gcc 8.3.0 was apparently clever enough to optimize away the previously
used os_memset() to explicitly clear a stack buffer that contains keys
when that clearing happened just before returning from the function.
Since memset_s() is not exactly portable (or commonly available yet..),
use a less robust mechanism that is still pretty likely to prevent
current compilers from optimizing the explicit clearing of the memory
away.
Signed-off-by: Jouni Malinen <j@w1.fi>
Commit 4b16c15bbc ("EAP-pwd server: Use os_get_random() for
unpredictable token") replaced use of os_random(), i.e., of random(),
with os_get_random(), but forgot to remove the now unused srandom()
call. Clean up the implementation and remove that unneeded code.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
data->inbuf allocation might fail and if that were to happen, the next
fragment in the exchange could have resulted in NULL pointer
dereference. Unexpected fragment with more bit might also be able to
trigger this. Fix that by explicitly checking for data->inbuf to be
available before using it.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
None of the ECC groups supported in the implementation had a cofactor
greater than 1, so these checks are unreachable and for all cases, the
cofactor is known to be 1. Furthermore, RFC 5931 explicitly disallow use
of ECC groups with cofactor larger than 1, so this checks cannot be
needed for any curve that is compliant with the RFC.
Remove the unneeded group cofactor checks to simplify the
implementation.
Signed-off-by: Jouni Malinen <j@w1.fi>
Binary presentations of element and scalar can be written directly to
the allocated commit message buffer instead of having to first write
them into temporary buffers just to copy them to the actual message
buffer.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
RFC 5931 has these conditions as MUST requirements, so better follow
them explicitly even if the rand,mask == 0 or rand+mask == 0 or 1 cases
are very unlikely to occur in practice while generating random values
locally.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This adds an explicit check for 0 < x,y < prime based on RFC 5931,
2.8.5.2.2 requirement. The earlier checks might have covered this
implicitly, but it is safer to avoid any dependency on implicit checks
and specific crypto library behavior. (CVE-2019-9498 and CVE-2019-9499)
Furthermore, this moves the EAP-pwd element and scalar parsing and
validation steps into shared helper functions so that there is no need
to maintain two separate copies of this common functionality between the
server and peer implementations.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
When processing an EAP-pwd Commit frame, verify that the peer's scalar
and elliptic curve element differ from the one sent by the server. This
prevents reflection attacks where the adversary reflects the scalar and
element sent by the server. (CVE-2019-9497)
The vulnerability allows an adversary to complete the EAP-pwd handshake
as any user. However, the adversary does not learn the negotiated
session key, meaning the subsequent 4-way handshake would fail. As a
result, this cannot be abused to bypass authentication unless EAP-pwd is
used in non-WLAN cases without any following key exchange that would
require the attacker to learn the MSK.
Signed-off-by: Mathy Vanhoef <mathy.vanhoef@nyu.edu>
When processing an EAP-pwd Commit frame, the peer's scalar and element
(elliptic curve point) were not validated. This allowed an adversary to
bypass authentication, and impersonate any user if the crypto
implementation did not verify the validity of the EC point.
Fix this vulnerability by assuring the received scalar lies within the
valid range, and by checking that the received element is not the point
at infinity and lies on the elliptic curve being used. (CVE-2019-9498)
The vulnerability is only exploitable if OpenSSL version 1.0.2 or lower
is used, or if LibreSSL or wolfssl is used. Newer versions of OpenSSL
(and also BoringSSL) implicitly validate the elliptic curve point in
EC_POINT_set_affine_coordinates_GFp(), preventing the attack.
Signed-off-by: Mathy Vanhoef <mathy.vanhoef@nyu.edu>
These changes add support for salted password databases to EAP-pwd per
RFC 8146. This commits introduces the framework for enabling this and
the salting mechanisms based on SHA-1, SHA256, and SHA512 hash
algorithms.
