Jouni Malinen [Sun, 14 Feb 2010 14:08:15 +0000 (16:08 +0200)]
EAP-TNC: Add Flags field into fragment acknowledgement
TNC IF-T is somewhat unclear on this are, but
draft-hanna-nea-pt-eap-00.txt, which is supposed to define the same
protocol, is clearer on the Flags field being included.
This change breaks interoperability with the old implementation if
EAP-TNC fragmentation is used. The old version would not accept
the acknowledgement message with the added Flags octet while the
new version accepts messagss with with both options.
Jouni Malinen [Sat, 13 Feb 2010 16:03:52 +0000 (18:03 +0200)]
EAP-TNC server: Fix processing when last message is fragmented
If the last message from the EAP-TNC server was fragmented, the
fragment processing lost the DONE/FAIL state and did not know how
to handle the final ACK from the peer. Fix this by remembering the
earlier DONE/FAIL state when fragmenting a frame.
Jouni Malinen [Sat, 13 Feb 2010 16:00:39 +0000 (18:00 +0200)]
EAP-TNC: Accept fragment ack frame with Flags field
TNC IF-T specification is unclear on the exact contents of the fragment
acknowledgement frame. An interoperability issue with the tncs@fhh
implementation was reported by Arne Welzel
<arne.welzel@stud.fh-hannover.de> due to the different interpretations
of the specification. Relax EAP-TNC server/peer validation rules to
accept fragmentation acknowledgement frames to include the Flags field
to avoid this issue.
Masashi Honma [Sat, 13 Feb 2010 11:59:29 +0000 (13:59 +0200)]
bsd: Aggregate ioctl routines
This patch aggregates ioctls.
First is SIOCS80211. The SIOCS80211's arguments has 3 couples.
1-1. i_len, i_data
1-2. i_val
1-3. i_len, i_data, i_val (currently only IEEE80211_IOC_APPIE)
There were 3 routines for each cases. This patch aggregates these to
one.
Second is SIOCG80211. The SIOCG80211 returns 2 type of value.
2-1. i_len
2-2. i_val
There were 2 routines for each cases. This patch aggregates these to
one.
I have tested on both FreeBSD 8.0 and NetBSD 5.0.1 with these cases.
Masashi Honma [Sat, 13 Feb 2010 11:48:52 +0000 (13:48 +0200)]
bsd: Use same field name between hostapd and wpa_supplicant
This patch modifies field name of struct bsd_driver_data to use
same name of struct wpa_driver_bsd_data. This is a preparation of
unifying struct bsd_driver_data and struct wpa_driver_bsd_data.
Jouni Malinen [Sat, 13 Feb 2010 09:14:23 +0000 (11:14 +0200)]
Add TLS client events, server probing, and srv cert matching
This allows external programs (e.g., UI) to get more information
about server certificate chain used during TLS handshake. This can
be used both to automatically probe the authentication server to
figure out most likely network configuration and to get information
about reasons for failed authentications.
The follow new control interface events are used for this:
CTRL-EVENT-EAP-PEER-CERT
CTRL-EVENT-EAP-TLS-CERT-ERROR
In addition, there is now an option for matching the server certificate
instead of the full certificate chain for cases where a trusted CA is
not configured or even known. This can be used, e.g., by first probing
the network and learning the server certificate hash based on the new
events and then adding a network configuration with the server
certificate hash after user have accepted it. Future connections will
then be allowed as long as the same server certificate is used.
Authentication server probing can be done, e.g., with following
configuration options:
eap=TTLS PEAP TLS
identity=""
ca_cert="probe://"
Example set of control events for this:
CTRL-EVENT-EAP-STARTED EAP authentication started
CTRL-EVENT-EAP-PROPOSED-METHOD vendor=0 method=21
CTRL-EVENT-EAP-METHOD EAP vendor 0 method 21 (TTLS) selected
CTRL-EVENT-EAP-PEER-CERT depth=0 subject='/C=US/ST=California/L=San Francisco/CN=Server/emailAddress=server@kir.nu' hash=5a1bc1296205e6fdbe3979728efe3920798885c1c4590b5f90f43222d239ca6a
CTRL-EVENT-EAP-TLS-CERT-ERROR reason=8 depth=0 subject='/C=US/ST=California/L=San Francisco/CN=Server/emailAddress=server@kir.nu' err='Server certificate chain probe'
CTRL-EVENT-EAP-FAILURE EAP authentication failed
Jouni Malinen [Fri, 12 Feb 2010 19:13:51 +0000 (21:13 +0200)]
Fix TLS in/out buffer freeing
The previous version could end leaking memory since os_free() was used
instead of wpabuf_free(). In addition, this could potentially have
triggered a crash if the TLS context were being freed when pending
input data where still in the buffer (though, this may not be possible
to trigger in practice).
