]>
Commit | Line | Data |
---|---|---|
f1c236f8 | 1 | OpenSSL CHANGES |
651d0aff RE |
2 | _______________ |
3 | ||
c5e8580e | 4 | Changes between 0.9.6 and 0.9.7 [xx XXX 2000] |
a43cf9fa | 5 | |
88ce56f8 DSH |
6 | *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}() |
7 | in a single operation. X509_get0_pubkey_bitstr() extracts the public_key | |
8 | structure from a certificate. X509_pubkey_digest() digests tha public_key | |
9 | contents: this is used in various key identifiers. | |
10 | [Steve Henson] | |
11 | ||
8cff6331 DSH |
12 | *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign |
13 | keyUsage if basicConstraints absent for a CA. | |
14 | [Steve Henson] | |
15 | ||
903872d6 RL |
16 | *) Make SMIME_write_PKCS7() write mail header values with a format that |
17 | is more generally accepted (no spaces before the semicolon), since | |
18 | some programs can't parse those values properly otherwise. Also make | |
19 | sure BIO's that break lines after each write do not create invalid | |
20 | headers. | |
21 | [Richard Levitte] | |
22 | ||
b8470240 DSH |
23 | *) Make sk_sort() tolerate a NULL argument. |
24 | [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>] | |
25 | ||
50d51991 DSH |
26 | *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates |
27 | passed by the function are trusted implicitly. If any of them signed the | |
28 | reponse then it is assumed to be valid and is not verified. | |
29 | [Steve Henson] | |
30 | ||
a342cc5a DSH |
31 | *) Zero the premaster secret after deriving the master secret in |
32 | DH ciphersuites. | |
33 | [Steve Henson] | |
34 | ||
a43cf9fa DSH |
35 | *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT |
36 | to data. This was previously part of the PKCS7 ASN1 code. This | |
37 | was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures. | |
38 | [Steve Henson, reported by Kenneth R. Robinette | |
39 | <support@securenetterm.com>] | |
40 | ||
41 | *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1 | |
42 | routines: without these tracing memory leaks is very painful. | |
43 | Fix leaks in PKCS12 and PKCS7 routines. | |
44 | [Steve Henson] | |
45 | ||
75802000 UM |
46 | *) Fix for Irix with NO_ASM. |
47 | ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>] | |
48 | ||
ae0665b8 BM |
49 | *) Add some EVP_add_digest_alias registrations (as found in |
50 | OpenSSL_add_all_digests()), to SSL_library_init() | |
51 | aka OpenSSL_add_ssl_algorithms(). This provides improved | |
52 | compatibility with peers using X.509 certificates | |
53 | with unconventional AlgorithmIdentifier OIDs. | |
54 | [Bodo Moeller] | |
55 | ||
893b76c5 UM |
56 | *) ./config script fixes. |
57 | [Ulf Moeller, Richard Levitte] | |
58 | ||
ba8e2824 DSH |
59 | *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new(). |
60 | Previously it initialised the 'type' argument to V_ASN1_UTCTIME which | |
61 | effectively meant GeneralizedTime would never be used. Now it | |
62 | is initialised to -1 but X509_time_adj() now has to check the value | |
63 | and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or | |
64 | V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime. | |
65 | [Steve Henson, reported by Kenneth R. Robinette | |
66 | <support@securenetterm.com>] | |
67 | ||
8e8972bb DSH |
68 | *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously |
69 | result in a zero length in the ASN1_INTEGER structure which was | |
70 | not consistent with the structure when d2i_ASN1_INTEGER() was used | |
71 | and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER() | |
72 | to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER() | |
73 | where it did not print out a minus for negative ASN1_INTEGER. | |
74 | [Steve Henson] | |
75 | ||
57108f0a BM |
76 | *) Fix 'openssl passwd -1'. |
77 | [Bodo Moeller] | |
78 | ||
73758d43 DSH |
79 | *) Add summary printout to ocsp utility. The various functions which |
80 | convert status values to strings have been renamed to: | |
81 | OCSP_response_status_str(), OCSP_cert_status_str() and | |
82 | OCSP_crl_reason_str() and are no longer static. New options | |
83 | to verify nonce values and to disable verification. OCSP response | |
