Commit Graph
1164 Commits
Author SHA1 Message Date
Frank Denis 1686da3d3c Remove the non-IETF versions of crypto_aead_xchacha20poly1305 2017-01-13 19:24:48 +01:00
Frank Denis 4e8832ed57 Merge branch 'master' of https://github.com/jedisct1/libsodium
* 'master' of https://github.com/jedisct1/libsodium:
  Indent
  xchacha20poly1305: optimize and be compatible with ietf chacha20poly1305 (#461)
2016-12-27 21:03:12 +01:00
Frank DENIS 24fd77ded3 Indent 2016-12-24 02:24:24 +01:00
Jason A. Donenfeld 6abad20323 xchacha20poly1305: optimize and be compatible with ietf chacha20poly1305 (#461)
Due to SSL, the IETF version of chacha20poly1305 is going to be the one
that's in libraries places. While the 12-byte nonce thing is a little
weird, it has other benefits, like adding padding to the auth tag, which
might help fend off certain attacks.

But more importantly, since chacha20poly1305 in the IETF construction is
lots of places, it would be useful to be able to build xchacha20poly1305
out of it. Fortunately it's very easy to make hchacha20 (either
stand-alone, or out of the normal chacha20 block function), and then
that can be composed with an existing library's chacha20poly1305. It
looks a bit like this:

    xchacha20poly1305(input, key, nonce) {
        new_key = hchacha20(key, nonce)
        return chacha20poly1305(input, new_key, nonce + 16)
    }

This is also an efficient way to do it, since it means hchacha20 must
only be computed once.

Unfortuantely, non-IETF xchacha20poly1305 means that you deprive
virtually all other libraries that only support the more common
IETF construction the ability the ability to interoperate with
libsodium, through the simple construction. Rather, it forces
everyone to reimplement the AEAD part.

So, this commit adds a xchacha20poly1305 that uses the IETF construction
with the padding.

