Richard Levitte [Wed, 4 May 2016 12:44:10 +0000 (14:44 +0200)]
Check return of PEM_write_* functions and report possible errors
Reviewed-by: Matt Caswell <matt@openssl.org>
(Merged from https://github.com/openssl/openssl/pull/1025)
(cherry picked from commit c73aa309049c4f04ec81f0f1cf552eab8456a16e)
The name length limit check in x509_name_ex_d2i() includes
the containing structure as well as the actual X509_NAME. This will
cause large CRLs to be rejected.
Fix by limiting the length passed to ASN1_item_ex_d2i() which will
then return an error if the passed X509_NAME exceeds the length.
Only treat an ASN1_ANY type as an integer if it has the V_ASN1_INTEGER
tag: V_ASN1_NEG_INTEGER is an internal only value which is never used
for on the wire encoding.
Thanks to David Benjamin <davidben@google.com> for reporting this bug.
Matt Caswell [Mon, 25 Apr 2016 08:06:29 +0000 (09:06 +0100)]
Ensure EVP_EncodeUpdate handles an output length that is too long
With the EVP_EncodeUpdate function it is the caller's responsibility to
determine how big the output buffer should be. The function writes the
amount actually used to |*outl|. However this could go negative with a
sufficiently large value for |inl|. We add a check for this error
condition.
Reviewed-by: Richard Levitte <levitte@openssl.org>
Matt Caswell [Fri, 4 Mar 2016 10:17:17 +0000 (10:17 +0000)]
Avoid overflow in EVP_EncodeUpdate
An overflow can occur in the EVP_EncodeUpdate function which is used for
Base64 encoding of binary data. If an attacker is able to supply very large
amounts of input data then a length check can overflow resulting in a heap
corruption. Due to the very large amounts of data involved this will most
likely result in a crash.
Internally to OpenSSL the EVP_EncodeUpdate function is primarly used by the
PEM_write_bio* family of functions. These are mainly used within the
OpenSSL command line applications, so any application which processes
data from an untrusted source and outputs it as a PEM file should be
considered vulnerable to this issue.
User applications that call these APIs directly with large amounts of
untrusted data may also be vulnerable.
Issue reported by Guido Vranken.
CVE-2016-2105
Reviewed-by: Richard Levitte <levitte@openssl.org>
Matt Caswell [Thu, 28 Apr 2016 09:46:55 +0000 (10:46 +0100)]
Prevent EBCDIC overread for very long strings
ASN1 Strings that are over 1024 bytes can cause an overread in
applications using the X509_NAME_oneline() function on EBCDIC systems.
This could result in arbitrary stack data being returned in the buffer.
Matt Caswell [Thu, 3 Mar 2016 23:36:23 +0000 (23:36 +0000)]
Fix encrypt overflow
An overflow can occur in the EVP_EncryptUpdate function. If an attacker is
able to supply very large amounts of input data after a previous call to
EVP_EncryptUpdate with a partial block then a length check can overflow
resulting in a heap corruption.
Following an analysis of all OpenSSL internal usage of the
EVP_EncryptUpdate function all usage is one of two forms.
The first form is like this:
EVP_EncryptInit()
EVP_EncryptUpdate()
i.e. where the EVP_EncryptUpdate() call is known to be the first called
function after an EVP_EncryptInit(), and therefore that specific call
must be safe.
The second form is where the length passed to EVP_EncryptUpdate() can be
seen from the code to be some small value and therefore there is no
possibility of an overflow.
Since all instances are one of these two forms, I believe that there can
be no overflows in internal code due to this problem.
It should be noted that EVP_DecryptUpdate() can call EVP_EncryptUpdate()
in certain code paths. Also EVP_CipherUpdate() is a synonym for
EVP_EncryptUpdate(). Therefore I have checked all instances of these
calls too, and came to the same conclusion, i.e. there are no instances
in internal usage where an overflow could occur.
This could still represent a security issue for end user code that calls
this function directly.
TJ Saunders [Wed, 23 Mar 2016 18:55:53 +0000 (11:55 -0700)]
Issue #719:
If no serverinfo extension is found in some cases, do not abort the handshake,
but simply omit/skip that extension.
