Reorder terms in a way that no invalid pointers are generated with
pos+len operations. end-pos is always defined (with a valid pos pointer)
while pos+len could end up pointing beyond the end pointer which would
be undefined behavior.
Signed-off-by: Jouni Malinen <j@w1.fi>
All the other web_connection_parse_*() functions were already doing
this, so make the GET handler consistent as well.
Signed-off-by: Jouni Malinen <j@w1.fi>
Use shorter variables for storing the attribute lengths and group these
variables together to allow compiler to pack them more efficiently. This
reduces the struct size from 960 bytes to 760 bytes in 64-bit builds.
This reduces stack use in number of functions.
Signed-off-by: Jouni Malinen <j@w1.fi>
There is no need to maintain two concurrent instances of struct
wps_parse_attr in this function. Share a single structure for parsing
both IEs.
Signed-off-by: Jouni Malinen <j@w1.fi>
This is needed to allow new operation to be started after an error
without having to wait for the AP entry to time out.
Signed-off-by: Jouni Malinen <j@w1.fi>
No need to have a common failure handler if it is used from only a
single location and that lcoation does not even need the memory freeing
step.
Signed-off-by: Jouni Malinen <j@w1.fi>
The hbp pointer is moved to the next space already earlier in this code
path, so the while loop here did not really do anything.
Signed-off-by: Jouni Malinen <j@w1.fi>
Commit 7da4f4b499 ('WPS: Check maximum
HTTP body length earlier in the process') added too strict check for
body length allocation. The comparison of new_alloc_nbytes against
h->max_bytes did not take into account that HTTPREAD_BODYBUF_DELTA was
added to previous allocation even if that ended up going beyond
h->max_bytes. This ended up rejecting some valid HTTP operations, e.g.,
when checking AP response to WPS ER setting selected registrar.
Fix this by taking HTTPREAD_BODYBUF_DELTA into account.
Signed-off-by: Jouni Malinen <j@w1.fi>
Incorrect number of bytes was skipped from the beginning of the line
which resulted in the loop skipping spaces doing nothing. However, the
following operation was simply looking for the max-age parameter with
os_strstr(), so this did not have any effect on functionality. Fix the
number of bytes to skip and remove the unneeded loop to skip spaces.
Signed-off-by: Jouni Malinen <j@w1.fi>
This is similar to the earlier commit
b363121a20 ('WPS: Reject invalid
credential more cleanly'), but for the AP cases where AP settings are
being replaced. Previously, the new settings were taken into use even if
the invalid PSK/passphrase had to be removed. Now, the settings are
rejected with such an invalid configuration.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Use a local variable and check the record payload length validity before
writing it into record->payload_length in hopes of getting rid of a
bogus static analyzer warning. The negative return value was sufficient
to avoid record->payload_length being used, but that seems to be too
complex for some analyzers. (CID 122668)
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
While this is already enforced in practice due to the limits on the
maximum control interface command length and total_length bounds
checking here, this explicit check on payload_length value may help
static analyzers understand the code better. (CID 122668)
Signed-off-by: Jouni Malinen <j@w1.fi>
It was possible for the 32-bit record->total_length value to end up
wrapping around due to integer overflow if the longer form of payload
length field is used and record->payload_length gets a value close to
2^32. This could result in ndef_parse_record() accepting a too large
payload length value and the record type filter reading up to about 20
bytes beyond the end of the buffer and potentially killing the process.
This could also result in an attempt to allocate close to 2^32 bytes of
heap memory and if that were to succeed, a buffer read overflow of the
same length which would most likely result in the process termination.
In case of record->total_length ending up getting the value 0, there
would be no buffer read overflow, but record parsing would result in an
infinite loop in ndef_parse_records().
Any of these error cases could potentially be used for denial of service
attacks over NFC by using a malformed NDEF record on an NFC Tag or
sending them during NFC connection handover if the application providing
the NDEF message to hostapd/wpa_supplicant did no validation of the
received records. While such validation is likely done in the NFC stack
that needs to parse the NFC messages before further processing,
hostapd/wpa_supplicant better be prepared for any data being included
here.
Fix this by validating record->payload_length value in a way that
detects integer overflow. (CID 122668)
Signed-off-by: Jouni Malinen <j@w1.fi>
The 32-bit version of payload length field may not be 32-bit aligned in
the message buffer, so use WPA_GET_BE32() to read it instead of ntohl().
