GitHub/exynos8895/android_kernel_samsung_universal8895.git
14 years agomwl8k: bump the transmit wait timeout to 5 seconds
Lennert Buytenhek [Fri, 8 Jan 2010 17:28:34 +0000 (18:28 +0100)]
mwl8k: bump the transmit wait timeout to 5 seconds

While it is reasonable to expect that at least one transmit ring entry
will be processed per second while we are waiting for the transmit
rings to drain, the firmware can end up doing batching of transmit ring
status writeback, which means that the transmit rings can appear stuck
for more than a second at a time.

Bump the TX drain wait timeout up from 1 to 5 seconds to account for
this.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: undo transmit queue 0/1 swapping in mwl8k_cmd_set_edca_params()
Lennert Buytenhek [Fri, 8 Jan 2010 17:28:14 +0000 (18:28 +0100)]
mwl8k: undo transmit queue 0/1 swapping in mwl8k_cmd_set_edca_params()

The comment and code in mwl8k_cmd_set_edca_params() suggest that the
mapping between SET_EDCA_PARAMS queue numbers and transmit rings isn't
actually 1:1, while tests show that the mapping is in fact 1:1.  So,
get rid of the transmit queue 0/1 swapping.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: minor cleanups
Lennert Buytenhek [Fri, 8 Jan 2010 17:27:59 +0000 (18:27 +0100)]
mwl8k: minor cleanups

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: do not transmit frames on unconfigured 4-addr vlan interfaces
Felix Fietkau [Fri, 8 Jan 2010 17:15:13 +0000 (18:15 +0100)]
mac80211: do not transmit frames on unconfigured 4-addr vlan interfaces

If frames are transmitted on 4-addr ap vlan interfaces with no station,
they end up being transmitted unencrypted, even if the ap interface
uses WPA. This patch add some sanity checking to make sure that this
does not happen.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: allow station updates on ap interfaces for vlan stations
Felix Fietkau [Fri, 8 Jan 2010 17:10:58 +0000 (18:10 +0100)]
mac80211: allow station updates on ap interfaces for vlan stations

Since the per-vif station changes, sta_info_get on the ap sdata no
longer returns entries for stations on ap vlans. This causes issues
with hostapd, which currently always passes the ap interface name to
nl80211 calls. This patch provides bug compatibility with the earlier
versions until hostapd is fixed.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: use nullfunc frames for 4-addr sta detection
Felix Fietkau [Fri, 8 Jan 2010 17:06:26 +0000 (18:06 +0100)]
mac80211: use nullfunc frames for 4-addr sta detection

To detect incoming 4-addr stations, hostapd needs to receive a 4-addr
data frame from the remote station, so that it can create the AP VLAN
for it. With this patch, the mlme code emits a 4-addr nullfunc frame
immediately after assoc. On the AP side it also drops 4-addr nullfunc
frames to the cooked monitor mode interface, if the interface hasn't
been fully set up to receive 4-addr data frames yet.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Fix monitor mode handling
Sujith [Fri, 8 Jan 2010 05:06:13 +0000 (10:36 +0530)]
ath9k: Fix monitor mode handling

mac80211 passes appropriate flags indicating whether
monitor mode is being used. Use this to set the HW opmode.

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Add debugfs file for RX errors
Sujith [Fri, 8 Jan 2010 05:06:11 +0000 (10:36 +0530)]
ath9k: Add debugfs file for RX errors

This file can be used to track frame reception errors.
PHY error counts are also added.

Location: ath9k/phy#/recv

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Fix queue handling
Sujith [Fri, 8 Jan 2010 05:06:09 +0000 (10:36 +0530)]
ath9k: Fix queue handling

The TX queues have to be stopped during an
internal reset. Not handling this would result
in packet loss - fix this.

