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>
Move the identical function used by both SAE and EAP-pwd to
src/utils/common.c to avoid duplicated implementation.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
These functions can be used to help implement constant time operations
for various cryptographic operations that must minimize externally
observable differences in processing (both in timing and also in
internal cache use, etc.).
This is related to CVE-2019-9494 and CVE-2019-9495.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Merge the practically copy-pasted implementations in wpa_supplicant and
hostapd into a single shared implementation in dpp.c for managing
configurator and boostrapping information. This avoid unnecessary code
duplication and provides a convenient location for adding new global DPP
data.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Add a constraint on the base64 encoded buffer length to avoid an integer
overflow in the output length calculation.
common.c:1087:16: runtime error: unsigned integer overflow: 0 - 1 cannot be represented in type 'size_t' (aka 'unsigned long')
Signed-off-by: Jouni Malinen <j@w1.fi>
Split the if/while loop condition into two independent steps so that
in_size-- happens only in the case in_size is nonzero. This gets rid of
unnecessary UBSan warnings.
common.c:1087:16: runtime error: unsigned integer overflow: 0 - 1 cannot be represented in type 'size_t' (aka 'unsigned long')
common.c:1076:16: runtime error: unsigned integer overflow: 0 - 1 cannot be represented in type 'size_t' (aka 'unsigned long')
common.c:1119:16: runtime error: unsigned integer overflow: 0 - 1 cannot be represented in type 'size_t' (aka 'unsigned long')
Signed-off-by: Jouni Malinen <j@w1.fi>
This results in an UBSan warning that can be avoided easily.
os_unix.c:524:3: runtime error: null pointer passed as argument 2, which is declared to never be null
Signed-off-by: Jouni Malinen <j@w1.fi>
UBSan testing with WPA_TRACE=y ended up hitting an unaligned access for
struct os_alloc_trace in os_program_deinit() because of the
dl_list_for_each() design that looked like dereferencing the member
element of the list head which is something that does not exist.
Get the first entry from the list using dl_list_first() so that the
empty list special case is covefred and compare item pointers instead of
struct dl_list pointers to check whether the end of the loop has been
reached.
Signed-off-by: Jouni Malinen <j@w1.fi>
Add test-tls program that can be used for fuzzing the internal TLS
client and server implementations. This tool can write client or server
messages into a file as an initialization step and for the fuzzing step,
that file (with potential modifications) can be used to replace the
internally generated message contents.
The TEST_FUZZ=y build parameter is used to make a special build where a
hardcoded random number generator and hardcoded timestamp are used to
force deterministic behavior for the TLS operations.
Signed-off-by: Jouni Malinen <j@w1.fi>
EV_SET() for EV_ADD used a specific filter type, but that same filter
type was not provided to the matching EV_DELETE case. This resulted in
the kernel rejecting the deletion with "Invalid argument". Fix this by
setting the same filter type for both operations.
Fixes: f9982b3212 ("Implement kqueue(2) support via CONFIG_ELOOP_KQUEUE")
Signed-off-by: Jouni Malinen <j@w1.fi>
The previous implementation did not work if the first registered socket
had fd > 16 or if the fd was more than double the largest value used in
previous registrations. Those cases could result in too small a memory
allocation being used and writes/reads beyond the end of that buffer.
This fix is applicable to CONFIG_ELOOP_EPOLL=y and CONFIG_ELOOP_KQUEUE=y
builds.
Fixes: f0356ec85c ("eloop: Add epoll option for better performance")
Signed-off-by: Jouni Malinen <j@w1.fi>
The SSL_METHOD patching hack to get proper OCSP validation for Hotspot
2.0 OSU needs cannot be used with OpenSSL 1.1.0 and newer since the
SSL_METHOD structure is not exposed anymore. Fall back to using the
incomplete CURLOPT_SSL_VERIFYSTATUS design to fix the build.
Signed-off-by: Ben Greear <greearb@candelatech.com>
SKM_sk_num() is not available anymore, so use DEFINE_STACK_OF() to get
the appropriate accessor functions.
Signed-off-by: Ben Greear <greearb@candelatech.com>
The direct ssl->ctx access are not allowed anymore in newer OpenSSL
versions, so use the SSL_get_SSL_CTX() helper for this.
Signed-off-by: Ben Greear <greearb@candelatech.com>
When syslog logging is used output from wpa_hexdump_ascii() was silently
discarded. This patch enables wpa_hexdump_ascii() to print data to
syslog but without ASCII decoding.
Signed-off-by: Andrey Kartashev <andrey.kartashev@afconsult.com>
On Linux this flag will make sure that no file descriptor is
accidentally leaked into potential child processes. While this is not a
problem right now, it is considered to be good practice these days when
dealing with file descriptors on Linux.
Signed-off-by: Karol Babioch <karol@babioch.de>
Previously the file permissions for the debug log file were not
explicitly set. Instead it was implicitly relying on a secure umask,
which in most cases would result in a file that is world-readable. This
is a violation of good practices, since not every user should have
access to sensitive information that might be contained in the debug log
file.
