30 Oct, 2009
1 commit
28 Oct, 2009
8 commits
-
This makes the mesh housekeeping timer work properly on big endian
systems.Signed-off-by: Rui Paulo
Signed-off-by: Javier Cardona
Signed-off-by: John W. Linville -
Mesh portals proxy traffic for nodes external to the mesh. When a
proxied frame is received by a mesh interface, it should update its mesh
portal table. This was only happening for unicast frames. With this
change we also learn about mesh portals from proxied multicast frames.Signed-off-by: Javier Cardona
Signed-off-by: John W. Linville -
Replace netif_tx_{start,stop,wake}_all_queues with the single-queue
equivalents (i.e. netif_{start,stop,wake}_queue). Since we are down to
a single queue, these should peform slightly better.Signed-off-by: John W. Linville
-
Get rid of cookies in cfg80211_send_XXX() functions.
Signed-off-by: Holger Schurig
Signed-off-by: John W. Linville -
When hostapd injects a frame, e.g. an authentication or association
response, mac80211 looks for a suitable access point virtual interface
to associate the frame with based on its source address. This makes it
possible e.g. to correctly assign sequence numbers to the frames.A small typo in the ethernet address comparison statement caused a
failure to find a suitable ap interface. Sequence numbers on such
frames where therefore left unassigned causing some clients
(especially windows-based 11b/g clients) to reject them and fail to
authenticate or associate with the access point. This patch fixes the
typo in the address comparison statement.Signed-off-by: Björn Smedman
Reviewed-by: Johannes Berg
Cc: stable@kernel.org
Signed-off-by: John W. Linville -
Fix a typo in the description of hwmp_route_info_get(), no function
changes.Signed-off-by: Andrey Yurovsky
Signed-off-by: John W. Linville -
When association fails, we should stay authenticated,
which in mac80211 is represented by the existence of
the mlme work struct, so we cannot free that, instead
we need to just set it to idle.(Brought to you by the hacking session at Kernel Summit 2009 in Tokyo,
Japan. -- JWL)Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Recent commit "mac80211: fix logic error ibss merge bssid check" fixed
joining of ibss cell when static bssid is provided. In this case
ifibss->bssid is set before the cell is joined and comparing that address
to a bss should thus always succeed. Unfortunately this change broke the
other case of joining a ibss cell without providing a static bssid where
the value of ifibss->bssid is not set before the cell is joined.Since ifibss->bssid may be set before or after joining the cell we do not
learn anything by comparing it to a known bss. Remove this check.Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville
14 Oct, 2009
1 commit
13 Oct, 2009
3 commits
-
ieee80211_rx() must be called with softirqs disabled
since the networking stack requires this for netif_rx()
and some code in mac80211 can assume that it can not
be processing its own tasklet and this call at the same
time.It may be possible to remove this requirement after a
careful audit of mac80211 and doing any needed locking
improvements in it along with disabling softirqs around
netif_rx(). An alternative might be to push all packet
processing to process context in mac80211, instead of
to the tasklet, and add other synchronisation.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
When a scan completes, we call ieee80211_sta_find_ibss(),
which is also called from other places. When the scan was
done in software, there's no problem as both run from the
single-threaded mac80211 workqueue and are thus serialised
against each other, but with hardware scan the completion
can be in a different context and race against callers of
this function from the workqueue (e.g. due to beacon RX).
So instead of calling ieee80211_sta_find_ibss() directly,
just arm the timer and have it fire, scheduling the work,
which will invoke ieee80211_sta_find_ibss() (if that is
appropriate in the current state).Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Signed-off-by: Felix Fietkau
Acked-by: Johannes Berg
Signed-off-by: John W. Linville
12 Oct, 2009
1 commit
08 Oct, 2009
3 commits
-
Signed-off-by: John W. Linville
-
kfree_skb() should be used to free struct sk_buff pointers.