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Cleanup init/deinit routines
Sujith [Fri, 8 Jan 2010 05:06:07 +0000 (10:36 +0530)]
ath9k: Cleanup init/deinit routines

The device initialization and termination functions
were messy and convoluted. Introduce helper functions
to clarify init_softc() and simplify things in general.

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Cleanup Powersave flags
Sujith [Fri, 8 Jan 2010 05:06:05 +0000 (10:36 +0530)]
ath9k: Cleanup Powersave flags

sc_flags has slowly become a kitchen sink over time.
Move powersave related flags to a separate variable.

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Add new file init.c
Sujith [Fri, 8 Jan 2010 05:06:02 +0000 (10:36 +0530)]
ath9k: Add new file init.c

Move initialization/de-initialization related
code to this file.

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Add a new file for GPIO
Sujith [Fri, 8 Jan 2010 05:06:00 +0000 (10:36 +0530)]
ath9k: Add a new file for GPIO

Move all LED/RFKILL/BTCOEX related code
to gpio.c

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agocfg80211: Fix country IE parsing for single channel triplets
Luis R. Rodriguez [Thu, 7 Jan 2010 22:24:57 +0000 (17:24 -0500)]
cfg80211: Fix country IE parsing for single channel triplets

This enhances the way we parse country IEs to minimize
the number of regulatory rules that we create. It also fixes
our current implementation which treated country IE triplets
with only one channel as one independed regulatory rule even
though adjecent rules were also being provided.

Without this patch APs which send country IE information with
a channel triplet for each individual channel will force cfg80211
to deny HT40 operation as a regulatory rule would have been created
independently for each channel and as such configured only for
20 MHz operation.

Although 802.11n APs which send country IEs triplets in this fassion
are likely rare Benoit reports this against the Ubiquity NanoStation M5,
with Country "FR" and HT40 enabled.

Since we now have a helper which parses the triplets in intermediate
steps we now take care extra care to process padding.

Reported-by: Benoit PAPILLAULT <benoit.papillault@free.fr>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agocfg80211: process the max power on a country IE
Luis R. Rodriguez [Thu, 7 Jan 2010 22:24:56 +0000 (17:24 -0500)]
cfg80211: process the max power on a country IE

The max power from each country IE triplet was being ignored.
This fix isn't critical as CRDA was always being used for the lower
limit, but we should process it in case the AP still wants to
decrease power output even more for whatever reason.

Reported-by: Benoit PAPILLAULT <benoit.papillault@free.fr>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agocfg80211: add debug print when we drop a bogus country IE
Luis R. Rodriguez [Thu, 7 Jan 2010 22:24:54 +0000 (17:24 -0500)]
cfg80211: add debug print when we drop a bogus country IE

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agob43: LP-PHY: note and explain specs inconsistency
Rafał Miłecki [Thu, 7 Jan 2010 13:09:27 +0000 (14:09 +0100)]
b43: LP-PHY: note and explain specs inconsistency

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: fix a few work bugs
Johannes Berg [Wed, 6 Jan 2010 14:30:58 +0000 (15:30 +0100)]
mac80211: fix a few work bugs

Kalle and Lennert reported problems with the new work
code, and at least Kalle's problem I was able to trace
to a missing jiffies initialisation.

I also ran into a problem where occasionally I couldn't
connect, which seems fixed with kicking the work items
after scanning.

Finally, also add some sanity checking code to verify
that we're not adding work items while an interface is
down -- that case could lead to something similar to
what Lennert was seeing.

There still seems to be a race condition that we're
trying to figure out separately.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Tested-by: Lennert Buytenhek <buytenh@marvell.com>
Tested-by: Kalle Valo <kalle.valo@iki.fi>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agocfg80211: Store IEs from both Beacon and Probe Response frames
Jouni Malinen [Wed, 6 Jan 2010 14:19:24 +0000 (16:19 +0200)]
cfg80211: Store IEs from both Beacon and Probe Response frames

Store information elements from Beacon and Probe Response frames in
separate buffers to allow both sets to be made available through
nl80211. This allows user space applications to get access to IEs from
Beacon frames even if we have received Probe Response frames from the
BSS. Previously, the IEs from Probe Response frames would have
overridden the IEs from Beacon frames.