Signed-off-by: Jouni Malinen <j@w1.fi>
The debug information from httpread can be helpful in figuring out error
cases in general and as such, should be enabled by default. Get rid of
the hardcoded httpread_debug value that would require source code
changes to enable.
Signed-off-by: Jouni Malinen <j@w1.fi>
There is no need to continue processing a HTTP body when it becomes
clear that the end result would be over the maximum length.
Signed-off-by: Jouni Malinen <j@w1.fi>
Verify that ncopy parameter to memcpy is not negative. While this is not
supposed to be needed, it is a good additional protection against
unknown implementation issues.
Signed-off-by: Jouni Malinen <j@w1.fi>
strtoul() return value may end up overflowing the int h->chunk_size and
resulting in a negative value to be stored as the chunk_size. This could
result in the following memcpy operation using a very large length
argument which would result in a buffer overflow and segmentation fault.
This could have been used to cause a denial service by any device that
has been authorized for network access (either wireless or wired). This
would affect both the WPS UPnP functionality in a WPS AP (hostapd with
upnp_iface parameter set in the configuration) and WPS ER
(wpa_supplicant with WPS_ER_START control interface command used).
Validate the parsed chunk length value to avoid this. In addition to
rejecting negative values, we can also reject chunk size that would be
larger than the maximum configured body length.
Thanks to Kostya Kortchinsky of Google security team for discovering and
reporting this issue.
Signed-off-by: Jouni Malinen <j@w1.fi>
Handling of WPS RF band for 60 GHz was missing. Add it in all relevant
places and also map "AES" as the cipher to GCMP instead of CCMP when
operating on the 60 GHz band.
Signed-off-by: Hamad Kadmany <qca_hkadmany@qca.qualcomm.com>
By analysing objdump output some read only structures were found in
.data section. To help compiler further optimize code declare these
as const.
Signed-off-by: Mikael Kanstrup <mikael.kanstrup@sonymobile.com>
There is no need to try to derive DH shared key with a peer that tries
to use too short or too long DH Public Key. Previously, such cases ended
up implicitly getting rejected by the DH operations failing to produce
matching results. That is unnecessarily, so simply reject the message
completely if it does not have a Public Key with valid length. Accept
couple of octets shorter value to be used to avoid interoperability
issues if there are implementations that do not use zero-padding
properly.
Signed-off-by: Jouni Malinen <j@w1.fi>
This enforces variable length strings Manufacturer, Model Name, Model
Number, and Serial Number to be within the maximum length defined in the
WSC specification. While none of the existing users for these within
hostapd/wpa_supplicant had problems with longer strings, it is good to
ensure the strings are not longer to avoid potential issues at higher
layer components.
Signed-off-by: Jouni Malinen <j@w1.fi>
This program can be used to run fuzzing tests for areas related to P2P
message parsing and processing. p2p-fuzzer allows data files to be used
to inject Probe Response and Action frames for processing by the P2P
module.
Signed-off-by: Jouni Malinen <j@w1.fi>
This modifies couple of code segments that replaced control characters
in strings with '_' to use a common helper function.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
While it looks like all the users of this parsed attribute were able to
handle longer SSID values, there is no valid use case for these and to
avoid any potential future issues, enforce maximum length (32 bytes) on
the SSID during parsing.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
While it looks like all the users of this parsed attribute were able to
handle longer Device Name values, there is no valid use case for these
and to avoid any potential issues in upper layer components, enforce
maximum length (32 bytes) on the Device Name during parsing.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This provides additional WPS definitions and rules for negotiating use
of P2PS default PIN configuration method.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Commit ce7b56afab ('WPS: Fix an
interoperability issue with mixed mode and AP Settings') added code to
filter M7 Authentication/Encryption Type attributes into a single bit
value in mixed mode (WPA+WPA2) cases to work around issues with Windows
7. This workaround was lost in commit
d7a15d5953 ('WPS: Indicate current AP
settings in M7 in unconfigurated state') that fixed unconfigured state
values in AP Settings, but did not take into account the earlier
workaround for mixed mode.
Re-introduce filtering of Authentication/Encryption Type attributes for
M7 based on the current AP configuration. In other words, merge those
two earlier commits together to include both the earlier workaround the
newer fix.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
This helps hwsim test cases by avoiding undesired state from previously
executed test cases affecting following tests.
Signed-off-by: Jouni Malinen <j@w1.fi>
The preceding dl_list_len() check guarantees that dl_list_first()
returns an entry and not NULL. However, that seems to be a bit too
difficult path to follow for static analyzers, so add an explicit check
for the dl_list_first() return value to be non-NULL to avoid warnings.
