Commit Graph

131 Commits

Author SHA1 Message Date
sam
c0c5a03a41 correct comparison for null ptr
Noticed by:	Coverity Prevent analysis tool
2005-03-16 20:42:00 +00:00
sam
9b7f94f833 avoid potential array index by -1
Noticed by:	Coverity Prevent analysis tool
2005-03-16 20:40:48 +00:00
sam
0ea8247d55 eliminate use after free in debug code
Noticed by:	Coverity Prevent analysis tool
2005-03-16 20:39:05 +00:00
sam
bddcc80ceb kill dead code
Noticed by:	Coverity Prevent analysis tool
2005-03-16 20:35:23 +00:00
sam
228aadaa15 honor any desired bssid when creating an ibss
Prodded by:	David Young
Obtained from:	netbsd
2005-03-08 17:04:14 +00:00
wpaul
8afa2c4c26 Recently, it was reported to me that you could provoke a double fault
panic with the NDISulator if you did "ifconfig ndis0 10.0.0.1/24,"
whereas "ifconfig ndis0 10.0.0.1/24 up" worked fine. The double fault
was caused by the ifconfig thread running out of kernel stack space.
(This was partly due to the NDIsulator using a couple of big buffers on
the stack, but even after fixing that the double fault persisted.)

It turns out that ndis_init() is called in both cases, but in the first
case the code path passes through ieee80211_ioctl(), and it turns out
ieee80211_ioctl() consumes a whopping 2400 bytes of stack space.
Apparently, gcc -O2 causes the ieee80211_ioctl_get80211() routine to
be inlined into ieee80211_ioctl(), and for some reason which I do not
fully understand, this causes ieee80211_ioctl() to consume an extra 2K
of stack space.

To prevent this overly agressive optimization, ieee80211_ioctl_get80211()
is now declared with __attribute__ ((noinline)). With this change,
ieee80211_ioctl() now only reserves about 200 bytes of stack instead of 2400.
2005-03-03 17:35:05 +00:00
sam
9b2bcce14f mark timestamp for pending fragments
Noticed by:	Jeffrey D. Chung
2005-02-23 04:52:30 +00:00
sam
bfba17efd6 rev 1.19 fixed wpa supplicant but broke wpa authenticator; when operating
as an authenticator need to always check for the unicast key in the node
(as was the case before)

Submitted by:	Divy Le Ray
2005-02-10 17:00:48 +00:00
sam
9b44db51a1 propagate state kept in the bss node when re-creating the node
on state transitions; this is a stopgap measure, need to rethink
how we do management of this state

Identified by:	Divy Le Ray
2005-02-10 16:59:04 +00:00
sam
a1edfec034 correct check for unicast key being setup; wpa_supplicant in particular
uses only the global key state so looking in the per-node key slot is wrong

Submitted by:	Tai-hwa Liang
2005-02-03 16:21:07 +00:00
sam
217734da84 supply a default ic_reset method for drivers; the ioctl code expect this
method to always be setup

Submitted by:	Tai-hwa Liang
2005-01-27 17:39:17 +00:00
sam
e0b285228a update node reference count debug msgs to include the node address
since the mac address may not be sufficient to uniquely identify a node
2005-01-24 20:50:20 +00:00
sam
a55551873c clear IEEE80211_F_WMEUPDATE when building the initial beacon frame so we
don't do an update on the first beacon
2005-01-24 20:41:12 +00:00
sam
c504db8755 noop change so RUN->RUN transition isn't considered invalid
(it happens on ibss merge)
2005-01-24 20:39:29 +00:00
sam
d07ff12dc9 add macros to convert between txop's and usecs 2005-01-24 20:38:26 +00:00
sam
01ce092a3a o clarify that beacon interval settings are in TU's, not ms
o add macros to convert between TU's and ms
2005-01-24 19:39:07 +00:00
sam
459bcc3dba statically allocate the station/neighbor node table; the deferred
allocation scheme introduced a race condition during device state
transitions
2005-01-24 19:32:10 +00:00
sam
822cad40f8 handle potential stale values of bssid in neighbor nodes that
can occur after an ibss merge

Submitted by:	David Young
2005-01-22 20:33:37 +00:00
sam
7a9676cb41 when ssid suppression is enabled don't respond to probe requests
unless our ssid is specified
2005-01-22 20:31:22 +00:00
sam
78282d6a1d be consistent in naming inactivity timers;
net.wlan.X.inact -> net.wlan.X.inact_run
2005-01-22 20:29:23 +00:00
sam
c1c8fdfa03 o replace out-of-line copy of FCS w/ a flag that indicates the
frame includes FCS (requires applications to be updated, but since
  we weren't doing the out-of-line FCS stuff anyway app changes
  were needed already)
o add a flag to indicate padding exists between the 802.11 header and
  the payload (e.g. for Atheros cards)
o diff reducation against netbsd

MFC after:	1 week
2005-01-22 20:12:05 +00:00
sam
8ec9f729de when a station is timed out for inactivity, remove it from the table
so it isn't considered again
2005-01-20 02:59:21 +00:00
sam
3973fcd768 explicitly avoid timing out ourself due to inactivity; it
can easily happen if the bss is quiet
2005-01-20 02:54:18 +00:00
sam
2b292030fb fix refcnt leak in adhoc mode: entries in the neighbor table
created due to rx'd frames had an extra reference
2005-01-20 02:53:11 +00:00
sam
2138a8301b remove debug msg from ieee80211_iterate_nodes; it makes logs very noisy
as onoe rate control invokes this once a second
2005-01-18 20:35:39 +00:00
sam
ff12c179c8 avoid possible null pointer deref when refcnt debugging is enabled; the
node may be orphaned

Submitted by:	Tai-hwa Liang
2005-01-18 20:34:22 +00:00
sam
ae59d17866 Explicitly ignore ibss merge requests when the node is ic_bss. This can
happen on the first management frame received from a neighbor; we assume
any merge candidate will send more frames and those should be processed
with a suitable table entry.