Masashi Honma [Mon, 8 Feb 2010 19:18:09 +0000 (21:18 +0200)]
driver_bsd.c: Reduce code duplication (DELKEY)
This patch reduces code duplication between hostapd and wpa_supplicant
about IEEE80211_IOC_DELKEY. This is a preparation for AP mode
wpa_supplicant. This is a patch to
http://lists.shmoo.com/pipermail/hostap/2010-January/021030.html.
This patch does these automatically. Maybe there will be some
objections, like "hardware configuration is not hostapd/wpa_supplican's
work". So I will write the reasons why I made this patch.
1. For usability.
2. The first command fails when previous state is adhoc. This patch is
free from previous state.
3. Some driver wrappers configure these automatically (like nl80211).
4. I have wasted time trying to find out these command were needed :(
Hamish Guthrie [Mon, 8 Feb 2010 19:08:54 +0000 (21:08 +0200)]
driver_ps3: Remove legacy ps3 wpa driver
The ps3 wireless kernel driver has wireless extension support.
There is a legacy wpa_supplicant driver, and support for this
has been removed from the kernel driver, as no distributions
are using it.
Jouni Malinen [Mon, 25 Jan 2010 02:19:50 +0000 (18:19 -0800)]
Try to avoid some unnecessary roaming
When multiple APs are present in scan results with similar signal
strength, wpa_supplicant may end up bounching between them frequently
whenever new scan results are available (e.g., due to periodic scans
requested by NetworkManager). This can result in unnecessary roaming
and in case of the current cfg80211 version, to frequent network
disconnections.
Do not request a roam if the current BSS is still present in the scan
results and the selected BSS is in the same ESS and has only a slighly
stronger signal strength.
Jouni Malinen [Mon, 25 Jan 2010 02:11:30 +0000 (18:11 -0800)]
nl80211: Dump scan results in debug log if association command fails
This may help in debugging why cfg80211 refused the association
command since the scan results should include information about all
pending authentication and association states.
Jouni Malinen [Mon, 25 Jan 2010 02:07:34 +0000 (18:07 -0800)]
nl80211: Clear cfg80211 authentication data for old entries
cfg80211 has a limit on pending authentications, so we better clear
the entries that we do not care about to avoid hitting the limit
when roaming between multiple APs.
Witold Sowa [Sat, 16 Jan 2010 14:37:37 +0000 (16:37 +0200)]
dbus: Change WPA/RSNIE byte array props to dicts
Expose RSN and WPA properties for BSS objects containing information
about key management and cipher suites. Get rid of WPA/RSN/WPSIE
byte array properties and add IEs byte array property with all IE data
instead.
Jouni Malinen [Sat, 16 Jan 2010 14:11:05 +0000 (16:11 +0200)]
Fetch IEs from both Beacon and Probe Response frames if available
This allows the driver wrappers to return two sets of IEs, so that
the BSS code can use information from both Beacon and Probe Response
frames if needed. For example, some Cisco APs seem to include more
information in Wireless Provisioning Services IE when it is in the
Beacon frame.
Jouni Malinen [Sat, 16 Jan 2010 13:19:58 +0000 (15:19 +0200)]
hostapd: Detect bridge interface automatically
This makes the bridge parameter unnecessary for cases where the interface
is already in a bridge and sysfs is mounted to /sys so that the detection
code works.
For nl80211, the bridge parameter can be used to request the AP
interface to be added to the bridge automatically (brctl may refuse to
do this before hostapd has been started to change the interface mode).
If needed, the bridge interface is also created.
Jouni Malinen [Tue, 12 Jan 2010 18:01:09 +0000 (20:01 +0200)]
wext: Add cfg80211-specific optimization to avoid silly behavior
If the driver is detected to use cfg80211, we can rely on it being able
to disconnect with SIOCSIWMLME commands and to use empty SSID as a way
to stop it from associating when we are in progress of configuring the
driver for association. Consequently, we can remove the hack that uses
random 32-octet SSID to force disconnection and re-order association
commands to match the expectations that cfg80211 has for WEXT ioctls.
This gets rid of extra scan rounds and attempts to associate with the
silly 32-octet SSID.
Jouni Malinen [Sun, 10 Jan 2010 19:53:17 +0000 (21:53 +0200)]
IBSS RSN: Explicitly check addr != NULL before passing it to memcmp
idx == 0 should be enough to make sure that the addr is set, but
verify that this is indeed the case to avoid any potential issues if
auth_set_key() gets called incorrectly.
Jouni Malinen [Sun, 10 Jan 2010 19:48:27 +0000 (21:48 +0200)]
eloop: Clear timeout data during allocation
Better make sure the eloop_timeout data gets fully initialized. The
current code is filling in all the fields, but it is clearer to just
zero the buffer to make sure any new field added to the structure gets
initialized.
Jouni Malinen [Sun, 10 Jan 2010 17:00:25 +0000 (19:00 +0200)]
Use zero address when reporting unknown peer in SMK error
This avoids potential use of uninitialized stack memory when printing
out peer address based on SMK error message that does not include the
MAC address.
