- Apr 11, 2017
-
-
Frank Morgner authored
fixes segmentation fault in case of not enough memory for object creation CLA: trivial Reviewed-by:
Richard Levitte <levitte@openssl.org> Reviewed-by:
Rich Salz <rsalz@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3157) (cherry picked from commit 487a73de)
-
Richard Levitte authored
Clearing a misunderstanding. The routines c2i_uint64_int() and i2c_uint64_int() expect to receive that internal values are absolute and with a separate sign flag, and the x_int64.c code handles values that aren't absolute and have the sign bit embedded. We therefore need to convert between absolute and non-absolute values for the encoding of negative values to be correct. [extended tests] Reviewed-by:
Andy Polyakov <appro@openssl.org> (Merged from https://github.com/openssl/openssl/pull/3160) (cherry picked from commit 429223d1)
-
Todd Short authored
SSLv3 does not support TLS extensions, and thus, cannot provide any curves for ECDH(E). With the removal of the default (all) list of curves being used for connections that didn't provide any curves, ECDHE is no longer possible. Reviewed-by:
Kurt Roeckx <kurt@openssl.org> Reviewed-by:
-
Rich Salz authored
Reviewed-by:
Matt Caswell <matt@openssl.org> Reviewed-by:
-
Matt Caswell authored
Reviewed-by:
-
Benjamin Kaduk authored
RFC 7301 mandates that the server SHALL respond with a fatal "no_application_protocol" alert when there is no overlap between the client's supplied list and the server's list of supported protocols. In commit 06217867 we changed from ignoring non-success returns from the supplied alpn_select_cb() to treating such non-success returns as indicative of non-overlap and sending the fatal alert. In effect, this is using the presence of an alpn_select_cb() as a proxy to attempt to determine whether the application has configured a list of supported protocols. However, there may be cases in which an application's architecture leads it to supply an alpn_select_cb() but have that callback be configured to take no action on connections that do not have ALPN configured; returning SSL_TLSEXT_ERR_NOACK from the callback would be the natural way to do so. Unfortunately, the aforementioned behavior change also treated SSL_TLSEXT_ERR_NOACK as indicative of no overlap and terminated the connection; this change supplies special handling for SSL_TLSEXT_ERR_NOACK returns from the callback. In effect, it provides a way for a callback to obtain the behavior that would have occurred if no callback was registered at all, which was not possible prior to this change. Reviewed-by:
-
- Apr 10, 2017
-
-
Andy Polyakov authored
Reviewed-by:
-
Andy Polyakov authored
Reviewed-by:
-
- Apr 08, 2017
-
-
Andy Polyakov authored
PRIu64 is error-prone with BIO_printf, so introduce and stick to custom platform-neutral macro. 'll' allows to print 64-bit values on *all* supported platforms, but it's problematic with -Wformat -Werror. Hence use 'l' in identifiable LP64 cases. Reviewed-by:
-
Andy Polyakov authored
This might seem controversial, but it doesn't actually affect anything. Or rather it doesn't make worse cases when it was problematic [with code additions to 1.1.0]. One of such rare cases is 32-bit PA-RISC target with *vendor* compiler. Reviewed-by:
-
- Apr 07, 2017
-
-
Rich Salz authored
(cherry picked from commit dfc63ccebd8262648640fc6e35fffa132102b967) Reviewed-by:
-
Rich Salz authored
(cherry picked from commit 7ea8ceab4b21d93eb272a89bff73958010d5c2c4) Reviewed-by:
-
Richard Levitte authored
When configured no-engine, we still refered to rand_engine_lock. Rework the lock init code to avoid that. Reviewed-by:
-
Richard Levitte authored
This is especially harmful since OPENSSL_cleanup() has already called the RAND cleanup function Reviewed-by:
-
Richard Levitte authored
Reviewed-by:
-
- Apr 06, 2017
-
-
Qin Long authored
Under UEFI build environment, we may encounter the OSSL_SSIZE macro re-definition error in e_os2.h if any module call OpenSSL API directly by including "openssl/xxxx.h" (caused by the predefined _WIN32/_WIN64 macro, which should have been un-defined under OPENSSL_SYS_UEFI). Though it's not one recommended usage, this patch could still eliminate the possible build issue by refining the OSSL_SSIZE definition under OPENSSL_SYS_UEFI. Reviewed-by:
-
Richard Levitte authored
If no default method was yet given, RAND_get_rand_method() will set it up. Doing so just to clean it away seems pretty silly, so instead, use the default_RAND_meth variable directly. This also clears a possible race condition where this will try to init things, such as ERR or ENGINE when in the middle of a OPENSSL_cleanup. Fixes #3128 Reviewed-by:
-
- Apr 04, 2017
-
-
Andy Polyakov authored
Fixes GH#3116. Reviewed-by:
-
Richard Levitte authored
Bug uncovered by test [extended tests] Reviewed-by:
-
Matt Caswell authored
Credit to OSS-Fuzz for finding this. Reviewed-by:
-
Richard Levitte authored
This increases portability of SSL_SESSION files between architectures where the size of |long| may vary. Before this, SSL_SESSION files produced on a 64-bit long architecture may break on a 32-bit long architecture. Reviewed-by:
-
Richard Levitte authored
Reviewed-by:
-
Richard Levitte authored
Also Z varieties. Reviewed-by:
-
- Apr 03, 2017
-
-
Matt Caswell authored
dhparams correctly handles X9.42 params in PEM format. However it failed to correctly processes them when reading/writing DER format. Fixes #3102 Reviewed-by:
-
Matt Caswell authored
DHparams has d2i_DHparams_fp, d2i_DHxparams_bio etc, but the equivalent macros for DHxparams were omitted. Reviewed-by:
-
- Apr 02, 2017
-
-
Andy Polyakov authored
It seems to be problematic to probe processor capabilities with SIGILL on MacOS X. The problem should be limited to cases when application code is debugged, but crashes were reported even during normal execution... Reviewed-by:
-
- Apr 01, 2017
-
-
Andy Polyakov authored
Configure started with 'require 5.10.0', but if executed by older perl it failed with "might be runaway multi-line // string" instead of naturally expected "Perl v5.10.0 required--this is only v5.x.y". Reviewed-by:
-
Andy Polyakov authored
Reviewed-by:
-
- Mar 31, 2017
-
-
Sebastian Andrzej Siewior authored
The man pages for ciphers, s_client and s_server mention the tls1_3 option but it is not implemented in the 1.1.0 branch. Thus remove it to avoid confusion. Signed-off-by:
Sebastian Andrzej Siewior <sebastian@breakpoint.cc> Reviewed-by:
-
Jon Spillett authored
Reviewed-by:
-
- Mar 30, 2017
-
-
FdaSilvaYY authored
Backport of 69687aa8 (Merged from #3069) Reviewed-by:
-
- Mar 29, 2017
-
-
Andy Polyakov authored
Reviewed-by:
-
Andy Polyakov authored
Reviewed-by:
-
Andy Polyakov authored
Reviewed-by:
-
Andy Polyakov authored
Reviewed-by:
-
Steven Collison authored
These ciphers don't appear to be documented anywhere. Given the performance[1] benefits I think it makes sense to expose them. [1] https://software.intel.com/sites/default/files/open-ssl-performance-paper.pdf Reviewed-by:
-
Steven Collison authored
These were added to the help in ad775e04 but not the pods. Reviewed-by:
-
Jon Spillett authored
Reviewed-by:
-
Rich Salz authored
Reviewed-by:
-
- Mar 28, 2017
-
-
FdaSilvaYY authored
Reviewed-by:
-