This feature is of somewhat limited use since most protocols include
the same (or extended) information in Probe Response frames. However,
there are couple of exceptions where the IEs from Beacon frames could
be of some use: TIM IE is only included in Beacon frames (and it would
be needed to figure out the DTIM period used in the BSS) and at least
some implementations of Wireless Provisioning Services seem to include
the full IE only in Beacon frames).

The new BSS attribute for scan results is added to allow both the IE
sets to be delivered. This is done in a way that maintains the
previously used behavior for applications that are not aware of the
new NL80211_BSS_BEACON_IES attribute.

Signed-off-by: Jouni Malinen <j@w1.fi>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agowl1251: fix sleep related error paths in wl1251_op_bss_info_changed()
Kalle Valo [Tue, 5 Jan 2010 18:17:10 +0000 (20:17 +0200)]
wl1251: fix sleep related error paths in wl1251_op_bss_info_changed()

In various cases wl1251_op_bss_info_changed() did not call elp_sleep()
after an error was noticed. Fix it by using correct goto label.

The bug was a theoretical one, in practise it doesn't matter because
if commands start returning errors there will be lots of other problems.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agowl1251: use mac80211 provided channel parameters in scanning
Kalle Valo [Tue, 5 Jan 2010 18:17:03 +0000 (20:17 +0200)]
wl1251: use mac80211 provided channel parameters in scanning

The number of channels to be used in scan was hard coded in wl1251. The
proper way is to use the channels array provided by mac80211.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agowl1251: get probe request template from mac80211
Kalle Valo [Tue, 5 Jan 2010 18:16:57 +0000 (20:16 +0200)]
wl1251: get probe request template from mac80211

Instead of creating the template in driver, get it from mac80211 instead.
Thanks to this, three functions can be now removed.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agowl1251: cleanup scanning code
Kalle Valo [Tue, 5 Jan 2010 18:16:51 +0000 (20:16 +0200)]
wl1251: cleanup scanning code

The current scanning code wasn't following the preferred style. Move code
related to scan and trigger scan to commans to wl1251_cmd.c. Because
there's now less code in wl1251_hw_scan(), the function can be now
merged with wl1251_op_hw_scan().

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: use Probe Request template when sending a direct scan
Kalle Valo [Tue, 5 Jan 2010 18:16:44 +0000 (20:16 +0200)]
mac80211: use Probe Request template when sending a direct scan

As mac80211 now has a separate function for creating Probe Request templates,
better to use it when sending direct Probe Requests to an AP. Only the
bssid needs to be updated in the template before sending it.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: create Probe Request template
Kalle Valo [Tue, 5 Jan 2010 18:16:38 +0000 (20:16 +0200)]
mac80211: create Probe Request template

Certain type of hardware, for example wl1251 and wl1271, need a template
for the Probe Request. Create a function ieee80211_probereq_get() which
creates the template and drivers send it to hardware.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agowl1251: get PS Poll and Nullfunc templates from mac80211
Kalle Valo [Tue, 5 Jan 2010 18:16:32 +0000 (20:16 +0200)]
wl1251: get PS Poll and Nullfunc templates from mac80211

Now that mac80211 creates templates for PS Poll and Nullfunc frames, use
them instead of creating our own.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: use PS Poll and Nullfunc templates when sending such frames
Kalle Valo [Tue, 5 Jan 2010 18:16:26 +0000 (20:16 +0200)]
mac80211: use PS Poll and Nullfunc templates when sending such frames

To avoid duplicate code, use ieee80211_[pspoll|nullfunc]_get() to get
templates for PS Poll and Nullfunc frames in mlme.c.