84 | printout format cleaned up. | |
85 | [Steve Henson] | |
86 | ||
e8af92fc DSH |
87 | *) Add additional OCSP certificate checks. These are those specified |
88 | in RFC2560. This consists of two separate checks: the CA of the | |
89 | certificate being checked must either be the OCSP signer certificate | |
90 | or the issuer of the OCSP signer certificate. In the latter case the | |
91 | OCSP signer certificate must contain the OCSP signing extended key | |
92 | usage. This check is performed by attempting to match the OCSP | |
93 | signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash | |
94 | in the OCSP_CERTID structures of the response. | |
95 | [Steve Henson] | |
96 | ||
81f169e9 DSH |
97 | *) Initial OCSP certificate verification added to OCSP_basic_verify() |
98 | and related routines. This uses the standard OpenSSL certificate | |
99 | verify routines to perform initial checks (just CA validity) and | |
100 | to obtain the certificate chain. Then additional checks will be | |
101 | performed on the chain. Currently the root CA is checked to see | |
102 | if it is explicitly trusted for OCSP signing. This is used to set | |
103 | a root CA as a global signing root: that is any certificate that | |
104 | chains to that CA is an acceptable OCSP signing certificate. | |
105 | [Steve Henson] | |
106 | ||
dfebac32 BM |
107 | *) New '-extfile ...' option to 'openssl ca' for reading X.509v3 |
108 | extensions from a separate configuration file. | |
109 | As when reading extensions from the main configuration file, | |
110 | the '-extensions ...' option may be used for specifying the | |
111 | section to use. | |
112 | [Massimiliano Pala <madwolf@comune.modena.it>] | |
113 | ||
6308af19 DSH |
114 | *) Change PKCS12_key_gen_asc() so it can cope with non null |
115 | terminated strings whose length is passed in the passlen | |
116 | parameter, for example from PEM callbacks. This was done | |
117 | by adding an extra length parameter to asc2uni(). | |
118 | [Steve Henson, reported by <oddissey@samsung.co.kr>] | |
119 | ||
5782ceb2 DSH |
120 | *) New OCSP utility. Allows OCSP requests to be generated or |
121 | read. The request can be sent to a responder and the output | |
122 | parsed, outputed or printed in text form. Not complete yet: | |
123 | still needs to check the OCSP response validity. | |
124 | [Steve Henson] | |
125 | ||
c67cdb50 BM |
126 | *) New subcommands for 'openssl ca': |
127 | 'openssl ca -status <serial>' prints the status of the cert with | |
128 | the given serial number (according to the index file). | |
129 | 'openssl ca -updatedb' updates the expiry status of certificates | |
130 | in the index file. | |
131 | [Massimiliano Pala <madwolf@comune.modena.it>] | |
132 | ||
d199858e BM |
133 | *) New '-newreq-nodes' command option to CA.pl. This is like |
134 | '-newreq', but calls 'openssl req' with the '-nodes' option | |
135 | so that the resulting key is not encrypted. | |
136 | [Damien Miller <djm@mindrot.org>] | |
137 | ||
10a2975a RL |
138 | *) New configuration for the GNU Hurd. |
139 | [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte] | |
140 | ||
9b4dc830 DSH |
141 | *) Initial code to implement OCSP basic response verify. This |
142 | is currently incomplete. Currently just finds the signer's | |
143 | certificate and verifies the signature on the response. | |
144 | [Steve Henson] | |
145 | ||
673b3fde BM |
146 | *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in |
147 | value of OPENSSLDIR. This is available via the new '-d' option | |
148 | to 'openssl version', and is also included in 'openssl version -a'. | |
149 | [Bodo Moeller] | |
150 | ||
c06648f7 BM |
151 | *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn |
152 | call failed, free the DSA structure. | |
153 | [Bodo Moeller] | |
154 | ||
a5435e8b BM |
155 | *) Allowing defining memory allocation callbacks that will be given |
156 | file name and line number information in additional arguments | |
157 | (a const char* and an int). The basic functionality remains, as | |
158 | well as the original possibility to just replace malloc(), | |
159 | realloc() and free() by functions that do not know about these | |