Check for already-registered serverinfo callbacks during serverinfo
registration.
Update SSL_CTX_use_serverinfo() documentation to mention the need to reload the
same serverinfo per certificate, for servers with multiple server certificates.
Reviewed-by: Matt Caswell <matt@openssl.org> Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Rich Salz <rsalz@openssl.org>
This adds an explicit limit to the size of an X509_NAME structure. Some
part of OpenSSL (e.g. TLS) already effectively limit the size due to
restrictions on certificate size.
The traditional private key encryption algorithm doesn't function
properly if the IV length of the cipher is zero. These ciphers
(e.g. ECB mode) are not suitable for private key encryption
anyway.
IBM argues that in certain scenarios capability query is really
expensive. At the same time it's asserted that query results can
be safely cached, because disabling CPACF is incompatible with
reboot-free operation.
Harden ASN.1 BIO handling of large amounts of data.
If the ASN.1 BIO is presented with a large length field read it in
chunks of increasing size checking for EOF on each read. This prevents
small files allocating excessive amounts of data.
CVE-2016-2109
Thanks to Brian Carpenter for reporting this issue.
Andy Polyakov [Thu, 31 Mar 2016 16:47:17 +0000 (18:47 +0200)]
PPC assembly pack: remove branch hints.
As it turns out branch hints grew as kind of a misconception. In
addition their interpretation by GNU assembler is affected by
assembler flags and can end up with opposite meaning on different
processors. As we have to loose quite a lot on misinterprerations,
especially on newer processors, we just omit them altogether.
David Benjamin [Mon, 14 Mar 2016 19:03:07 +0000 (15:03 -0400)]
Fix memory leak on invalid CertificateRequest.
Free up parsed X509_NAME structure if the CertificateRequest message
contains excess data.
The security impact is considered insignificant. This is a client side
only leak and a large number of connections to malicious servers would
be needed to have a significant impact.
This was found by libFuzzer.
Reviewed-by: Emilia Käsper <emilia@openssl.org> Reviewed-by: Stephen Henson <steve@openssl.org>
Todd Short [Sat, 5 Mar 2016 13:47:55 +0000 (08:47 -0500)]
Fix ALPN
* Perform ALPN after the SNI callback; the SSL_CTX may change due to
that processing
* Add flags to indicate that we actually sent ALPN, to properly error
out if unexpectedly received.
* document ALPN functions
* unit tests
Matt Caswell [Mon, 14 Mar 2016 17:06:19 +0000 (17:06 +0000)]
Fix a potential double free in EVP_DigestInit_ex
There is a potential double free in EVP_DigestInit_ex. This is believed
to be reached only as a result of programmer error - but we should fix it
anyway.
Richard Levitte [Wed, 9 Mar 2016 10:36:32 +0000 (11:36 +0100)]
Touch the correct variables for the system; shlib_wrap.sh on Solaris
If there is cause to think LD_LIBRARY_PATH_32 and LD_PRELOAD_32 are
appropriate variables to touch, do so. Otherwise, touch the usual
LD_LIBRARY_PATH and LD_PRELOAD. This covers for older installations
that don't have a mix of 32-bit and 64-bit libs.
Andy Polyakov [Fri, 4 Mar 2016 10:39:11 +0000 (11:39 +0100)]
bn/asm/x86[_64]-mont*.pl: complement alloca with page-walking.
Some OSes, *cough*-dows, insist on stack being "wired" to
physical memory in strictly sequential manner, i.e. if stack
allocation spans two pages, then reference to farmost one can
be punishable by SEGV. But page walking can do good even on
other OSes, because it guarantees that villain thread hits
the guard page before it can make damage to innocent one...
PVK files with abnormally large length or salt fields can cause an
integer overflow which can result in an OOB read and heap corruption.
However this is an rarely used format and private key files do not
normally come from untrusted sources the security implications not
significant.
Fix by limiting PVK length field to 100K and salt to 10K: these should be
more than enough to cover any files encountered in practice.
At the same time remove miniscule bias in final subtraction.
Performance penalty varies from platform to platform, and even with
key length. For rsa2048 sign it was observed to be 4% for Sandy
Bridge and 7% on Broadwell.