Compile-tested only.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: add functions to create PS Poll and Nullfunc templates
Kalle Valo [Tue, 5 Jan 2010 18:16:19 +0000 (20:16 +0200)]
mac80211: add functions to create PS Poll and Nullfunc templates

Some hardware, for example wl1251 and wl1271, handle the transmission
of power save related frames in hardware, but the driver is responsible
for creating the templates. It's better to create the templates in mac80211,
that way all drivers can benefit from this.

Add two new functions, ieee80211_pspoll_get() and ieee80211_nullfunc_get()
which drivers need to call to get the frame. Drivers are also responsible
for updating the templates after each association.

Also new struct ieee80211_hdr_3addr is added to ieee80211.h to make it
easy to calculate length of the Nullfunc frame.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agort2x00: Make rt2800_init_led static and don't export it.
Gertjan van Wingerde [Wed, 30 Dec 2009 10:36:34 +0000 (11:36 +0100)]
rt2x00: Make rt2800_init_led static and don't export it.

It is only used within the rt2800lib module itself.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agort2x00: Unify rt2800 WPDMA ready waiting functions.
Gertjan van Wingerde [Wed, 30 Dec 2009 10:36:32 +0000 (11:36 +0100)]
rt2x00: Unify rt2800 WPDMA ready waiting functions.

The rt2800pci_wait_wpdma_ready and rt2800usb_wait_wpdma_ready functions are
exactly the same, so unify them into rt200lib.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agort2x00: RF3052 is a valid RF chipset for USB devices as well.
Gertjan van Wingerde [Wed, 30 Dec 2009 10:36:31 +0000 (11:36 +0100)]
rt2x00: RF3052 is a valid RF chipset for USB devices as well.

The RF3052 chipset is now also being integrated onto USB devices, so
allow the RF chipset and don't treat it as PCI/SOC only.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agonl80211: New command for setting TX rate mask for rate control
Jouni Malinen [Tue, 29 Dec 2009 10:59:45 +0000 (12:59 +0200)]
nl80211: New command for setting TX rate mask for rate control

Add a new NL80211_CMD_SET_TX_BITRATE_MASK command and related
attributes to provide support for setting TX rate mask for rate
control. This uses the existing cfg80211 set_bitrate_mask operation
that was previously used only with WEXT compat code (SIOCSIWRATE). The
nl80211 command allows more generic configuration of allowed rates as
a mask instead of fixed/max rate.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agocfg80211/mac80211: Use more generic bitrate mask for rate control
Jouni Malinen [Wed, 6 Jan 2010 11:09:08 +0000 (13:09 +0200)]
cfg80211/mac80211: Use more generic bitrate mask for rate control

Extend struct cfg80211_bitrate_mask to actually use a bitfield mask
instead of just a single fixed or maximum rate index. This change
itself does not modify the behavior (except for debugfs files), but it
prepares cfg80211 and mac80211 for a new nl80211 command for setting
which rates can be used in TX rate control.

Since frames are now going through the rate control algorithm
unconditionally, the internal IEEE80211_TX_INTFL_RCALGO flag can now
be removed. The RC implementations can use the rate_idx_mask value to
optimize their behavior if only a single rate is enabled.

The old max_rate_idx in struct ieee80211_tx_rate_control is maintained
(but commented as deprecated) for backwards compatibility with existing
RC implementations. Once these implementations have been updated to
use the more generic rate_idx_mask, the max_rate_idx value can be
removed.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: Select lowest rate based on basic rate set in AP mode
Jouni Malinen [Tue, 29 Dec 2009 10:59:19 +0000 (12:59 +0200)]
mac80211: Select lowest rate based on basic rate set in AP mode

If the basic rate set is configured to not include the lowest rate
(e.g., basic rate set = 6, 12, 24 Mbps in IEEE 802.11g mode), the AP
should not send out broadcast frames at 1 Mbps. This type of
configuration can be used to optimize channel usage in cases where
there is no need for backwards compatibility with IEEE 802.11b-only
devices.