While we're at it, we redefine xchacha20poly1305 in terms of
chacha20poly1305, which gives the same output, but computes one less
hchacha20 and is generally a lot cleaner and simpler to understand.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
2016-12-24 02:17:33 +01:00
Frank Denis c5735ef215 Merge branch 'master' of https://github.com/jedisct1/libsodium
* 'master' of https://github.com/jedisct1/libsodium:
  Check if atomic operations are supported
  Remove a whitespace following trailing backslash in a Makefile
2016-12-17 19:00:59 +01:00
Frank Denis d54b0b8d69 Do not include xchacha20poly1305 in minimal mode 2016-12-17 18:59:17 +01:00
Frank Denis db97a35502 Check if atomic operations are supported 2016-12-16 16:37:12 +01:00
Frank Denis 4c6f704084 Remove a whitespace following trailing backslash in a Makefile 2016-12-16 16:20:30 +01:00
Frank Denis 9d2ac5f747 Correct an assertion and prefer compile-time assertions 2016-12-11 20:28:03 +01:00
Frank Denis 9979762bbe Indent 2016-12-11 00:01:40 +01:00
Winston Durand 9cae7b6b7c fixed GCC2 bug seeing empty statement (#449) 2016-11-30 06:52:18 +01:00
Frank Denis 157c4a80c1 + crypto_aead_xchacha20poly1305 2016-11-26 21:29:26 +01:00
Frank Denis 8b7f03ddf7 Indent 2016-11-26 21:16:42 +01:00
Frank Denis 184110ccc5 + crypto_box_curve25519xchacha20poly1305_* 2016-11-26 21:06:23 +01:00
Frank Denis 54a1357ce3 Indent 2016-11-26 20:24:58 +01:00
Frank Denis 2ace041fd9 Add secretbox_xchacha20poly1305_easy 2016-11-26 19:45:24 +01:00
Frank Denis d4f384e388 Make crypto_secretbox_xsalsa20poly1305_open() as __warn_unused_result__ 2016-11-26 19:44:51 +01:00
Frank Denis 669ed597d0 Rename box_x*poly1305.c -> secretbox_x*poly1305.c for consistency 2016-11-26 14:12:47 +01:00
Frank Denis 2848984edf + secretbox_xchacha20poly1305 2016-11-26 14:04:23 +01:00
Frank Denis a86ac590d6 Reformat to make the style more consistent 2016-11-26 13:40:34 +01:00
Frank Denis 5eed910c11 Cast the scalar instead of the coefficient 2016-10-30 01:13:22 +02:00
Frank Denis 71f0693ee7 Argon2i: fix encoding issues
For compatibility with hashes might have been encoded using other libraries.
2016-10-26 22:50:38 +02:00
Frank Denis aff4aaeabf Change the garbage value to 0xdb
If that garbage value becomes the LSB of a pointer, the pointer is more
likely to be unaligned, an trigger more bugs.
2016-10-15 18:54:56 +02:00
Frank Denis 49741c59e8 Allows RANDOMBYTES_DEFAULT_IMPLEMENTATION to be overriden 2016-10-13 22:57:01 +02:00
Frank Denis 583c16707c + crypto_stream_xchacha20 2016-09-30 22:57:56 +02:00
Frank Denis 42dc78b38b Indent 2016-09-30 08:40:15 +02:00
Frank Denis b20d227f37 Avoid collision with a possibly existing int128 type definition 2016-09-30 08:36:50 +02:00
Frank Denis 53ee1fe758 Remove commented out code and avoid inconsistent indentation 2016-09-30 08:30:22 +02:00
Frank Denis f257413772 uint32 -> uint32_t 2016-09-30 08:26:24 +02:00
Robert Spychala 94ea419247 add preprocessor flag to skip blocking /dev/random during libsodium init (#429) 2016-09-20 21:13:07 +02:00
Frank Denis 26e8b0253f Argon2: check that m_cost/t_cost/lanes decode to uint32 2016-09-18 09:33:35 +02:00
Frank Denis 6035c0779b Back to dev mode 2016-08-04 02:28:21 +02:00
Frank Denis 2f4f718cd9 Remove dev flag 2016-07-31 16:34:11 +02:00
Frank Denis 19a9d18b9c _MSC_VER > 1600 -> _MSC_VER >= 1700 for consistency 2016-07-24 19:58:00 +02:00
Jan-E 6b739fc821 Fix VS2010 (and VC9) x64 build 2016-07-24 02:13:22 +02:00
Frank Denis 89918e94f1 crit_{enter,leave} can fail 2016-07-06 12:03:08 +02:00
Frank Denis 29492143ab Warn if the library is being compiled in a custom way 2016-07-02 10:07:38 +02:00
Frank Denis 2cc0bab0e3 Update comment 2016-06-29 15:31:23 +02:00
Frank Denis 648f46d22a Expose sodium_crit_enter() and sodium_crit_leave() internally 2016-06-29 15:28:15 +02:00
Frank Denis 5a3ff833fd Slightly change how the length of argon2 strings is checked 2016-06-19 23:26:08 +02:00
Frank Denis 6fad3644b5 Nits 2016-06-15 16:00:59 +02:00
Aaron Zauner 77c05fce0c document why RtlGenRandom is used 2016-06-15 21:50:33 +08:00
Frank Denis aaf6854edf Remove extra space 2016-06-08 08:34:45 +02:00
Frank Denis 9c12da0362 Don't include <immintrin.h> if it is not needed
Some environments provide <wmmintrin.h> but not <immintrin.h>
2016-05-26 06:08:26 -07:00
Frank Denis fa4e4bf174 Fix & simplify MADV_DO{NO}DUMP alternatives 2016-05-18 22:12:07 +02:00
Lev Serebryakov 7a4c4459f8 Support madvise() on FreeBSD
FreeBSD have madvise() behaviors equivalent to MADV_DONTDUMP and MADV_DODUMP but with its own names.
Add definitions for these behaviors used in sodium_mlock() and sodium_munlock() if FreeBSD names are found and Linux ones don't.
2016-05-18 23:04:09 +03:00
Frank Denis f01299a91b Indent 2016-05-17 23:05:04 +02:00
Samuel Neves 998bacf375 don't crash on Win32 2016-05-17 21:40:56 +01:00
Frank Denis efb81c7290 CRLF 2016-05-17 01:41:06 +02:00
Thomas Waldmann b9c266181b fix avx2 feature detection, fixes #395
cpuid needed to get called with EAX = 7 to get the "extended features"
(not with EAX = 1 for the "features").
2016-05-17 01:28:03 +02:00