Signed-off-by: Jouni Malinen <j@w1.fi>
The internal entropy pool was previously used to prevent 4-way handshake
in AP mode from completing before sufficient entropy was available to
allow secure keys to be generated. This commit extends that workaround
for boards that do not provide secure OS level PRNG (e.g., /dev/urandom
does not get enough entropy) for the most critical WPS operations by
rejecting AP-as-enrollee case (use of AP PIN to learn/modify AP
configuration) and new PSK/passphrase generation. This does not have any
effect on devices that have an appropriately working OS level PRNG
(e.g., /dev/random and /dev/urandom on Linux).
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
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 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: Jouni Malinen <j@w1.fi>
While this call cannot really fail, check the return value to be more
consistent with all the other wps_build_wfa_ext() calls.
Signed-off-by: Jouni Malinen <j@w1.fi>
The implementation here was doing what it was supposed to, but the code
was indented in a way that made it quite confusing in the context of a
single line if statement body.
Signed-off-by: Jouni Malinen <j@w1.fi>
In theory, gmtime() could return NULL if the year value would not fit
into an integer. However, that cannot really happen with the current
time() value in practice. Anyway, clean up static analyzer reports by
checking for this corner case.
Signed-off-by: Jouni Malinen <j@w1.fi>
Must check that UDN was present before trying to parse it. Avoid a NULL
pointer dereference by checking the result before using os_strstr() when
parsing device description from an AP.
Signed-off-by: Jouni Malinen <j@w1.fi>
The previous check for dl_list_len() or having an entry from the list is
sufficient, but some static analyzers cannot figure out that
dl_list_first() and dl_list_last() will return non-NULL in this type of
cases. Avoid invalid reports by explicitly checking for NULL.
Signed-off-by: Jouni Malinen <j@w1.fi>
Now that WPS 2.0 support is enabled unconditionally, WEP and Shared auth
type are not allowed. This made some of the older code unused and that
can now be removed to clean up the implementation. There is still one
place where WEP is allowed for testing purposes: wpa_supplicant as
Registrar trying to configure an AP to use WEP. That is now only allowed
in CONFIG_TESTING_OPTIONS=y builds, though.
Signed-off-by: Jouni Malinen <j@w1.fi>
There is not much point in building devices with WPS 1.0 only supported
nowadays. As such, there is not sufficient justification for maintaining
extra complexity for the CONFIG_WPS2 build option either. Remove this by
enabling WSC 2.0 support unconditionally.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
While the device itself may support WPA/WPA2-Enterprise, enrollment of
credentials for EAP authentication is not supported through WPS. As
such, there is no need to claim support for these capabilities within
WPS information.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
When wpa_psk_file is used instead of wpa_psk/wpa_passphrase, each WPS
Enrollee was given a unique PSK. This did not work for the
station-as-Registrar case where ER would learn the current AP settings
instead of enrolling itself (i.e., when using the AP PIN instead of
station PIN). That case can be covered with a similar design, so
generate a per-device PSK when building M7 as an AP in wpa_psk_file
configuration.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
There is no need to use a for loop here since the h variable is
set identically at the beginning of the body anyway.
Signed-off-by: Jouni Malinen <j@w1.fi>
record->type == NULL case was handled through the record->type_length
comparison. While this was correct, it is a bit difficult for static
analyzers to understand, so add an extra check for NULL to avoid false
reports on this.
Signed-off-by: Jouni Malinen <j@w1.fi>
UPnP-arch-DeviceArchitecture describe ErrorCode 412 to be used for the
case where no un-epxired subscription matches. This used to return 200
which is not strictly speaking correct even though it is unlikely to
cause any problems.
Signed-off-by: Jouni Malinen <j@w1.fi>
The caller of the GET parser is checking this already and the GET case
was the only one that ended up doing the duplicated validation step.
Signed-off-by: Jouni Malinen <j@w1.fi>
The previous parser would have skipped a WFA vendor extension attribute
that includes only a single zero-length subelement. No such subelement
has been defined so far, so this does not really affect any
functionality, but better make the parser address this correctly should
such an element ever be added.
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Previously, unconfigured state was forcing the best supported
authentication and encryption state to be shown in WPS messages,
including AP Settings in M7 in case the AP acts as an Enrollee. This is
not really correct for the AP Settings case, so change that one to
indicate the currently configured state.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Commit bd3a373767 added a mechanism to use
AP Channel attribute from within a Credential attribute to optimize
scans. However, this design is not actually used with the WPS NFC use
cases. With configuration token, the AP Channel attribute is in the same
container with the Credential attribute (and that was also handled in
the previous implementation). With connection handover, AP Channel
information is outside the Credential attribute as well.