In AP mode, mac80211 was unconditionally using the lowest rate for
Beacon frames and similarly, with all rate control algorithms that use
rate_control_send_low(), the lowest rate ended up being used for all
broadcast frames (and all unicast frames that are sent before
association). Change this to take into account the basic rate
configuration in AP mode, i.e., use the lowest rate in the basic rate
set instead of the lowest supported rate when selecting the rate.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: quit addba_resp_timer if Tx BA session is torn down
Zhu Yi [Mon, 28 Dec 2009 08:57:15 +0000 (16:57 +0800)]
mac80211: quit addba_resp_timer if Tx BA session is torn down

Make addba_resp_timer aware the HT_AGG_STATE_REQ_STOP_BA_MSK mask
so that when ___ieee80211_stop_tx_ba_session() is issued the timer
will quit. Otherwise when suspend happens before the timer expired,
the timer handler will be called immediately after resume and
messes up driver status.

Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: Implement mac80211 callback set_coverage_class
Lukáš Turek [Mon, 21 Dec 2009 21:50:51 +0000 (22:50 +0100)]
ath5k: Implement mac80211 callback set_coverage_class

The callback sets slot time as specified in IEEE 802.11-2007 section
17.3.8.6 (for 20MHz channels only for now) and raises ACK and CTS
timeouts accordingly. The values are persistent, they are restored after
device reset.

Signed-off-by: Lukas Turek <8an@praha12.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: Reimplement clock rate to usec conversion
Lukáš Turek [Mon, 21 Dec 2009 21:50:50 +0000 (22:50 +0100)]
ath5k: Reimplement clock rate to usec conversion

The original code was correct in 802.11a mode only, 802.11b/g uses
different clock rates. The new code uses values taken from FreeBSD HAL
and should be correct for all modes including turbo modes.

The former rate calculation was used by slope coefficient calculation
function ath5k_hw_write_ofdm_timings. However, this function requires
the 802.11a values even in 802.11g mode. Thus the use of
ath5k_hw_htoclock was replaced by hardcoded values. Possibly the slope
coefficient calculation is not related to clock rate at all.

Signed-off-by: Lukas Turek <8an@praha12.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: Fix functions for getting/setting slot time
Lukáš Turek [Mon, 21 Dec 2009 21:50:49 +0000 (22:50 +0100)]
ath5k: Fix functions for getting/setting slot time

Functions ath5k_hw_get_slot_time and ath5k_hw_set_slot_time were
converting microseconds to clocks only for AR5210, although it's needed
for all supported devices. The conversion was moved outside the
hardware-specific branches.

The original code also limited minimum slot time to 9, while turbo modes
use 6, this was fixed too.

Signed-off-by: Lukas Turek <8an@praha12.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: Add new callback set_coverage_class
Lukáš Turek [Mon, 21 Dec 2009 21:50:48 +0000 (22:50 +0100)]
mac80211: Add new callback set_coverage_class

Mac80211 callback to driver set_coverage_class() sets slot time and ACK
timeout for given IEEE 802.11 coverage class. The callback is optional,
but it's essential for long distance links.

Signed-off-by: Lukas Turek <8an@praha12.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agonl80211: Add new WIPHY attribute COVERAGE_CLASS
Lukáš Turek [Mon, 21 Dec 2009 21:50:47 +0000 (22:50 +0100)]
nl80211: Add new WIPHY attribute COVERAGE_CLASS

The new attribute NL80211_ATTR_WIPHY_COVERAGE_CLASS sets IEEE 802.11
Coverage Class, which depends on maximum distance of nodes in a
wireless network. It's required for long distance links (more than a few
hundred meters).

The attribute is now ignored by two non-mac80211 drivers, rndis and
iwmc3200wifi, together with WIPHY_PARAM_RETRY_SHORT and
WIPHY_PARAM_RETRY_LONG. If it turns out to be a problem, we could split
set_wiphy_params callback or add new capability bits.