Jouni Malinen [Sun, 10 Jan 2010 16:54:41 +0000 (18:54 +0200)]
Avoid a theoretical use-after-free in WPA auth sm init
wpa_sm_step() could theoretically free the statemachine, but it does
not do it in this particular case. Anyway, the code can be cleaned to
verify whether the state machine data is still available after the
wpa_sm_step() call.
Masashi Honma [Sat, 9 Jan 2010 09:04:44 +0000 (11:04 +0200)]
FreeBSD: Fix driver_bsd.c build
On FreeBSD 8.0, driver_bsd.c build fails because of changes from
older versions of FreeBSD. The error messages are below:
In file included from ../src/drivers/driver_bsd.c:38:
/usr/include/net80211/ieee80211_crypto.h:94: error: 'IEEE80211_TID_SIZE'
undeclared here (not in a function)
../src/drivers/driver_bsd.c: In function 'wpa_driver_bsd_set_wpa_ie':
../src/drivers/driver_bsd.c:968: error: 'IEEE80211_IOC_OPTIE' undeclared (first
use in this function)
../src/drivers/driver_bsd.c:968: error: (Each undeclared identifier is reported
only once
../src/drivers/driver_bsd.c:968: error: for each function it appears in.)
gmake: *** [../src/drivers/driver_bsd.o] Error 1
Jouni Malinen [Fri, 8 Jan 2010 22:38:09 +0000 (00:38 +0200)]
Fix PKCS#12 use with OpenSSL 1.0.0
Add 40-bit RC2 CBC explicitly since OpenSSL 1.0.0 does not seem to that
anymore with PKCS12_PBE_add(). Furthermore, at least 1.0.0-beta4 crashes
if the needed cipher is not registered when parsing the PKCS#12 data
(this crashing part should be fixed in newer 1.0.0 versions)
Following bug reports are related to the issue:
https://bugzilla.redhat.com/show_bug.cgi?id=541924
https://bugzilla.redhat.com/show_bug.cgi?id=538851
http://rt.openssl.org/Ticket/Display.html?id=2127
http://rt.openssl.org/Ticket/Display.html?id=2128
When the wme_ac_??_cw{min,max} parameters aren't specified in
hostapd.conf, hostapd uses an incorrect set of default values, as the
defaults are in 2^x-1 form instead of in log form. This patch changes
them over to the expected log form.
Jouni Malinen [Wed, 6 Jan 2010 09:55:21 +0000 (11:55 +0200)]
dbus: Fix prop_changed_flags allocation to count properties correctly
There was an extra semicolon that broke the calculation of registered
properties and resulted in obj_desc->prop_changed_flags not being
allocated long enough for all the flags.
Jouni Malinen [Wed, 6 Jan 2010 09:50:43 +0000 (11:50 +0200)]
dbus: Fix flush_object_timeout_handler context to be consistent
The path pointer used as the timeout_ctx was not constant; the path
string itself may have been the same, but the pointer certainly was not
since it was sometimes from stack and sometimes from the dynamically
allocated buffer in obj_desc. This caused some of the eloop timeout
cancellations not to find the timeout. Fix this by using the obj_desc
as the timeout context data.
Witold Sowa [Mon, 4 Jan 2010 15:15:57 +0000 (17:15 +0200)]
dbus: Aggregate PropertiesChanged signals
Instead of sending PropertiesChanged signals for each changed
property separately, mark properties as changed and send aggregated
PropertiesChanged signals for each interface in each object.
Aggregated PropertiesChanged signal is sent
- for all object after responding on DBus call
- for specified object after manual call to
wpa_dbus_flush_object_changed_properties() function
- for each object separately after short timeout (currently 5 ms)
which starts when first property in object is marked changed
Marcel Holtmann [Mon, 4 Jan 2010 14:33:44 +0000 (16:33 +0200)]
dbus: Replace StateChanged with PropertiesChanged signal
The actual supplicant state is exposed via a property on the interface
object. So having a separate signal StateChanged for notifying about
changes is a bad idea. The standard PropertiesChanged signal should be
used for this.
The advantage of StateChanged signal was that it includes the previous
state, but not even NetworkManager is making use of this. And tracking
the old state via the property and this signal is easily possible anyway.
Marcel Holtmann [Mon, 4 Jan 2010 14:22:26 +0000 (16:22 +0200)]
dbus: Use simple strings for debug level
Instead of using some magic integer values that really only mean
something to WPA internal code, just use simple strings. Possible
values are "msgdump", "debug", "info", "warning" and "error" which
map directly to WPA debugging support.
Jouni Malinen [Sun, 3 Jan 2010 20:08:26 +0000 (22:08 +0200)]
Share a single Linux ioctl helper fo setting interface up/down
Number of Linux driver wrappers included this more or less identical
function, so lets add a new helper file to be able to share some more
code between the driver wrappers.