Signed-off-by: Dan Harkins <dharkins@lounge.org>
This leads to cleaner code overall, and also reduces the size
of the hostapd and wpa_supplicant binaries (in hwsim test build
on x86_64) by about 2.5 and 3.5KiB respectively.
The mechanical conversions all over the code were done with
the following spatch:
@@
expression SIZE, SRC;
expression a;
@@
-a = os_malloc(SIZE);
+a = os_memdup(SRC, SIZE);
<...
if (!a) {...}
...>
-os_memcpy(a, SRC, SIZE);
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
RFC 5931 Section 2.8.5.1 does not list the Prep field as something that
the server validates to match the Request. However, the supplicant side
has to use the same pre-processing mechanism for the password for the
authentication to work, so we may as well as enforce this field to match
the requested value now that wpa_supplicant implementation is fixed to
copy the value from the request.
Signed-off-by: Jouni Malinen <j@w1.fi>
Do not use os_random() that uses a low quality PRNG to generate the
anti-clogging token. The construction can be improved upon by replacing
it with a call to os_get_random(), which uses a high quality PRNG. While
the RFC 5931 explictly recommends not to do this ("SHOULD NOT be from a
source of random entropy"), it does still mandate unpredicability ("MUST
be unpredictable"). The anti-clogging token is most unpredictable when
it is taken from a high quality PRNG.
Signed-off-by: Nick Lowe <nick.lowe@lugatech.com>
Free the allocated structure in error cases to remove need for each EAP
method to handle the error cases separately. Each registration function
can simply do "return eap_server_method_register(eap);" in the end of
the function.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
All but the last fragment had their length checked against the remaining
room in the reassembly buffer. This allowed a suitably constructed last
fragment frame to try to add extra data that would go beyond the buffer.
The length validation code in wpabuf_put_data() prevents an actual
buffer write overflow from occurring, but this results in process
termination. (CVE-2015-5314)
Signed-off-by: Jouni Malinen <j@w1.fi>
The cleanup code will handle this, but it is more robust to make sure
this is cleared to zero when allocating a new buffer.
Signed-off-by: Jouni Malinen <j@w1.fi>
The remaining number of bytes in the message could be smaller than the
Total-Length field size, so the length needs to be explicitly checked
prior to reading the field and decrementing the len variable. This could
have resulted in the remaining length becoming negative and interpreted
as a huge positive integer.
In addition, check that there is no already started fragment in progress
before allocating a new buffer for reassembling fragments. This avoid a
potential memory leak when processing invalid message.
Signed-off-by: Jouni Malinen <j@w1.fi>
The length of the received Commit and Confirm message payloads was not
checked before reading them. This could result in a buffer read
overflow when processing an invalid message.
Fix this by verifying that the payload is of expected length before
processing it. In addition, enforce correct state transition sequence to
make sure there is no unexpected behavior if receiving a Commit/Confirm
message before the previous exchanges have been completed.
Thanks to Kostya Kortchinsky of Google security team for discovering and
reporting this issue.
Signed-off-by: Jouni Malinen <j@w1.fi>
This extends EAP-pwd server support to allow NtHash version of password
storage in addition to full plaintext password.
Signed-off-by: Jouni Malinen <j@w1.fi>
This gets registered in tls_openssl.c from tls_init(), so there is no
need for EAP-pwd implementation to register explicitly. This avoids some
corner cases where OpenSSL resources do not get fully freed on exit.
Signed-off-by: Jouni Malinen <j@w1.fi>
Use an explicit memset call to clear any configuration parameter and
dynamic data that contains private information like keys or identity.
This brings in an additional layer of protection by reducing the length
of time this type of private data is kept in memory.
Signed-off-by: Jouni Malinen <j@w1.fi>
This makes the EAP-pwd server and peer implementations more robust
should OpenSSL fail to derive random number for some reason. While this
is unlikely to happen in practice, the implementation better be prepared
for this should something unexpected ever happen. See
http://jbp.io/2014/01/16/openssl-rand-api/#review-of-randbytes-callers
for more details.