Signed-off-by: Lukas Turek <8an@praha12.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: re-enable ps by default for new single chip families
Luis R. Rodriguez [Fri, 18 Dec 2009 16:26:04 +0000 (11:26 -0500)]
ath9k: re-enable ps by default for new single chip families

The newer single chip hardware family of chipsets have not been
experiencing issues with power saving set by default with recent
fixes merged (even into stable). The remaining issues are only
reported with AR5416 and since enabling PS by default can increase
power savings considerably best to take advantage of that feature
as this has been tested properly.

For more details on this issue see the bug report:

http://bugzilla.kernel.org/show_bug.cgi?id=14267

We leave AR5416 with PS disabled by default, that seems to require
some more work.

Cc: stable@kernel.org
Cc: Peter Stuge <peter@stuge.se>
Cc: Justin P. Mattock <justinmattock@gmail.com>
Cc: Kristoffer Ericson <kristoffer.ericson@gmail.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6
John W. Linville [Wed, 6 Jan 2010 22:22:54 +0000 (17:22 -0500)]
Merge git://git./linux/kernel/git/linville/wireless-2.6

Conflicts:
net/mac80211/scan.c
net/mac80211/wme.c

14 years agoRevert "mac80211: replace netif_tx_{start,stop,wake}_all_queues"
John W. Linville [Wed, 6 Jan 2010 20:39:39 +0000 (15:39 -0500)]
Revert "mac80211: replace netif_tx_{start,stop,wake}_all_queues"

This reverts commit 53623f1a09c7a7d23b74f0f7d93dba0ebde1006b.

This was inadvertantly missed in "mac80211: fix skb buffering issue",
and is required with that patch to restore proper queue operation.

Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: fix-up build breakage in 2.6.33
John W. Linville [Wed, 6 Jan 2010 20:35:49 +0000 (15:35 -0500)]
mac80211: fix-up build breakage in 2.6.33

"mac80211: fix skb buffering issue" is based on what will become 2.6.34,
so it includes an incompatible usage of sta_info_get.  This patch will
need to be effectively reverted when merging for 2.6.34.

Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agowireless: fix build breakage when CONFIG_CFG80211_REG_DEBUG not set
John W. Linville [Tue, 5 Jan 2010 22:57:20 +0000 (17:57 -0500)]
wireless: fix build breakage when CONFIG_CFG80211_REG_DEBUG not set

Bad macro definition in "cfg80211: add a regulatory debug print"...

Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6
John W. Linville [Tue, 5 Jan 2010 22:18:59 +0000 (17:18 -0500)]
Merge git://git./linux/kernel/git/linville/wireless-2.6

Conflicts:
net/mac80211/iface.c

14 years agomac80211: fix ampdu_action tx_start ssn
Christian Lamparter [Sun, 3 Jan 2010 23:52:56 +0000 (00:52 +0100)]
mac80211: fix ampdu_action tx_start ssn

The start_seq_num is taken from the station's tid_seq[tid].
This is fine, except tid_seq sequence counter is shifted
by 4 bits to accommodate for frame fragmentation.

Both (iwlagn & ath9k) were unaffected by this minor glitch,
because they don't read the *ssn for the AMPDU_TX_START action.

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: pass GET_HW_SPEC capability bitmask up the stack
Lennert Buytenhek [Mon, 4 Jan 2010 20:58:40 +0000 (21:58 +0100)]
mwl8k: pass GET_HW_SPEC capability bitmask up the stack

This enables HT association and AMPDU in the receive direction for
STA firmware images on hardware that supports it.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: trivial rx-only ampdu implementation
Lennert Buytenhek [Mon, 4 Jan 2010 20:58:12 +0000 (21:58 +0100)]
mwl8k: trivial rx-only ampdu implementation