CVE-2016-0702
Reviewed-by: Richard Levitte <levitte@openssl.org> Reviewed-by: Rich Salz <rsalz@openssl.org>
(cherry picked from master)
Matt Caswell [Fri, 19 Feb 2016 11:38:25 +0000 (11:38 +0000)]
Add a test for SSLv2 configuration
SSLv2 should be off by default. You can only turn it on if you have called
SSL_CTX_clear_options(SSL_OP_NO_SSLv2) or
SSL_clear_options(SSL_OP_NO_SSLv2). You should not be able to inadvertantly
turn it on again via SSL_CONF without having done that first.
Viktor Dukhovni [Thu, 18 Feb 2016 02:07:48 +0000 (21:07 -0500)]
Disable SSLv2 default build, default negotiation and weak ciphers.
SSLv2 is by default disabled at build-time. Builds that are not
configured with "enable-ssl2" will not support SSLv2. Even if
"enable-ssl2" is used, users who want to negotiate SSLv2 via the
version-flexible SSLv23_method() will need to explicitly call either
of:
SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
or
SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
as appropriate. Even if either of those is used, or the application
explicitly uses the version-specific SSLv2_method() or its client
or server variants, SSLv2 ciphers vulnerable to exhaustive search
key recovery have been removed. Specifically, the SSLv2 40-bit
EXPORT ciphers, and SSLv2 56-bit DES are no longer available.
Matt Caswell [Mon, 22 Feb 2016 10:27:18 +0000 (10:27 +0000)]
Fix BN_hex2bn/BN_dec2bn NULL ptr/heap corruption
In the BN_hex2bn function the number of hex digits is calculated using
an int value |i|. Later |bn_expand| is called with a value of |i * 4|.
For large values of |i| this can result in |bn_expand| not allocating any
memory because |i * 4| is negative. This leaves ret->d as NULL leading
to a subsequent NULL ptr deref. For very large values of |i|, the
calculation |i * 4| could be a positive value smaller than |i|. In this
case memory is allocated to ret->d, but it is insufficiently sized
leading to heap corruption. A similar issue exists in BN_dec2bn.
This could have security consequences if BN_hex2bn/BN_dec2bn is ever
called by user applications with very large untrusted hex/dec data. This is
anticipated to be a rare occurrence.
All OpenSSL internal usage of this function uses data that is not expected
to be untrusted, e.g. config file data or application command line
arguments. If user developed applications generate config file data based
on untrusted data then it is possible that this could also lead to security
consequences. This is also anticipated to be a rare.
Matt Caswell [Thu, 25 Feb 2016 13:09:46 +0000 (13:09 +0000)]
Fix memory issues in BIO_*printf functions
The internal |fmtstr| function used in processing a "%s" format string
in the BIO_*printf functions could overflow while calculating the length
of a string and cause an OOB read when printing very long strings.
Additionally the internal |doapr_outch| function can attempt to write to
an OOB memory location (at an offset from the NULL pointer) in the event of
a memory allocation failure. In 1.0.2 and below this could be caused where
the size of a buffer to be allocated is greater than INT_MAX. E.g. this
could be in processing a very long "%s" format string. Memory leaks can also
occur.
These issues will only occur on certain platforms where sizeof(size_t) >
sizeof(int). E.g. many 64 bit systems. The first issue may mask the second
issue dependent on compiler behaviour.
These problems could enable attacks where large amounts of untrusted data
is passed to the BIO_*printf functions. If applications use these functions
in this way then they could be vulnerable. OpenSSL itself uses these
functions when printing out human-readable dumps of ASN.1 data. Therefore
applications that print this data could be vulnerable if the data is from
untrusted sources. OpenSSL command line applications could also be
vulnerable where they print out ASN.1 data, or if untrusted data is passed
as command line arguments.
Libssl is not considered directly vulnerable. Additionally certificates etc
received via remote connections via libssl are also unlikely to be able to
trigger these issues because of message size limits enforced within libssl.
