23 Apr, 2009
40 commits
-
The maximum sleep interval, for powersave purposes, is
determined by the DTIM period (it may not be larger)
and the required networking latency (it must be small
enough to fulfil those constraints).This makes mac80211 calculate the maximum sleep interval
based on those constraints, and pass it to the driver.
Then the driver should instruct the device to sleep at
most that long.Note that the device is responsible for aligning the
maximum sleep interval between DTIMs, we make sure it's
not longer but it needs to make sure it's between them.Also, group some powersave documentation together and
make it more explicit that we support managed mode only,
and no IBSS powersaving (yet).Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Make the JOIN_IBSS command look at the beacon interval
attribute to see if the user requested a specific beacon
interval, if not default to 100 TU (wext too).Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Just setting IEEE80211_CONF_CHANGE_PS should be sufficient
for changes in the power saving things. The driver already
tells us whether it wants notification of dynps via the
"have dynps support" hw flag.Signed-off-by: Johannes Berg
Reviewed-by: Kalle Valo
Signed-off-by: John W. Linville -
Stephen Rothwell reported these warnings from a 32-bit build:
net/mac80211/mlme.c:1771: warning: left shift count >= width of type
net/mac80211/mlme.c:1772: warning: left shift count >= width of type
net/mac80211/mlme.c:1773: warning: left shift count >= width of type
net/mac80211/mlme.c:1774: warning: left shift count >= width of type
net/mac80211/mlme.c:1775: warning: left shift count >= width of typeThis shows a bug in my code -- BIT(X) uses just "1 << X" which means
a 32-bit integer on 32-bit platforms, but the code here needs a u64
on all platforms. Fix this by using "1ULL << X" instead of BIT(X).Thanks Stephen!
Reported-by: Stephen Rothwell
Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Private structure is allocated by wiphy_new now, so use
wiphy_free instead of kfree.Signed-off-by: Jussi Kivilinna
Signed-off-by: John W. Linville -
rndis_wext_link_change() might be called from rndis_command() at
initialization stage and priv->workqueue/priv->work have not been
initialized yet. This causes invalid opcode at rndis_wext_bind on
some brands of bcm4320.Fix by initializing workqueue/workers in rndis_wext_bind() before
rndis_command is used.Signed-off-by: Jussi Kivilinna
Signed-off-by: John W. Linville -
sparse complains, correctly, about these:
drivers/net/wireless/rndis_wlan.c:418:21: warning: symbol 'rndis_config_ops' was not declared. Should it be static?
drivers/net/wireless/rndis_wlan.c:423:6: warning: symbol 'rndis_wiphy_privid' was not declared. Should it be static?Fix that.
Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
With the RCU locking here we sleep while in an atomic context,
since we can sleep just use mutex locking for the interface
list instead of RCU. Sorry, seems I didn't get that in my UML
test.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Allow user to config the device all the time but only allow commiting
these changes to card if the card is up and running.Signed-off-by: Mohamed Abbas
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
Patch removes the "manual radio disable" parameter
as there is no usage scenario of disabling radio using this module parameter.
User can use iwconfig's txpower to enable and disable radio.This module parameter also does not work as expected. During module load
the status of radio is set, the radio is not actually disabled. Even so,
the moment mac80211 requests the interface to be up the radio will be
enabled again.Signed-off-by: Abhijeet Kolekar
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
Function rs_stay_in_table was flushing the rate scale table way to early. time_after
macro in expecting long value and was failing because we passing u32 value.Signed-off-by: Mohamed Abbas
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
This patch clean up the code for NL80211_IFTYPE_MONITOR mode,
priv->iw_mode is set in add_interface, but add_interface is never called
for monitor mode.
The only way mac80211 informs us about monitor mode is through
configuring filter; since iw_mode will never set to
NL80211_IFTYPE_MONITOR, modify and remove all the code refer to
NL80211_IFTYPE_MONITOR and replace with iwl_is_monitor_mode() function
call.Signed-off-by: Wey-Yi Guy
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
This patch calculates interrupt statistics for debugfs.