Signed-off-by: Roel Kluin
Acked-by: Johannes Berg
Cc: stable@kernel.org
Signed-off-by: John W. Linville -
When receiving data frames, we can send them only to
the interface they belong to based on transmitting
station (this doesn't work for probe requests). Also,
don't try to handle other frames for AP_VLAN at all
since those interface should only receive data.Additionally, the transmit side must check that the
station we're sending a frame to is actually on the
interface we're transmitting on, and not transmit
packets to functions that live on other interfaces,
so validate that as well.Another bug fix is needed in sta_info.c where in the
VLAN case when adding/removing stations we overwrite
the sdata variable we still need.Signed-off-by: Johannes Berg
Cc: stable@kernel.org
Signed-off-by: John W. Linville
05 Oct, 2009
1 commit
-
Instead of having to modify every non-mac80211 for device type assignment,
do this inside the netdev notifier callback of cfg80211. So all drivers
that integrate with cfg80211 will export a proper device type.Signed-off-by: Marcel Holtmann
Signed-off-by: David S. Miller
30 Sep, 2009
1 commit
-
Consider the following step-by step:
1. A STA authenticates and associates with the AP and exchanges
traffic.
2. The STA reports to the AP that it is going to PS state.
3. Some time later the STA device goes to the stand-by mode (not only
its wi-fi card, but the device itself) and drops the association state
without sending a disassociation frame.
4. The STA device wakes up and begins authentication with an
Auth frame as it hasn't been authenticated/associated previously.At the step 4 the AP "remembers" the STA and considers it is still in
the PS state, so the AP buffers frames, which it has to send to the STA.
But the STA isn't actually in the PS state and so it neither checks
TIM bits nor reports to the AP that it isn't power saving.
Because of that authentication/[re]association fails.To fix authentication/[re]association stage of this issue, Auth, Assoc
Resp and Reassoc Resp frames are transmitted disregarding of STA's power
saving state.N.B. This patch doesn't fix further data frame exchange after
authentication/[re]association. A patch in hostapd is required to fix
that.Signed-off-by: Igor Perminov
Signed-off-by: John W. Linville
29 Sep, 2009
1 commit
-
It's useful to know the MAC address when being
disassociated; fix a typo (missing colon) and
move some messages so we get them only when they
are actually taking effect.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville
23 Sep, 2009
2 commits
-
When the DTIM setting is read from beacons, mac80211 will
assume it is 1 if the TIM IE is not present or the value
is 0. This sounds fine, but the same function processes
probe responses as well, which don't have a TIM IE. This
leads to overwriting any values previously parsed out of
beacon frames.Thus, instead of checking for the presence of the TIM IE
when setting the default, simply check whether the DTIM
period value is valid already. If the TIM IE is not there
then the value cannot be valid (it is initialised to 0)
and probe responses received after beacons will not lead
to overwriting an already valid value.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
The contention window is supposed to be a power of two minus one, i.e.
15, 31, 63, 127... minstrel_rate_init() forgets to subtract 1, so the
sequence becomes 15, 32, 66, 134...Bug reported by Dan Halperin
Signed-off-by: Pavel Roskin
Signed-off-by: John W. Linville
12 Sep, 2009
1 commit
-
The Ethernet framing is used for a lot of devices these days. Most
prominent are WiFi and WiMAX based devices. However for userspace
application it is important to classify these devices correctly and
not only see them as Ethernet devices. The daemons like HAL, DeviceKit
or even NetworkManager with udev support tries to do the classification
in userspace with a lot trickery and extra system calls. This is not
good and actually reaches its limitations. Especially since the kernel
does know the type of the Ethernet device it is pretty stupid.To solve this problem the underlying device type needs to be set and
then the value will be exported as DEVTYPE via uevents and available
within udev.# cat /sys/class/net/wlan0/uevent
DEVTYPE=wlan
INTERFACE=wlan0
IFINDEX=5This is similar to subsystems like USB and SCSI that distinguish
between hosts, devices, disks, partitions etc.The new SET_NETDEV_DEVTYPE() is a convenience helper to set the actual
device type. All device types are free form, but for convenience the
same strings as used with RFKILL are choosen.Signed-off-by: Marcel Holtmann
Signed-off-by: David S. Miller
03 Sep, 2009
2 commits
-
We can just display this upon enabling mac80211 with an
'if MAC80211 != n' check.Cc: Johannes Berg
Signed-off-by: Luis R. Rodriguez
Signed-off-by: John W. Linville
02 Sep, 2009
1 commit
-
All instances of file_operations should be const.
