The Genesis of the OpenSSL rpject: |
651d0aff | 11581 | We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A. |
03f8b042 | 11582 | Young and Tim J. Hudson created while they were working for C2Net until |
651d0aff | 11583 | summer 1998. |
f1c236f8 | 11584 | [The OpenSSL Project] |
651d0aff | 11585 | |
1b24cca9 BM |
11586 | |
11587 | Changes between 0.9.0b and 0.9.1b [not released] | |
651d0aff RE |
11588 | |
11589 | *) Updated a few CA certificates under certs/ | |
11590 | [Eric A. Young] | |
11591 | ||
11592 | *) Changed some BIGNUM api stuff. | |
11593 | [Eric A. Young] | |
11594 | ||
11595 | *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, | |
11596 | DGUX x86, Linux Alpha, etc. | |
11597 | [Eric A. Young] | |
11598 | ||
11599 | *) New COMP library [crypto/comp/] for SSL Record Layer Compression: | |
11600 | RLE (dummy implemented) and ZLIB (really implemented when ZLIB is | |
11601 | available). | |
11602 | [Eric A. Young] | |
11603 | ||
11604 | *) Add -strparse option to asn1pars program which parses nested | |
11605 | binary structures | |
11606 | [Dr Stephen Henson <shenson@bigfoot.com>] | |
11607 | ||
11608 | *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs. | |
11609 | [Eric A. Young] | |
11610 | ||
11611 | *) DSA fix for "ca" program. | |
11612 | [Eric A. Young] | |
11613 | ||
11614 | *) Added "-genkey" option to "dsaparam" program. | |
11615 | [Eric A. Young] | |
11616 | ||
11617 | *) Added RIPE MD160 (rmd160) message digest. | |
11618 | [Eric A. Young] | |
11619 | ||
11620 | *) Added -a (all) option to "ssleay version" command. | |
11621 | [Eric A. Young] | |
11622 | ||
11623 | *) Added PLATFORM define which is the id given to Configure. | |
11624 | [Eric A. Young] | |
11625 | ||
11626 | *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking. | |
11627 | [Eric A. Young] | |
11628 | ||
11629 | *) Extended the ASN.1 parser routines. | |
11630 | [Eric A. Young] | |
11631 | ||
11632 | *) Extended BIO routines to support REUSEADDR, seek, tell, etc. | |
11633 | [Eric A. Young] | |
11634 | ||
11635 | *) Added a BN_CTX to the BN library. | |
11636 | [Eric A. Young] | |
11637 | ||
11638 | *) Fixed the weak key values in DES library | |
11639 | [Eric A. Young] | |
11640 | ||
11641 | *) Changed API in EVP library for cipher aliases. | |
11642 | [Eric A. Young] | |
11643 | ||
11644 | *) Added support for RC2/64bit cipher. | |
11645 | [Eric A. Young] | |
11646 | ||
11647 | *) Converted the lhash library to the crypto/mem.c functions. | |
11648 | [Eric A. Young] | |
11649 | ||
11650 | *) Added more recognized ASN.1 object ids. | |
11651 | [Eric A. Young] | |
11652 | ||
11653 | *) Added more RSA padding checks for SSL/TLS. | |
11654 | [Eric A. Young] | |
11655 | ||
11656 | *) Added BIO proxy/filter functionality. | |
11657 | [Eric A. Young] | |
11658 | ||
11659 | *) Added extra_certs to SSL_CTX which can be used | |
11660 | send extra CA certificates to the client in the CA cert chain sending | |
11661 | process. It can be configured with SSL_CTX_add_extra_chain_cert(). | |
11662 | [Eric A. Young] | |
11663 | ||
11664 | *) Now Fortezza is denied in the authentication phase because | |
11665 | this is key exchange mechanism is not supported by SSLeay at all. | |
11666 | [Eric A. Young] | |
11667 | ||
11668 | *) Additional PKCS1 checks. | |
11669 | [Eric A. Young] | |
11670 | ||
11671 | *) Support the string "TLSv1" for all TLS v1 ciphers. | |
11672 | [Eric A. Young] | |
11673 | ||
11674 | *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the | |
11675 | ex_data index of the SSL context in the X509_STORE_CTX ex_data. | |
11676 | [Eric A. Young] | |
11677 | ||
11678 | *) Fixed a few memory leaks. | |
11679 | [Eric A. Young] | |
11680 | ||
11681 | *) Fixed various code and comment typos. | |
11682 | [Eric A. Young] | |
11683 | ||
11684 | *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 | |
11685 | bytes sent in the client random. | |
11686 | [Edward Bishop <ebishop@spyglass.com>] | |
11687 |