The lazy-initialisation of BN_MONT_CTX was serialising all threads, as noted by
Daniel Sands and co at Sandia. This was to handle the case that 2 or more
threads race to lazy-init the same context, but stunted all scalability in the
case where 2 or more threads are doing unrelated things! We favour the latter
case by punishing the former. The init work gets done by each thread that finds
the context to be uninitialised, and we then lock the "set" logic after that
work is done - the winning thread's work gets used, the losing threads throw
away what they've done.
(Imported from upstream's bf43446835)
It's not clear whether this inconsistency could lead to an actual
computation error, but it involved a BIGNUM being passed around the
montgomery logic in an inconsistent state. This was found using flags
-DBN_DEBUG -DBN_DEBUG_RAND, and working backwards from this assertion
in 'ectest';
ectest: bn_mul.c:960: BN_mul: Assertion `(_bnum2->top == 0) ||
(_bnum2->d[_bnum2->top - 1] != 0)' failed
(Imported from upstream's 3cc546a3bb)
When looking for an extension we need to set the last found
position to -1 to properly search all extensions.
PR#3309
(Imported from upstream's 5cd5e0219d)
A missing bounds check in the handling of the TLS heartbeat extension
can be used to reveal up to 64k of memory to a connected client or
server.
Thanks for Neel Mehta of Google Security for discovering this bug and to
Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
preparing the fix (CVE-2014-0160)
(Imported from upstream's 7e840163c0)
Don't clear verification errors from the error queue unless
SSL_BUILD_CHAIN_FLAG_CLEAR_ERROR is set.
If errors occur during verification and
SSL_BUILD_CHAIN_FLAG_IGNORE_ERROR is set return 2 so applications can
issue warnings.
(Imported from upstream's 2dd6976f6d)
New flags to build certificate chains. The can be used to rearrange
the chain so all an application needs to do is add all certificates
in arbitrary order and then build the chain to check and correct them.
Add verify error code when building chain.
(Imported from upstream's c5ea65b157)
Improve CBC decrypt and CTR by ~13/16%, which adds up to ~25/33%
improvement over "pre-Silvermont" version. [Add performance table to
aesni-x86.pl].
(Imported from upstream's b347341c75)
Include self-signed flag in certificates by checking SKID/AKID as well as
issuer and subject names. Although this is an incompatible change it should
have little impact in pratice because self-issued certificates that are not
self-signed are rarely encountered.
(Imported from upstream's c00f8d697a)
When a chain is complete and ends in a trusted root checks are also performed
on the TA and the callback notified with ok==1. For consistency do the same for
chains where the TA is not self signed.
(Imported from upstream's b07e4f2f46)
If an application calls the macro SSL_CTX_get_extra_chain_certs
return either the old "shared" extra certificates or those associated
with the current certificate.
This means applications which call SSL_CTX_use_certificate_chain_file
and retrieve the additional chain using SSL_CTX_get_extra_chain_certs
will still work. An application which only wants to check the shared
extra certificates can call the new macro
SSL_CTX_get_extra_chain_certs_only
(Imported from upstream's e0d4272a58 and
3bff195dca)
This change adds a new function, BN_bn2bin_padded, that attempts, as
much as possible, to serialise a BIGNUM in constant time.
This is used to avoid some timing leaks in RSA decryption.
Fix a bug in handling of 128 byte long PSK identity in
psk_client_callback.
OpenSSL supports PSK identities of up to (and including) 128 bytes in
length. PSK identity is obtained via the psk_client_callback,
implementors of which are expected to provide a NULL-terminated
identity. However, the callback is invoked with only 128 bytes of
storage thus making it impossible to return a 128 byte long identity and
the required additional NULL byte.
This CL fixes the issue by passing in a 129 byte long buffer into the
psk_client_callback. As a safety precaution, this CL also zeroes out the
buffer before passing it into the callback, uses strnlen for obtaining
the length of the identity returned by the callback, and aborts the
handshake if the identity (without the NULL terminator) is longer than
128 bytes.
https://rt.openssl.org/Ticket/Display.html?id=2608
Previously, this input to the base64 code:
================================================================================-
Would cause the output length of EVP_DecodeUpdate to be negative. When
that happened in the base64 BIO, it would crash. In PEM decoding, the
ASN.1 code actually maintains signed lengths and manages to simply error
out!
Move ECC SSL extensions to the end.
WebSphere Application Server 7.0 appears to be intolerant of an empty
extension at the end. To that end, also ensure we never send an empty
padding extension.
Adam Langley
Adam Langley