Explicitly set sane default file permissions in case the file is newly
created.
Unfortunately the fopen(3) function does not provide such a facility, so
the approach needs to be changed in the following way:
1) The file descriptor needs to be created manually using the open(3)
function with the correct flags and the desired mode set.
2) fdopen(3) can then be used on the file descriptor to associate a file
stream with it.
Note: This modification will not change the file permissions of any
already existing debug log files, and only applies to newly created
ones.
Signed-off-by: Karol Babioch <karol@babioch.de>
With new restriction in Android, if PATH env variable doesn't have
correct path of 'am' binary, execv() fails to launch wpadebug browser
(am starts, but something seems to fail within its internal processing).
This commit is a workaround to use execve() with custom environment PATH
which includes "/system/bin;/vendor/bin" to handle the cases where
hs20-osu-client fails to launch wpadebug browser through /system/bin/am.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
This fixes some validation issues against DM_ddf DTD that were caused by
the conversion from the internal tree structure to TNDS. Only the leaf
nodes are supposed to have the Value node.
Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
If the second json_alloc_token() call failed to allocate memory,
json_parse() missed the first allocation on the error path. Assign the
root pointer earlier for that case to avoid the potential memory leak.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Undefine the __bitwise macro before defining it to avoid conflicts
with the one from linux/types.h; the same is done some lines above
when __CHECKER__ is defined. Fixes the following warning:
In file included from ../src/l2_packet/l2_packet_linux.c:15:0:
hostap/src/utils/common.h:438:0: warning: "__bitwise" redefined
#define __bitwise
In file included from /usr/include/linux/filter.h:9:0,
from ../src/l2_packet/l2_packet_linux.c:13:
/usr/include/linux/types.h:21:0: note: this is the location of the previous definition
#define __bitwise __bitwise__
Signed-off-by: Beniamino Galvani <bgalvani@redhat.com>
Use the SSL_get_SSL_CTX() helper instead of dereferencing SSL* since
struct ssl_st is not exposed in public header files anymore.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
The previous implementation was able to parse arrays of objects, but not
arrays of other types of items.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Introduction of the new base64 helper function changed the backtraces
for these OOM test cases and resulted in test failures. Update the test
scripts to work with the new implementation.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Previously, only SHA1 hash -based server certificate matching was used,
but the OCSP response may use SHA256 instead of SHA1, so check the match
with both hash functions, if needed.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Replace the previously used extension of IEEE 802.11 managed Ethertype
89-0d (originally added for Remote Request/Response in IEEE 802.11r)
with Ethertype 88-b7 (OUI Extended EtherType) for FT inter-AP
communication. The new design uses a more properly assigned identifier
for the messages.
This assigns the OUI 00:13:74 vendor-specific subtype 0x0001 for the new
hostapd AP-to-AP communication purposes. Subtypes 1 (PULL), 2 (RESP),
and 3 (PUSH) are also assigned in this commit for the R0KH-R1KH
protocol.
This breaks backward compatibility, i.e., hostapd needs to be updated on
all APs at the same time to allow FT to remain functional.
Signed-off-by: Michael Braun <michael-dev@fami-braun.de>
If the uuid configuration parameter is not set, wpa_supplicant generates
an UUID automatically to allow WPS operations to proceed. This was
previously always using an UUID generated from the MAC address. This
commit adds an option to use a random UUID instead. The type of the
automatically generated UUID is set with the auto_uuid parameter: 0 =
based on MAC address (default; old behavior), 1 = random UUID.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Previously, the Value node was not added if value of a node could not be
fetched. This can cause interoperability issues, so address that in the
same way as an empty length value, i.e., by adding a Value node with
zero-length contents.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
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>
This can be used to clean the code and reduce size by converting
os_malloc() followed by os_memcpy() cases to use a single function call.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
When ASLR is enabled, like it is by default on many distros now,
the trace code doesn't work right.
Fix this by looking up the start of the executable mapping and
subtracing it from all the lookups.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
HTTP response was previously sent as a plaintext without the HTTP header
on port 12345. By default Android webview/Chrome assumes plaintext as
HTTP/0.9 data. Android webview/Chrome has removed support of HTTP/0.9
request/response on non-standard ports, i.e., other than port 80. This
results in error while opening URL 'http://localhost:12345/'.
Fix this by prefixing the HTTP response with the HTTP/1.1 header.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
We can only send module specific messages to syslog and not debug
messages printed with wpa_printf. Add an extra command line parameter
'-s' to allow it. The feature is enabled with compile flag
CONFIG_DEBUG_SYSLOG as for wpa_supplicant and behaves in the same manner
as the wpa_supplicant -s command line argument.
Signed-off-by: Wojciech Dubowik <Wojciech.Dubowik@neratec.com>
This conversion will be done several times in the code, so add a helper
function that does this conversion.
Signed-off-by: Avrahams Stern <avraham.stern@intel.com>