Signed-off-by: Abhijeet Kolekar
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
Patch adds debugfs to 3945. Also fix debugfs registration in iwlagn
to return error code if it fails.Signed-off-by: Abhijeet Kolekar
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
If passive scanning is requested we should not ask the
microcode to do active scanning after detecting traffic
on a channel -- that should only be used when an active
scan is requested but some channels are marked passive.Signed-off-by: Johannes Berg
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
This modifies iwlwifi to
* no longer build its own probe request, but use mac80211's
* therefore, support arbitrary scan IEs (up to the max len)
* support multiple scan SSIDs
* support passive scanningSigned-off-by: Johannes Berg
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
This limit applies to current (APIv1 and APIv2) 3945 firmware only, not
supported firmware of any of the other cards.Signed-off-by: Johannes Berg
Signed-off-by: Reinette Chatre
Signed-off-by: John W. Linville -
With this patch ar9170 is capable of receiving aggregated 802.11n frames
and sniffing on most networks without having a "debug message overhead".(Includes phy initialization requested by
Johannes Berg -- JWL)Signed-off-by: Christian Lamparter
Signed-off-by: John W. Linville -
The TIM IE must not be shorter than 4 bytes, so verify that
when parsing it and use the proper type. To ease that adjust
struct ieee80211_tim_ie to have a virtual bitmap of size
at least 1.Also check that the TIM IE is actually present before trying
to parse it!Because other people may need the function, make it a static
inline in ieee80211.h.(The original "mac80211: validate TIM IE length" was a minimal fix for
2.6.30. This purports to be the full, correct fix. -- JWL)Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
The fact that these are exported is a technical detail
of the conversion period -- we don't want anybody to
start relying on these. Ultimately we want things to
use cfg80211 only, and once everything that is in wext
is converted to cfg80211 drivers will not need to touch
wext _at all_.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
When we leave an IBSS, we should clear the SSID and not just the
BSSID, but since WEXT allows configuring while the interface is
down we must not clear it when leaving due to taking the iface
down, so some complications are needed.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Add new nl80211 attributes that can be used with NL80211_CMD_SET_WIPHY
and NL80211_CMD_GET_WIPHY to manage fragmentation/RTS threshold and
retry limits.Since these values are stored in struct wiphy, remove the local copy
from mac80211 where feasible (frag & rts threshold). The retry limits
are currently needed in struct ieee80211_conf, but these could be
eventually removed since the driver should have access to the values
in struct wiphy.Signed-off-by: Jouni Malinen
Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
This patch adds support for configuring HT40 channels
and receiving HT40 to ar9170. Receiving aggregation
doesn't seem to work right now, so it's not enabled.
Same goes for TX aggregation, but that probably needs
even more work.With this, I can receive roughly 33 Mbits/sec.
The HT capabilities are a little odd, I tried following
otus here -- in particular having SGI_40 but not SGI_20
is a little weird but afaict that's what otus does.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Supporting wireless extension nickname is pointless
and no other modern driver supports this, so remove
it.Signed-off-by: Johannes Berg
Acked-by: Jussi Kivilinna
Signed-off-by: John W. Linville -
Trying to separate header files into net/wireless.h and
net/cfg80211.h has been a source of confusion. Remove
net/wireless.h (because there also is the linux/wireless.h)
and subsume everything into net/cfg80211.h -- except the
definitions for regulatory structures which get moved to
a new header net/regulatory.h.The "new" net/cfg80211.h is now divided into sections.
There are no real changes in this patch but code shuffling
and some very minor documentation fixes.I have also, to make things reflect reality, put in a
copyright line for Luis to net/regulatory.h since that
is probably exclusively written by him but was formerly
in a file that only had my copyright line.Signed-off-by: Johannes Berg
Cc: Luis R. Rodriguez
Signed-off-by: John W. Linville -
This converts mac80211 to the new cfg80211 IBSS API, the
wext handling functions are called where appropriate.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
This adds IBSS API along with (preliminary) wext handlers.
The wext handlers can only do IBSS so you need to call them
from your own wext handlers if the mode is IBSS.The nl80211 API requires
* an SSID
* a channel (frequency) for the case that a new IBSS
has to be createdIt optionally supports
* a flag to fix the channel
* a fixed BSSIDThe cfg80211 code also takes care to leave the IBSS before
the netdev is set down. If wireless extensions are used, it
also caches values when the interface is down and instructs
the driver to join when the interface is set up.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Since we have ->deauth and ->disassoc we can support the
wext SIWMLME call directly without driver wext handlers.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
When listing all wireless netdevs in the system this
is useful to print which wiphy they belong to. Just
add the attribute, any program that doesn't care will
just ignore it.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
This patch sets IEEE80211_TX_CTL_CLEAR_PS_FILT for outgoing
frames for a half-wake station.this is necessary if one wants to get ps-poll working properly with a p54 ap.