AMPDU receive doesn't need any special handling, so let's enable
this before tackling the transmit side.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: pass in HT capabilities and rates when associating
Lennert Buytenhek [Mon, 4 Jan 2010 20:57:59 +0000 (21:57 +0100)]
mwl8k: pass in HT capabilities and rates when associating

Pass the AP's MCS rate mask to SET_RATE when associating, and make
UPDATE_STADB pass in the peer's HT caps and rates when adding a new
hardware station database entry.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: allow setting HT channels
Lennert Buytenhek [Mon, 4 Jan 2010 20:56:46 +0000 (21:56 +0100)]
mwl8k: allow setting HT channels

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: add support for 88w8363 in STA mode
Lennert Buytenhek [Mon, 4 Jan 2010 20:56:19 +0000 (21:56 +0100)]
mwl8k: add support for 88w8363 in STA mode

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: fix changed flags handling in mwl8k_bss_info_changed()
Lennert Buytenhek [Mon, 4 Jan 2010 20:56:07 +0000 (21:56 +0100)]
mwl8k: fix changed flags handling in mwl8k_bss_info_changed()

Previously, mwl8k_bss_info_changed() would refuse to do anything if
the 'changed' argument indicated that the association status hadn't
changed.  Fix this up so that it will allow changing things like the
preamble type, the slot time and the CTS-to-self protection method
without having to reassociate.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: honor peer rate set
Lennert Buytenhek [Mon, 4 Jan 2010 20:55:52 +0000 (21:55 +0100)]
mwl8k: honor peer rate set

When calling SET_RATE, SET_AID, or when creating a station database
entry for our AP, pass in the AP's rate set instead of just blindly
enabling all legacy rates, so as to end up doing the right thing when
talking to 11b-only APs.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: move ->peer_id from mwl8k_vif to mwl8k_sta
Lennert Buytenhek [Mon, 4 Jan 2010 20:55:42 +0000 (21:55 +0100)]
mwl8k: move ->peer_id from mwl8k_vif to mwl8k_sta

For STA firmware, move the per-peer hardware station ID to the
driver-private part of struct ieee80211_sta, where it belongs.

(Since issuing a hardware station database maintenance command sleeps,
we can't hold a reference to the ieee80211_sta * across the command,
and since we won't know the station ID until after the command
completes, we need to re-lookup the sta when the command is done to
write the returned station ID back to its driver-private part.)

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: remove mwl8k_vif::bssid, which is now useless
Lennert Buytenhek [Mon, 4 Jan 2010 20:55:21 +0000 (21:55 +0100)]
mwl8k: remove mwl8k_vif::bssid, which is now useless

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: handle station database update for AP's sta entry via ->sta_notify()
Lennert Buytenhek [Mon, 4 Jan 2010 20:55:12 +0000 (21:55 +0100)]
mwl8k: handle station database update for AP's sta entry via ->sta_notify()

Inserting and removing a hardware station database entry for the AP
when we are in managed mode is currently done in ->bss_info_changed().

To prepare for adding AP mode support, implement the ->sta_notify()
driver method, and let that handle inserting and removing the hardware
station database entry for our AP instead.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: move struct peer_capability_info to its only user
Lennert Buytenhek [Mon, 4 Jan 2010 20:54:41 +0000 (21:54 +0100)]
mwl8k: move struct peer_capability_info to its only user

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: make the tx ring drain status messages somewhat more friendly
Lennert Buytenhek [Mon, 4 Jan 2010 20:54:25 +0000 (21:54 +0100)]
mwl8k: make the tx ring drain status messages somewhat more friendly

Old:
> phy0: timeout waiting for tx rings to drain (9 -> 5 pkts), retrying
> phy0: timeout waiting for tx rings to drain (5 -> 2 pkts), retrying
> phy0: tx rings drained

New:
> phy0: waiting for tx rings to drain (9 -> 5 pkts)
> phy0: waiting for tx rings to drain (5 -> 2 pkts)
> phy0: tx rings drained