160 | additional arguments. To register and find out the current | |
161 | settings for extended allocation functions, the following | |
162 | functions are provided: | |
65a22e8e RL |
163 | |
164 | CRYPTO_set_mem_ex_functions | |
165 | CRYPTO_set_locked_mem_ex_functions | |
166 | CRYPTO_get_mem_ex_functions | |
167 | CRYPTO_get_locked_mem_ex_functions | |
168 | ||
a5435e8b BM |
169 | These work the same way as CRYPTO_set_mem_functions and friends. |
170 | CRYPTO_get_[locked_]mem_functions now writes 0 where such an | |
171 | extended allocation function is enabled. | |
172 | Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where | |
173 | a conventional allocation function is enabled. | |
174 | [Richard Levitte, Bodo Moeller] | |
65a22e8e | 175 | |
cbf0f45f DSH |
176 | *) Fix to uni2asc() to cope with zero length Unicode strings. |
177 | These are present in some PKCS#12 files. | |
178 | [Steve Henson] | |
179 | ||
3c914840 GT |
180 | *) Finish off removing the remaining LHASH function pointer casts. |
181 | There should no longer be any prototype-casting required when using | |
56a67adb GT |
182 | the LHASH abstraction, and any casts that remain are "bugs". See |
183 | the callback types and macros at the head of lhash.h for details | |
184 | (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example). | |
3c914840 GT |
185 | [Geoff Thorpe] |
186 | ||
599c0353 LJ |
187 | *) Add automatic query of EGD sockets in RAND_poll() for the unix variant. |
188 | If an EGD or PRNGD is running and enough entropy is returned, automatic | |
189 | seeding like with /dev/[u]random will be performed. | |
190 | Positions tried are: /etc/entropy, /var/run/egd-pool. | |
191 | [Lutz Jaenicke] | |
192 | ||
0c61e299 | 193 | *) Change the Unix RAND_poll() variant to be able to poll several |
361ef5f4 RL |
194 | random devices, as specified by DEVRANDOM, until a sufficient amount |
195 | of data has been collected. We spend at most 10 ms on each file | |
196 | (select timeout) and read in non-blocking mode. DEVRANDOM now | |
197 | defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom" | |
198 | (previously it was just the string "/dev/urandom"), so on typical | |
199 | platforms the 10 ms delay will never occur. | |
200 | Also separate out the Unix variant to its own file, rand_unix.c. | |
201 | For VMS, there's a currently-empty rand_vms.c. | |
0c61e299 RL |
202 | [Richard Levitte] |
203 | ||
0b33bc65 DSH |
204 | *) Move OCSP client related routines to ocsp_cl.c. These |
205 | provide utility functions which an application needing | |
206 | to issue a request to an OCSP responder and analyse the | |
207 | response will typically need: as opposed to those which an | |
208 | OCSP responder itself would need which will be added later. | |
209 | ||
210 | OCSP_request_sign() signs an OCSP request with an API similar | |
211 | to PKCS7_sign(). OCSP_response_status() returns status of OCSP | |
212 | response. OCSP_response_get1_basic() extracts basic response | |
213 | from response. OCSP_resp_find_status(): finds and extracts status | |
214 | information from an OCSP_CERTID structure (which will be created | |
215 | when the request structure is built). These are built from lower | |
216 | level functions which work on OCSP_SINGLERESP structures but | |
217 | wont normally be used unless the application wishes to examine | |
218 | extensions in the OCSP response for example. | |
219 | ||
220 | Replace nonce routines with a pair of functions. | |
221 | OCSP_request_add1_nonce() adds a nonce value and optionally | |
222 | generates a random value. OCSP_check_nonce() checks the | |
223 | validity of the nonce in an OCSP response. | |
224 | [Steve Henson] | |
225 | ||
226 | *) Change function OCSP_request_add() to OCSP_request_add0_id(). | |
8e961835 DSH |
227 | This doesn't copy the supplied OCSP_CERTID and avoids the |
228 | need to free up the newly created id. Change return type | |
229 | to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure. | |
230 | This can then be used to add extensions to the request. | |
231 | Deleted OCSP_request_new(), since most of its functionality | |
232 | is now in OCSP_REQUEST_new() (and the case insensitive name | |