Simplify implementation by removing the AP Channel within Credential
case. This allows wpas_wps_use_cred() to get the AP Channel from the
container instead of having to find this during credential iteration.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
This is not really used in practice and there is no need to maintain
unsed code that would only print debug log entries.
Signed-hostap: Jouni Malinen <j@w1.fi>
If CONFIG_WPS_TESTING=y is enabled in build configuration, the new
wps_corrupt_pkhash parameter (similar to wps_testing_dummy_cred) can be
used to request public key hash to be corrupted in all generated OOB
Device Password attributes. This can be used for testing purposes to
validate public key hash validation steps.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
When "P2P_SET nfc_tag 1" is used to enable the own NFC Tag for P2P, also
enable it for any running GO interface.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
This was already done for Registrar, but the Enrollee case did not
set config error properly if Registrar public key did not match the
hash received during NFC connection handover.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
"NFC_REPORT_HANDOVER {INIT,RESP} P2P <req> <sel>" can now be used to
report completed NFC negotiated connection handover in which the P2P
alternative carrier was selected.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
These functions can be used to build the WPS attributes for P2P NFC
connection handover messages.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
It is more useful to be able to build a single NFC carrier record
instead of the full connection handover request message to allow
external components to decide whether to negotiate which alternative
carrier is used. This updates the carrier record contents to the new
design to include Enrollee public key hash and provides this as a
carrier record instead of full message. An external program is expected
to be used to build the full NFC connection handover message with
potentially other alternative carrier records included.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
This allows DH key generation to be shared for other purposes than just
the case of OOB Device Password building. In addition, force the DH
public key buffer to be full 192 octets with zero padding to avoid
issues with the buffer being used in messages sent to a peer device.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
The new Device Password ID 7 is used to indicate that NFC connection
handover is used with DH public key hash from both devices being
exchanged over the NFC connection handover messages. This allows an
abbreviated M1-M2 handshake to be used since Device Password does not
need to be used when DH is authenticated with the out-of-band
information (validation of the public key against the hash).
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
When both the Registrar and Enrollee public key hashes are delivered
out-of-band (in NFC connection handover), use abbreviated WPS handshake
(skip M3-M8).
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Since the Enrollee can now get the public key hash from the Registrar,
there is need to validate this during the WPS protocol run.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Instead of terminating the WPS protocol immediately, go through an M2D
exchange to notify Enrollee of the public key hash mismatch.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
This old Version attribute is not really needed anymore for these use
cases with the assumption that there are no existing WPS+NFC
deployments. It was removed from the WSC specification, so make the
implementation match that change.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
Previously, wps_version_number was used only to test extensibility to
newer version numbers, but it can also be used to enable testing of
older versions (1.0), e.g., to avoid hitting some 2.0 specific
validation steps.
Signed-hostap: Jouni Malinen <j@w1.fi>
It is fine to try to cancel a registration that does not exist, so there
is no need to have the duplicated checks for eloop timeout and socket
registration.
Signed-hostap: Jouni Malinen <j@w1.fi>
The five second timeout to call wps_er_deinit_finish() could potentially
be left behind when removing the ER data based on some other event. This
could result in double-freeing of wps_er context killing the process,
e.g., if the WPS ER functionality is stopped while in the process of
unsubscribing from an AP and then restarted.
In addition, AP entries could still be present in the
er->ap_unsubscribing list when the deinit timeout hits. These entries
would still maintain HTTP context pointing to the ER which would be
freed here and as such, the following HTTP client callback could refer
to freed memory and kill the process. Fix this by freeing AP entries
from ap_unsubscribing list when ER is deinitialized from timeout even if
such AP entries have not completed unsubscription from UPnP events.
Signed-hostap: Jouni Malinen <jouni@qca.qualcomm.com>
The event debouncing isn't very accurate (since it doesn't
take sub-second resolution into account), but it should use
monotonic time anyway since it doesn't care about the wall
clock.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
The PBC ignore-start workaround just needs to check whether
the time is within 5 seconds, so should use monotonic time.
While at it, add a few more ifdefs to clearly separate the
code and variables needed.
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
PBC sessions are just time-stamped when activated, and
eventually time out, so should use monotonic time.
While at it, make the code use os_reltime_expired().
Signed-hostap: Johannes Berg <johannes.berg@intel.com>