Signed-off-by: Florent Daigniere <nextgens@freenetproject.org>
This makes the implementation less likely to provide useful timing
information to potential attackers from comparisons of information
received from a remote device and private material known only by the
authorized devices.
Signed-off-by: Florent Daigniere <nextgens@freenetproject.org>
This changes OpenSSL calls to explicitly clear the EC_POINT memory
allocations when freeing them. This adds an extra layer of security by
avoiding leaving potentially private keys into local memory after they
are not needed anymore. While some of these variables are not really
private (e.g., they are sent in clear anyway), the extra cost of
clearing them is not significant and it is simpler to just clear these
explicitly rather than review each possible code path to confirm where
this does not help.
Signed-off-by: Florent Daigniere <nextgens@freenetproject.org>
This changes OpenSSL calls to explicitly clear the bignum memory
allocations when freeing them. This adds an extra layer of security by
avoiding leaving potentially private keys into local memory after they
are not needed anymore. While some of these variables are not really
private (e.g., they are sent in clear anyway), the extra cost of
clearing them is not significant and it is simpler to just clear these
explicitly rather than review each possible code path to confirm where
this does not help.
Signed-off-by: Florent Daigniere <nextgens@freenetproject.org>
Reduce the amount of time keying material (MSK, EMSK, temporary private
data) remains in memory in EAP methods. This provides additional
protection should there be any issues that could expose process memory
to external observers.
Signed-off-by: Jouni Malinen <j@w1.fi>
Instead of using implicit limit based on 16-bit unsigned integer having
a maximum value of 65535, limit the maximum length of a fragmented
EAP-pwd message explicitly to 15000 bytes. None of the supported groups
use longer messages, so it is fine to reject any longer message without
even trying to reassemble it. This will hopefully also help in reducing
false warnings from static analyzers (CID 68124).
Signed-off-by: Jouni Malinen <j@w1.fi>
If invalid group was negotiated, compute_password_element() left some of
the data->grp pointer uninitialized and this could result in
segmentation fault when deinitializing the EAP method. Fix this by
explicitly clearing all the pointer with eap_zalloc(). In addition,
speed up EAP failure reporting in this type of error case by indicating
that the EAP method execution cannot continue anymore on the peer side
instead of waiting for a timeout.
Signed-off-by: Jouni Malinen <j@w1.fi>
If fragmentation is used, the temporary inbuf/outbuf could have been
leaked in error cases (e.g., reaching maximum number of roundtrips).
Signed-off-by: Jouni Malinen <j@w1.fi>
Remove the GPL notification text from EAP-pwd implementation per
approval from Dan Harkins who contributed these files.
(email from Dan Harkins <dharkins@lounge.org> dated
Wed, 4 Jan 2012 16:25:48 -0800)
Signed-hostap: Jouni Malinen <j@w1.fi>
This structure was not really used for anything apart from figuring out
length of the EAP-pwd header (and even that in a way that would not work
with fragmentation). Since the bitfields in the structure could have
been problematic depending on target endianness, remove this unnecessary
structure.
Signed-hostap: Jouni Malinen <j@w1.fi>
Another niceness of OpenSSL is that if the high-order bit of a 521-bit
big num is not set then BN_bn2bin() will just return 65 bytes instead of
66 bytes with the 1st (big endian, after all) being all zero. When this
happens the wrong number of octets are mixed into function H(). So
there's a whole bunch of "offset" computations and BN_bn2bin() dumps the
big number into a buffer + offset. That should be obvious in the patch
too.
The changes are:
1. the word "and" in the hunting-and-pecking string passed to the KDF
should be capitalized.
2. the primebitlen used in the KDF should be a short not an int.
3. the computation of MK in hostap is based on an older version of the
draft and is not the way it's specified in the RFC.
4. the group being passed into computation of the Commit was not in
network order.