233 | clash) apart from the ability to set the request name which | |
234 | will be added elsewhere. | |
235 | [Steve Henson] | |
236 | ||
bf0d176e DSH |
237 | *) Update OCSP API. Remove obsolete extensions argument from |
238 | various functions. Extensions are now handled using the new | |
239 | OCSP extension code. New simple OCSP HTTP function which | |
240 | can be used to send requests and parse the response. | |
241 | [Steve Henson] | |
242 | ||
ec5add87 DSH |
243 | *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new |
244 | ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN | |
245 | uses the special reorder version of SET OF to sort the attributes | |
246 | and reorder them to match the encoded order. This resolves a long | |
247 | standing problem: a verify on a PKCS7 structure just after signing | |
248 | it used to fail because the attribute order did not match the | |
249 | encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes: | |
250 | it uses the received order. This is necessary to tolerate some broken | |
251 | software that does not order SET OF. This is handled by encoding | |
252 | as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class) | |
253 | to produce the required SET OF. | |
254 | [Steve Henson] | |
255 | ||
a6574c21 RL |
256 | *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and |
257 | OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header | |
258 | files to get correct declarations of the ASN.1 item variables. | |
259 | [Richard Levitte] | |
260 | ||
ecbe0781 DSH |
261 | *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many |
262 | PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs: | |
263 | asn1_check_tlen() would sometimes attempt to use 'ctx' when it was | |
264 | NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i(). | |
265 | New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant | |
266 | ASN1_ITEM and no wrapper functions. | |
267 | [Steve Henson] | |
268 | ||
4e1209eb DSH |
269 | *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These |
270 | replace the old function pointer based I/O routines. Change most of | |
271 | the *_d2i_bio() and *_d2i_fp() functions to use these. | |
272 | [Steve Henson] | |
273 | ||
3f07fe09 RL |
274 | *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor |
275 | lines, recognice more "algorithms" that can be deselected, and make | |
276 | it complain about algorithm deselection that isn't recognised. | |
277 | [Richard Levitte] | |
278 | ||
78d3b819 | 279 | *) New ASN1 functions to handle dup, sign, verify, digest, pack and |
73e92de5 DSH |
280 | unpack operations in terms of ASN1_ITEM. Modify existing wrappers |
281 | to use new functions. Add NO_ASN1_OLD which can be set to remove | |
282 | some old style ASN1 functions: this can be used to determine if old | |
283 | code will still work when these eventually go away. | |
09ab755c DSH |
284 | [Steve Henson] |
285 | ||
ec558b65 DSH |
286 | *) New extension functions for OCSP structures, these follow the |
287 | same conventions as certificates and CRLs. | |
288 | [Steve Henson] | |
289 | ||
57d2f217 DSH |
290 | *) New function X509V3_add1_i2d(). This automatically encodes and |
291 | adds an extension. Its behaviour can be customised with various | |
292 | flags to append, replace or delete. Various wrappers added for | |
293 | certifcates and CRLs. | |
294 | [Steve Henson] | |
295 | ||
5755cab4 DSH |
296 | *) Fix to avoid calling the underlying ASN1 print routine when |
297 | an extension cannot be parsed. Correct a typo in the | |
298 | OCSP_SERVICELOC extension. Tidy up print OCSP format. | |
299 | [Steve Henson] | |
300 | ||
3880cd35 BM |
301 | *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c). |
302 | Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits | |
303 | when writing a 32767 byte record. | |
304 | [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>] | |
305 | ||
f640ee90 | 306 | *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c), |
126fe085 | 307 | obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}. |
f640ee90 BM |
308 | |
309 | (RSA objects have a reference count access to which is protected | |
