Commit Graph

12950 Commits

Author SHA1 Message Date
Andy Polyakov
5da05a26f2 cms-test.pl: "localize" /dev/null even further [as follow-up to VMS].
Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-02-02 15:27:07 +01:00
Richard Levitte
1d4d68570b Make the libssl opaque changes compile on VMS
Reviewed-by: Matt Caswell <matt@openssl.org>
2015-01-31 18:07:32 +00:00
Matt Caswell
78cc1f03e8 Add changes entry for opaquifying of libssl structures
Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-31 18:07:22 +00:00
Matt Caswell
0c2837564c Remove OPENSSL_NO_SSL_INTERN as it is now redundant - all internals
previously protected by this have been moved into non-public headers

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-31 18:07:11 +00:00
Matt Caswell
b6ba401497 Make libssl opaque. Move all structures that were previously protected by
OPENSSL_NO_SSL_INTERN into internal header files.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-31 18:06:45 +00:00
Ben Laurie
4de8385796 Build correctly for me on FreeBSD 10.
Reviewed-by: Rich Salz

Don't debug.
2015-01-30 22:23:17 +00:00
Rich Salz
02a938c953 Dead code removal: #if 0 asn1, pkcs7
Keep one #if 0 but rename the symbol to be more descriptive of what
it's doing (you can disable support for old broken Netscape software).

Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-30 15:35:49 -05:00
Rich Salz
75d0ebef2a Dead code clean: #if 0 removal in apps
Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-30 14:52:57 -05:00
Rich Salz
d6fbb19409 Dead code removal #if 0 engines
Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-30 13:24:35 -05:00
Rich Salz
6f1a93ad11 Dead code removal: #if 0 conf, dso, pqueue, threads
Mostly, but not completely, debugging print statements.
Some old logic kept for internal documentation reasons, perhaps.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-30 12:46:49 -05:00
Andy Polyakov
2e635aa81c modes/gcm128.c: harmonize ctx->ghash assignment, shortcut *_ctr32
in OPENSSL_SMALL_FOOTPRINT build, remove undesired reformat artefact
and inconsistency in pre-processor logic.

Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-01-30 16:37:21 +01:00
Andy Polyakov
b2991c081a modes/gcm128.c: fix OPENSSL_SMALL_FOOTPRINT compile failure
on affected platforms (PowerPC and AArch64).

For reference, minimalistic #ifdef GHASH is sufficient, because
it's never defined with OPENSSL_SMALL_FOOTPRINT and ctx->ghash
is never referred.

Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-01-30 16:36:27 +01:00
Richard Levitte
4938ebc406 Since SHA0 was completely removed, also remove the related test
Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-30 15:14:48 +01:00
Richard Levitte
4fdde1aa0c Update on the use of logical names for OpenSSL configuration
Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-30 14:44:46 +01:00
Richard Levitte
e00ab250c8 VMS exit codes weren't handled well enough and were unclear
Making a specific variable $failure_code and a bit of commenting in the
VMS section should help clear things up.

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-30 14:44:36 +01:00
Richard Levitte
09ebad72df VMS adjustments:
Add missing crypto modules and files to copy to crypto/install-crypto.com

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-30 14:44:27 +01:00
Richard Levitte
36ed7adfbc VMS adjustments:
test/cms-test.pl adjusted to handle NL: instead of /dev/null on VMS

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-30 14:44:18 +01:00
Richard Levitte
36759bb751 VMS build changes
crypto/crypto-lib.com:
 Remove all APPS building, as they are gone.
 Depend on the variable SDIRS that's defined by makevms.com.
 Remake the whole partial module list mechanism to check for variables with a counter.
 Define the logical name INTERNAL to allow for '#include "internal/foo.h"'.

makevms.com:
 Define SDIRS, to allow for removal of crypto modules and pass that information to crypto/crypto-lib.com.
 Allow for experimental modules.
 Update the allowed things to disable.
 Update the things disabled by default to match Configure.
 Update headers to be copied.