Emilia Kasper [Wed, 24 Feb 2016 11:59:59 +0000 (12:59 +0100)]
CVE-2016-0798: avoid memory leak in SRP
The SRP user database lookup method SRP_VBASE_get_by_user had confusing
memory management semantics; the returned pointer was sometimes newly
allocated, and sometimes owned by the callee. The calling code has no
way of distinguishing these two cases.
Specifically, SRP servers that configure a secret seed to hide valid
login information are vulnerable to a memory leak: an attacker
connecting with an invalid username can cause a memory leak of around
300 bytes per connection.
Servers that do not configure SRP, or configure SRP but do not configure
a seed are not vulnerable.
In Apache, the seed directive is known as SSLSRPUnknownUserSeed.
To mitigate the memory leak, the seed handling in SRP_VBASE_get_by_user
is now disabled even if the user has configured a seed.
Applications are advised to migrate to SRP_VBASE_get1_by_user. However,
note that OpenSSL makes no strong guarantees about the
indistinguishability of valid and invalid logins. In particular,
computations are currently not carried out in constant time.
David Woodhouse [Tue, 23 Feb 2016 14:00:03 +0000 (14:00 +0000)]
RT4175: Fix PKCS7_verify() regression with Authenticode signatures
This is a partial revert of commit c8491de39 ("GH354: Memory leak fixes"),
which was cherry-picked from commit 55500ea7c in OpenSSL 1.1.
That commit introduced a change in behaviour which is a regression for
software implementing Microsoft Authenticode — which requires a PKCS#7
signature to be validated against explicit external data, even though
it's a non-detached signature with its own embedded data.
The is fixed differently in OpenSSL 1.1 by commit 6b2ebe433 ("Add
PKCS7_NO_DUAL_CONTENT flag"), but that approach isn't viable in the
1.0.2 stable branch, so just comment the offending check back out again.
Signed-off-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Tim Hudson <tjh@openssl.org>
Corinna Vinschen [Sat, 16 Jan 2016 18:36:15 +0000 (19:36 +0100)]
Don't strip object files on Cygwin
Building for the Cygwin distro requires to be able to build debuginfo
files. This in turn requires to build object files without stripping.
The stripping is performed by the next step after building which creates
the debuginfo files.
Andy Polyakov [Wed, 3 Feb 2016 17:21:00 +0000 (18:21 +0100)]
util/mk1mf.pl: use LINK_CMD instead of LINK variable.
Trouble is that LINK variable assignment in make-file interferes with
LINK environment variable, which can be used to modify Microsoft's
LINK.EXE behaviour.
Matt Caswell [Tue, 2 Feb 2016 10:05:43 +0000 (10:05 +0000)]
Handle SSL_shutdown while in init more appropriately #2
Previous commit f73c737c7 attempted to "fix" a problem with the way
SSL_shutdown() behaved whilst in mid-handshake. The original behaviour had
SSL_shutdown() return immediately having taken no action if called mid-
handshake with a return value of 1 (meaning everything was shutdown
successfully). In fact the shutdown has not been successful.
Commit f73c737c7 changed that to send a close_notify anyway and then
return. This seems to be causing some problems for some applications so
perhaps a better (much simpler) approach is revert to the previous
behaviour (no attempt at a shutdown), but return -1 (meaning the shutdown
was not successful).
This also fixes a bug where SSL_shutdown always returns 0 when shutdown
*very* early in the handshake (i.e. we are still using SSLv23_method).
Hubert Kario [Mon, 1 Feb 2016 16:14:34 +0000 (11:14 -0500)]
GH554: Improve pkeyutl doc
Signed-off-by: Rich Salz <rsalz@openssl.org> Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 53619f9f40740ae8b256998574758aa191635db8)
Signed-off-by: Rich Salz <rsalz@akamai.com> Reviewed-by: Dr. Stephen Henson <steve@openssl.org>
(cherry picked from commit 8ab31975bacb9c907261088937d3aa4102e3af84)
Matt Caswell [Wed, 4 Nov 2015 17:30:22 +0000 (17:30 +0000)]
Fix bug in nistp224/256/521 where have_precompute_mult always returns 0
During precomputation if the group given is well known then we memcpy a
well known precomputation. However we go the wrong label in the code and
don't store the data properly. Consequently if we call have_precompute_mult
the data isn't there and we return 0.