Signed-off-by: Stephen Hemminger
Signed-off-by: David S. Miller
01 Sep, 2009
1 commit
-
Mostly just simple conversions:
* ray_cs had bogus return of NET_TX_LOCKED but driver
was not using NETIF_F_LLTX
* hostap and ipw2x00 had some code that returned value
from a called function that also had to change to return netdev_tx_tSigned-off-by: Stephen Hemminger
Signed-off-by: David S. Miller
29 Aug, 2009
9 commits
-
When the queued management work items are processed in
ieee80211_sta_work() an item could be removed. This could change the
anybusy from true to false, so we better check whether we can start a
new scan only after having processed the pending work first.Signed-off-by: Jouni Malinen
Acked-by: Johannes Berg
Signed-off-by: John W. Linville -
When an interface is taken down while a scan is
pending -- i.e. a scan request was accepted but
not yet acted upon due to other work being in
progress -- we currently do not properly cancel
that scan and end up getting stuck. Fix this by
doing better checks when an interface is taken
down.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
An integer overflow in the minstrel debug code prevented the
throughput to be displayed correctly. This patch fixes that,
by permutating operations like proposed by Pavel Roskin.Signed-off-by: Arnd Hannemann
Signed-off-by: John W. Linville -
In mac80211's RX path some of the warnings that
warn about drivers passing invalid status values
leak the skb, fix that by refactoring the code.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
I noticed that the throughput field of the minstrel_rate struct is never used,
so remove it.Signed-off-by: Arnd Hannemann
Signed-off-by: John W. Linville -
Due to the way the tasklets work in mac80211 there's
no need to ever disable them.However, we need to clear the pending packets when
taking down the last interface because otherwise
the tx_pending_tasklet might be queued if the
driver mucks with the queues (which it shouldn't).I've had a situation occasionally with ar9170 in
which ksoftirq was using 100% CPU time because
a disabled tasklet was scheduled, and I think that
was due to ar9170 receiving a packet while the
tasklet was disabled. That's strange and it really
should not do that for other reasons, but there's
no need to waste that much CPU time over it, it
should just warn instead.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
With the libipw naming scheme change, it is no longer necessary for
mac80211 to avoid the ieee80211_rx name clash.Reported-by: Johannes Berg
Signed-off-by: John W. Linville -
Since configure_filter can sleep now, any multicast
configuration needed to be postponed to a work struct.
This, however, lead to a problem that we could queue
the work, stop the device and then afterwards invoke
configure_filter which may lead to driver hangs and is
a bug. To fix this, we can just cancel the filter work
since it's unnecessary to do after stopping the hw.Since there are various places that call drv_stop, and
two of them do very similar things, the code for them
can be put into a shared function at the same time.Signed-off-by: Johannes Berg
Reported-by: Lennert Buytenhek
Tested-by: Lennert Buytenhek
Signed-off-by: John W. Linville -
The mesh config information element has changed significantly since draft 1.08
This patch brings it up to date.Thanks to Sam Leffler and Rui Paulo for identifying this.
Signed-off-by: Javier Cardona
Signed-off-by: John W. Linville
20 Aug, 2009
4 commits
-
"cfg80211: fix alignment problem in scan request"
introduced a bug into the error path, because now
we allocate the entire scan request and not just
the channel list (the channel list is allocated
together with the scan request) -- on errors we
thus also need to free the entire scan request.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
This expands on the current fwded_frames stat counter which should be equal to
the total of these two new counters. The new counters are called "fwded_mcast"
and "fwded_unicast".Signed-off-by: Daniel Walker
Signed-off-by: Javier Cardona
Signed-off-by: John W. Linville -
CC [M] net/mac80211/rc80211_pid_algo.o
net/mac80211/rc80211_pid_algo.c: In function ‘rate_control_pid_rate_init’:
net/mac80211/rc80211_pid_algo.c:304: warning: unused variable ‘si’Signed-off-by: John W. Linville
-
Mesh uses the tx failure average to compute the (m)path metric. This used to
be done inside the rate control module. This patch breaks the dependency
between the mesh stack and the rate control algorithm. Mesh will now work
independently of the chosen rate control algorithm.The mesh stack keeps a moving average of the average transmission losses for
each mesh peer station. If the fail average exceeds a certain threshold, the
peer link is marked as broken.Signed-off-by: Javier Cardona
Signed-off-by: John W. Linville