Signed-off-by: Christian Lamparter
Signed-off-by: John W. Linville -
PS-Polled frames must be sent with OUT_NOCANCEL flag set,
or the firmware will reject all of them, at the station is still blacklisted.Signed-off-by: Christian Lamparter
Signed-off-by: John W. Linville -
The following patch implements some control over the LED on RTL8187B and
RTL8187L devices. Triggers are registered for TX and RX. Whenever the
trigger event occurs, the LED is turned off for 1/20 second, then turned
back on.Note: For those RTL8187X devices that are built into the computer and have
a LED that is expected to be controlled with a radio switch, this patch will
not operate that LED. That will take a separate patch to be prepared later.Signed-off-by: Larry Finger
Signed-off-by: Herton Ronaldo Krzesinski
Tested-by: Hin-Tak Leung
Signed-off-by: John W. Linville -
We can allow scan requests in AP mode as long as the interface has not
yet been configured to send out Beacon frames (or if beaconing has
been disabled prior to the scan request). This makes it easier to scan
for neighboring BSSes during AP initialization and makes it possible
to run a scan without setting the interface down, if needed. Without
this change, the only available option would be to set the interface
down, move into station mode, and set the interface up, prior to
requesting the scan.Signed-off-by: Jouni Malinen
Signed-off-by: John W. Linville -
Document what mac80211 will do in the future to help save power.
We're not quite there yet, but a plan helps. Also, while at it,
fix the docs wrt. multicast traffic.Signed-off-by: Johannes Berg
Reviewed-by: Kalle Valo
Signed-off-by: John W. Linville -
Enable PS by default (depending on Kconfig) -- rely on drivers
to control the level using pm_qos. Due to the previous patch
we turn off PS when necessary due to latency requirements.This has a Kconfig symbol so people can, if they really want,
configure the default in their kernels. We may want to keep it
at "default y" only in wireless-testing for a while.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
Regardless of whether the hardware implements beacon filtering,
there's no need to process all beacons in software all the time
throughout the stack (mac80211 does a lot, then cfg80211, then
in the future possibly userspace).This patch implements the "best possible" beacon filtering in
mac80211. "Best possible" means that it can look for changes in
all requested information elements, and distinguish vendor IEs
by their OUI.In the future, we will add nl80211 API for userspace to request
information elements and vendor IE OUIs to watch -- drivers can
then implement the best they can do while software implements
it fully.It is unclear whether or not this actually saves CPU time, but
the data is all in the cache already so it should be fairly
cheap. The additional _testing_, however, has great benefit;
Without this, and on hardware that doesn't implement beacon
filtering, wrong assumptions about, for example, scan result
updates could quickly creep into code.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
When an application asks for a latency lower than the beacon interval
there's nothing we can do -- we need to stay awake and not have the
AP buffer frames for us. Add code to automatically calculate this
constraint in mac80211 so drivers need not concern themselves with it.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
When you have multiple virtual interfaces the current
implementation requires setting them up properly from
userspace, which is undesirable when we want to default
to power save mode. Keep track of powersave requested
from userspace per managed mode interface, and only
enable powersave globally when exactly one managed mode
interface is active and has powersave turned on.Second, only start the dynPS timer when PS is turned
on, and properly turn it off when PS is turned off.Third, fix the scan_sdata abuse in the dynps code.
Finally, also reorder the code and refactor the code
that enables PS or the dynps timer instead of having
it copied in two places.Signed-off-by: Johannes Berg
Signed-off-by: John W. Linville -
A detection function was added for identifying CF8381.
Signed-off-by: Marek Vasut
Acked-by: Dan Williams
Signed-off-by: John W. Linville -
This patch deactivates powersave in station mode.
It does not work correctly yet, so the code does more harm than good.
(I split the original patch and sent part of it for 2.6.30, which didn't
have the IEEE80211_HW_BEACON_FILTER flag. -- JWL)Reported-by: Johannes Berg
Signed-off-by: Christian Lamparter
Signed-off-by: John W. Linville