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-30 14:44:06 +01:00
Richard Levitte
132536f96e VMS adjustments:
catch up with the Unix build.
A number of new tests, among others test/tocsp.com
Define INTERNAL in ssl/ssl-lib.com to allow for '#include "internal/foo.h"'

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-30 14:43:57 +01:00
Richard Levitte
c168a027cf VMS adjustments:
Add new symbols that are longer than 31 chars to symhacks.
VMS doesn't have <sys/un.h>, reflect that in e_os.h.
MS_CALLBACK has been removed, ssl_task.c needs adjustment.

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-30 14:43:40 +01:00
Richard Levitte
be7b1097e2 dso_vms needs to add the .EXE extension if there is none already
Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-01-30 04:44:17 +01:00
Rich Salz
4d428cd250 Dead code removal: #if 0 bio, comp, rand
The start of removing dead code.
A remaining #if 0 in bss_conn.c needs more thought.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-29 21:38:57 -05:00
Rich Salz
33fc38ff8e Make output consistency: remove blank line
When you use "-s" in the make flag, you see that engines outputs
a blank line because EDIRS isn't set.  This is a debug echo that
isn't needed.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-29 12:09:14 -05:00
Richard Levitte
c6ef15c494 clang on Linux x86_64 complains about unreachable code.
Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-01-29 01:54:09 +01:00
Matt Caswell
7317192c64 Fix various windows compilation issues
Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-28 22:55:15 +00:00
Rich Salz
537bf4381b Fix int/unsigned compiler complaint
Reviewed-by: Matt Caswell <matt@openssl.org>
2015-01-28 15:41:14 -05:00
Rich Salz
68fd6dce73 Remove support for opaque-prf
An expired IETF Internet-Draft (seven years old) that nobody
implements, and probably just as good as NSA DRBG work.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-28 15:37:16 -05:00
Rich Salz
31b446e212 Add missing declaration for lh_node_usage_stats
Reviewed-by: Matt Caswell <matt@openssl.org>
2015-01-28 12:27:23 -05:00
Rich Salz
49b05c7d50 Rename index to idx to avoid symbol conflicts.
Picky compilers with old index() string functions.
Reviewed-by: Matt Caswell <matt@openssl.org>
2015-01-28 12:23:01 -05:00
Rich Salz
625a9baf11 Finish removal of DSS
Reviewed-by: Matt Caswell <matt@openssl.org>
2015-01-28 12:21:55 -05:00
Matt Caswell
55467a16c2 Fix warning on some compilers where variable index shadows a global
declaration

Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-01-28 10:57:14 +00:00
Matt Caswell
dc0e9a35fa Fix no-ocb for Windows
Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-28 10:39:15 +00:00
Matt Caswell
488ede07bd Rationalise testing of AEAD modes
Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-28 10:39:07 +00:00
Matt Caswell
e640fa0200 Harmonise use of EVP_CTRL_GET_TAG/EVP_CTRL_SET_TAG/EVP_CTRL_SET_IVLEN
Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-28 10:39:01 +00:00
Matt Caswell
d57d135c33 Replace EVP_CTRL_OCB_SET_TAGLEN with EVP_CTRL_SET_TAG for consistency with
CCM

Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-28 10:38:38 +00:00
Rich Salz
646e8c1d6b Dead code removal: Fortezza identifiers
Not interested in helping the NSA in the slightest.
And anyway, it was never implemented, #if'd out.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-27 21:00:03 -05:00
Rich Salz
1a5adcfb5e "#if 0" removal: header files
Remove all "#if 0" blocks from header files.

Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-27 17:44:12 -05:00
Rich Salz
63c574f6a6 OPENSSL_NO_XXX cleanup: OPENSSL_NO_BUF_FREELISTS
Remove OPENSSL_NO_BUF_FREELISTS. This was turned on by default,
so the work here is removing the 'maintain our own freelist' code.
Also removed a minor old Windows-multibyte/widechar conversion flag.

