Andy Polyakov [Fri, 1 Feb 2013 14:31:50 +0000 (15:31 +0100)]
ssl/*: remove SSL3_RECORD->orig_len to restore binary compatibility.
Kludge alert. This is arranged by passing padding length in unused
bits of SSL3_RECORD->type, so that orig_len can be reconstructed.
(cherry picked from commit 8bfd4c659f180a6ce34f21c0e62956b362067fba)
Adam Langley [Wed, 6 Feb 2013 15:50:42 +0000 (15:50 +0000)]
Fix for EXP-RC2-CBC-MD5
MD5 should use little endian order. Fortunately the only ciphersuite
affected is EXP-RC2-CBC-MD5 (TLS_RSA_EXPORT_WITH_RC2_CBC_40_MD5) which
is a rarely used export grade ciphersuite.
Add ordinal for CRYPTO_memcmp: since this will affect multiple
branches it needs to be in a "gap".
(cherry picked from commit 81ce0e14e72e8e255ad1bd9c7cfaa47a6291919c)
Andy Polyakov [Fri, 1 Feb 2013 08:59:56 +0000 (09:59 +0100)]
ssl/s3_cbc.c: md_state alignment portability fix.
RISCs are picky and alignment granted by compiler for md_state can be
insufficient for SHA512.
(cherry picked from commit 36260233e7e3396feed884d3f501283e0453c04f)
Andy Polyakov [Fri, 1 Feb 2013 08:55:43 +0000 (09:55 +0100)]
ssl/s3_cbc.c: uint64_t portability fix.
Break dependency on uint64_t. It's possible to declare bits as
unsigned int, because TLS packets are limited in size and 32-bit
value can't overflow.
(cherry picked from commit cab13fc8473856a43556d41d8dac5605f4ba1f91)
Ben Laurie [Mon, 28 Jan 2013 17:33:18 +0000 (17:33 +0000)]
Don't crash when processing a zero-length, TLS >= 1.1 record.
The previous CBC patch was bugged in that there was a path through enc()
in s3_pkt.c/d1_pkt.c which didn't set orig_len. orig_len would be left
at the previous value which could suggest that the packet was a
sufficient length when it wasn't.
(cherry picked from commit 6cb19b7681f600b2f165e4adc57547b097b475fd)
Ben Laurie [Mon, 28 Jan 2013 17:31:49 +0000 (17:31 +0000)]
Make CBC decoding constant time.
This patch makes the decoding of SSLv3 and TLS CBC records constant
time. Without this, a timing side-channel can be used to build a padding
oracle and mount Vaudenay's attack.
This patch also disables the stitched AESNI+SHA mode pending a similar
fix to that code.
In order to be easy to backport, this change is implemented in ssl/,
rather than as a generic AEAD mode. In the future this should be changed
around so that HMAC isn't in ssl/, but crypto/ as FIPS expects.
(cherry picked from commit e130841bccfc0bb9da254dc84e23bc6a1c78a64e)
Ben Laurie [Mon, 28 Jan 2013 17:30:38 +0000 (17:30 +0000)]
Add and use a constant-time memcmp.
This change adds CRYPTO_memcmp, which compares two vectors of bytes in
an amount of time that's independent of their contents. It also changes
several MAC compares in the code to use this over the standard memcmp,
which may leak information about the size of a matching prefix.
(cherry picked from commit 2ee798880a246d648ecddadc5b91367bee4a5d98)
Multiple copies of the ENGINE will cause problems when it is cleaned up as
the methods are stored in static structures which will be overwritten and
freed up more than once.
Set static methods to NULL when the ENGINE is freed so it can be reloaded.
Andy Polyakov [Fri, 27 Apr 2012 20:21:26 +0000 (20:21 +0000)]
ppccpuid.pl: branch hints in OPENSSL_cleanse impact small block performance
of digest algorithms, mosty SHA, on Power7. Mystery of century, why SHA,
why slower algorithm are affected more... [from HEAD].
PR: 2794
Submitted by: Ashley Lai
Check for potentially exploitable overflows in asn1_d2i_read_bio
BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
in CRYPTO_realloc_clean.
Thanks to Tavis Ormandy, Google Security Team, for discovering this
issue and to Adam Langley <agl@chromium.org> for fixing it. (CVE-2012-2110)
Fix for CMS/PKCS7 MMA. If RSA decryption fails use a random key and
continue with symmetric decryption process to avoid leaking timing
information to an attacker.
Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
this issue. (CVE-2012-0884)