310 | by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c], | |
311 | so they are meant to be shared between threads.) | |
126fe085 BM |
312 | [Bodo Moeller, Geoff Thorpe; original patch submitted by |
313 | "Reddie, Steven" <Steven.Reddie@ca.com>] | |
f640ee90 | 314 | |
9c67ab2f DSH |
315 | *) Make mkdef.pl parse some of the ASN1 macros and add apropriate |
316 | entries for variables. | |
5755cab4 | 317 | [Steve Henson] |
9c67ab2f | 318 | |
1456d186 BM |
319 | *) Fix a deadlock in CRYPTO_mem_leaks(). |
320 | [Bodo Moeller] | |
321 | ||
3ac82faa BM |
322 | *) Add functionality to apps/openssl.c for detecting locking |
323 | problems: As the program is single-threaded, all we have | |
324 | to do is register a locking callback using an array for | |
325 | storing which locks are currently held by the program. | |
3ac82faa BM |
326 | [Bodo Moeller] |
327 | ||
328 | *) Use a lock around the call to CRYPTO_get_ex_new_index() in | |
329 | SSL_get_ex_data_X509_STORE_idx(), which is used in | |
330 | ssl_verify_cert_chain() and thus can be called at any time | |
331 | during TLS/SSL handshakes so that thread-safety is essential. | |
332 | Unfortunately, the ex_data design is not at all suited | |
333 | for multi-threaded use, so it probably should be abolished. | |
334 | [Bodo Moeller] | |
335 | ||
2a86064f GT |
336 | *) Added Broadcom "ubsec" ENGINE to OpenSSL. |
337 | [Broadcom, tweaked and integrated by Geoff Thorpe] | |
338 | ||
2c15d426 DSH |
339 | *) Move common extension printing code to new function |
340 | X509V3_print_extensions(). Reorganise OCSP print routines and | |
c08523d8 | 341 | implement some needed OCSP ASN1 functions. Add OCSP extensions. |
2c15d426 DSH |
342 | [Steve Henson] |
343 | ||
de487514 DSH |
344 | *) New function X509_signature_print() to remove duplication in some |
345 | print routines. | |
346 | [Steve Henson] | |
347 | ||
06db4253 DSH |
348 | *) Add a special meaning when SET OF and SEQUENCE OF flags are both |
349 | set (this was treated exactly the same as SET OF previously). This | |
350 | is used to reorder the STACK representing the structure to match the | |
351 | encoding. This will be used to get round a problem where a PKCS7 | |
352 | structure which was signed could not be verified because the STACK | |
353 | order did not reflect the encoded order. | |
354 | [Steve Henson] | |
355 | ||
36f554d4 DSH |
356 | *) Reimplement the OCSP ASN1 module using the new code. |
357 | [Steve Henson] | |
358 | ||
2aff7727 DSH |
359 | *) Update the X509V3 code to permit the use of an ASN1_ITEM structure |
360 | for its ASN1 operations. The old style function pointers still exist | |
361 | for now but they will eventually go away. | |
362 | [Steve Henson] | |
363 | ||
9d6b1ce6 | 364 | *) Merge in replacement ASN1 code from the ASN1 branch. This almost |
5755cab4 DSH |
365 | completely replaces the old ASN1 functionality with a table driven |
366 | encoder and decoder which interprets an ASN1_ITEM structure describing | |
367 | the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is | |
368 | largely maintained. Almost all of the old asn1_mac.h macro based ASN1 | |
369 | has also been converted to the new form. | |
9d6b1ce6 DSH |
370 | [Steve Henson] |
371 | ||
8dea52fa BM |
372 | *) Change BN_mod_exp_recp so that negative moduli are tolerated |
373 | (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set | |
374 | so that BN_mod_exp_mont and BN_mod_exp_mont_word work | |
375 | for negative moduli. | |
376 | [Bodo Moeller] | |
377 | ||
378 | *) Fix BN_uadd and BN_usub: Always return non-negative results instead | |
379 | of not touching the result's sign bit. | |
380 | [Bodo Moeller] | |
381 | ||
80d89e6a BM |
382 | *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be |
383 | set. | |
384 | [Bodo Moeller] | |
385 | ||
f1919c3d GT |
386 | *) Changed the LHASH code to use prototypes for callbacks, and created |
387 | macros to declare and implement thin (optionally static) functions | |
388 | that provide type-safety and avoid function pointer casting for the | |
389 | type-specific callbacks. | |
390 | [Geoff Thorpe] | |
391 | ||
1946cd8b UM |
392 | *) Use better test patterns in bntest. |
393 |