Stepped on by:	Tai-hwa Liang
2005-01-18 20:30:16 +00:00
sam
c32261430c Simplify the logic for checking the scan candidates at the end of a scan.
Hold a lock on the table instead of futzing with reference counts which
was potentially dangerous except drivers were quiescent while we did this
so the table contents never changed.  Disable the hack logic for removing
scan candidates with multiple association failures; it's never done the
right thing and will be fixed correctly with background scanning goes in.
2005-01-18 20:21:29 +00:00
sam
6495462157 correct logic that caused beacon frames received in ibss mode to be
discarded when not scanning
2005-01-18 20:07:44 +00:00
sam
07d33c069a move beacon/probe response counting down to after we've decided whether or
not we're going to process the frame; this makes the counters reflect frames
actually processes instead of received (discarded frames were already counted)
2005-01-18 20:04:51 +00:00
sam
4c9a185523 when scanning is interrupted reset state so table entries go in the station
table and not the scan table

Noticed by:	Tai-hwa Liang
2005-01-18 19:59:40 +00:00
sam
5d86ecca9f do fixed rate check when considering if a scan candidate is suitable so when
it's time to join the bss we can't get an error
2005-01-18 19:52:36 +00:00
sam
8af7e8df05 Correct scan candidate selection logic for dual-band devices: prefer
candidate on 5Ghz channel to candidate on 2Ghz channel only when the
rssi are comparable (wasn't considering rssi).
2005-01-01 17:48:27 +00:00
sam
c5bfbdb959 remove netbsd rcsid lines; they are way out of date and we appear to be
diverging too much to make tracking these files worthwhile
2004-12-31 22:44:26 +00:00
sam
14fee97914 bump copyright for 2005 2004-12-31 22:42:38 +00:00
sam
141330c912 fixup inactivity timers:
o ic_inact_auth is a bad name, it's the inactivity threshold
  for being associated but not authorized; use it that way
o reset ni_inact when switching inactivity thresholds to
  minimize the race against the timer (don't want to lock
  for this stuff)
o change the inactivity probe threshold from a one-shot to
  cover a range: when below this threshold but not expired
  send a probe each inactivity interval; should probably
  guard against the interval being turned way down as this
  could cause us to spam the net with probes
2004-12-31 22:05:13 +00:00
sam
7c448457f6 Relearn how WPA keying is supposed to work and fix WPA+WME while
we're at it:
o WPA/802.11i has a unicast key and a group key; in station mode
  everything is sent with the unicast key--we were consulting the
  destination mac address and incorrectly using the group key
o (perpetuate fallback use of the default tx key to maintain
  compatibility with the way wpa_supplicant works)
o correct EAPOL encryption logic to check unicast key instead
  of assuming other state implies this
o move QoS encapsulation up to before enmic work so TKIP has the
  information required to calculate the pseudo-header
o do not do QoS-encapsulation of EAPOL frames as some ap's do the
  wrong thing with such frames (may need to revisit this if ap's
  start dropping non-QoS frames from stations assoc'd with QoS)
o move ieee80211_mbuf_adjust closer to its caller
2004-12-31 21:54:53 +00:00
sam
5a21ccb974 disable default use of wme until we can sort out interoperability issues;
users that want it can explicitly enable it
2004-12-31 21:48:15 +00:00
sam
fe8d9eb613 s/inline/__inline/ 2004-12-31 21:30:42 +00:00
sam
fe545f4516 compare pointers against NULL 2004-12-31 21:28:41 +00:00
sam
bb13d71061 make ap mode sta association debug msg as informative as sta mode 2004-12-31 21:27:53 +00:00
sam
95f27e450d do 11g erp station management in turboG mode too 2004-12-31 21:25:28 +00:00
sam
2cf886f5dc check if the node/station table is already present before creating one; this
can happen on an ibss merge
2004-12-31 21:22:12 +00:00
sam
886afd7287 include phy mode in scan debug msg 2004-12-31 21:19:59 +00:00
sam
e1832f531b accept IEEE80211_KEYIX_NONE for a default tx key 2004-12-31 21:13:35 +00:00
sam
d54bd57731 don't return prematurely from processing a beacon or probe response frame
when associated; this corrects a problem where doing a scan while associated
caused the associated ap to sometimes not be included in the scan set
2004-12-31 21:11:59 +00:00
sam
f249744499 change ieee80211_parse_wmeparams to return a unique value when the ie
is invalid so we can distinguish this from needing a parameter update;
fixes dynamic update of wme parameters
2004-12-31 21:09:43 +00:00
sam
59c07f15bb remove stray \n from debug msg 2004-12-31 21:05:05 +00:00
sam
d5e5ef7858 clear wep bit in 802.11 header after crypto decap so packet
capture doesn't see it
2004-12-31 21:04:05 +00:00
sam
8c3a1cf5aa o add net.wlan.X.driver_caps mib variable for setting the driver capabilities
flags for testing
o debug msg consistency fixups
2004-12-31 21:01:41 +00:00