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-27 16:43:53 -05:00
Rich Salz
daa48704cc OPENSSL_NO_XXX cleanup: NO_TLS, NO_TLS1
TLS and TLS1 are no longer optional.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-27 15:14:12 -05:00
Rich Salz
474e469bbd OPENSSL_NO_xxx cleanup: SHA
Remove support for SHA0 and DSS0 (they were broken), and remove
the ability to attempt to build without SHA (it didn't work).
For simplicity, remove the option of not building various SHA algorithms;
you could argue that SHA_224/256/384/512 should be kept, since they're
like crypto algorithms, but I decided to go the other way.
So these options are gone:
	GENUINE_DSA         OPENSSL_NO_SHA0
	OPENSSL_NO_SHA      OPENSSL_NO_SHA1
	OPENSSL_NO_SHA224   OPENSSL_NO_SHA256
	OPENSSL_NO_SHA384   OPENSSL_NO_SHA512

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-27 12:34:45 -05:00
Rich Salz
c73ad69017 OPENSSL_NO_xxx cleanup: RFC3779
Remove OPENSSL_NO_RFCF3779.

Also, makevms.com was ignored by some of the other cleanups, so
I caught it up.  Sorry I ignored you, poor little VMS...

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-27 10:19:14 -05:00
Rich Salz
a00ae6c46e OPENSSL_NO_xxx cleanup: many removals
The following compile options (#ifdef's) are removed:
    OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
    OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
    OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
    OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY

This diff is big because of updating the indents on preprocessor lines.

Reviewed-by: Richard Levitte <levitte@openssl.org>
2015-01-27 10:06:22 -05:00
Rich Salz
109f1031a8 OPENSSL_NO_xxx cleanup: DEC-CBCM removed
A DES algorithm mode, known attacks, no EVP support.
Flushed.

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-27 10:02:39 -05:00
Matt Caswell
8507474564 Provide documentation for all SSL(_CTX)?_(get|set)(_default)?_read_ahead
functions.

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-27 14:28:41 +00:00
Matt Caswell
f400241251 Remove explicit setting of read_ahead for DTLS. It never makes sense not to
use read_ahead with DTLS because it doesn't work. Therefore read_ahead needs
to be the default.

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-27 14:28:41 +00:00
Matt Caswell
8dd4ad0ff5 Make DTLS always act as if read_ahead is set. The actual value of read_ahead
is ignored for DTLS.

RT#3657

Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-27 14:28:41 +00:00
Andy Polyakov
2a4af9478d Configure: addendum to OPENSSL_NO_[RMD160|RIPEMD] harmonization.
Reviewed-by: Rich Salz <rsalz@openssl.org>
2015-01-27 12:23:42 +01:00
Andy Polyakov
2863d5f3cd des/asm/des_enc.m4: strip #ifdef OPENSSL_SYS_ULTRASPARC as part of
pre-processor controls cleanup. It doesn't mean that it no longer
works on UltraSPARC, only that it doesn't utilize sparcv9-specific
features like branch prediction hints and load in little-endian byte
order anymore. This "costs" ~3% in EDE3 performance regression on
UltraSPARC.

Reviewed-by: Rich Salz <rsalz@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
2015-01-27 12:18:15 +01:00
Rich Salz
3d0cf91807 Remove obsolete support for old code.
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Andy Polyakov <appro@openssl.org>
2015-01-26 11:04:26 -05:00
Rich Salz
abdd677125 Make OPENSSL_config truly ignore errors.
Per discussion: should not exit. Should not print to stderr.
Errors are ignored.  Updated doc to reflect that, and the fact
that this function is to be avoided.

Reviewed-by: Matt Caswell <matt@openssl.org>
Reviewed-by: Tim Hudson <tjh@openssl.org>
Reviewed-by: Viktor Dukhovni <viktor@openssl.org>
2015-01-25 21:07:20 -05:00