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: prevent freeing free IRQ if ieee80211_register_hw() fails
Lennert Buytenhek [Mon, 4 Jan 2010 20:54:08 +0000 (21:54 +0100)]
mwl8k: prevent freeing free IRQ if ieee80211_register_hw() fails

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: bail out if there is no AP firmware image support for this chip
Lennert Buytenhek [Mon, 4 Jan 2010 20:53:54 +0000 (21:53 +0100)]
mwl8k: bail out if there is no AP firmware image support for this chip

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agocfg80211: add a regulatory debug print
Luis R. Rodriguez [Mon, 4 Jan 2010 16:50:11 +0000 (11:50 -0500)]
cfg80211: add a regulatory debug print

Instead of sprinkling code with ifdef's define REG_DBG_PRINT() instead.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: update version number to 0.11
Lennert Buytenhek [Sat, 2 Jan 2010 09:31:50 +0000 (10:31 +0100)]
mwl8k: update version number to 0.11

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: fix ieee80211_change_mac() to use struct sockaddr
Kalle Valo [Wed, 30 Dec 2009 13:54:03 +0000 (15:54 +0200)]
mac80211: fix ieee80211_change_mac() to use struct sockaddr

Setting the mac address from user space was buggy. For example, when
executing this command:

ip link set wlan0 address 00:1f:df:88:cd:55

mac80211 used the address 01:00:00:1f:df:88 instead. It was shifted two
bytes.

The reason was that the addr (type of void *) provided to
ieee80211_change_mac() is actually of type struct sockaddr, not just the
mac address array. Also the call to eth_mac_addr() expects the address to
be struct sockaddr.

Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: fix a warning related to pointer conversion to u64 cookie
Kalle Valo [Wed, 30 Dec 2009 12:42:20 +0000 (14:42 +0200)]
mac80211: fix a warning related to pointer conversion to u64 cookie

On a 32 bit system (in this case an omap 3430 system) gcc warned about
pointer conversion:

net/mac80211/work.c: In function 'ieee80211_remain_on_channel_timeout':
net/mac80211/work.c:534: warning: cast from pointer to integer of different size
net/mac80211/work.c: In function 'ieee80211_remain_done':
net/mac80211/work.c:1030: warning: cast from pointer to integer of different size
net/mac80211/work.c: In function 'ieee80211_wk_remain_on_channel':
net/mac80211/work.c:1056: warning: cast from pointer to integer of different size
net/mac80211/work.c: In function 'ieee80211_wk_cancel_remain_on_channel':
net/mac80211/work.c:1072: warning: cast from pointer to integer of different size

Fix it by casting the pointers to unsigned long instead. This makes the
compiler happy again.

Compile-tested only.

Cc: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: No need to include WEXT headers here
Jouni Malinen [Tue, 29 Dec 2009 10:43:58 +0000 (12:43 +0200)]
mac80211: No need to include WEXT headers here

Remove the forgotten linux/wireless.h inclusion from mac80211.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwlwifi: remove linux/utsrelease.h dependency
Zhu Yi [Thu, 24 Dec 2009 07:31:10 +0000 (15:31 +0800)]
iwlwifi: remove linux/utsrelease.h dependency

Commit 250cce26d5d03337aec4ff8405121f026adb4a89 uses UTS_RELEASE
as the the in-tree iwlwifi driver version. However the inclusion
of generated/utsrelease.h makes it a unpleasant behaviour to
recompile the driver everytime when utsrelease.h is updated. In
fact, the driver module is already built with the UTS_RELEASE
information via vermagic of modinfo. Mark the in-tree driver
with the version string "in-tree" to distinguish with those old
out-of-tree drivers.

Reported-by: David Miller <davem@davemloft.net>
Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agort2x00: Properly request tx headroom for alignment operations.
Gertjan van Wingerde [Wed, 30 Dec 2009 10:36:30 +0000 (11:36 +0100)]
rt2x00: Properly request tx headroom for alignment operations.

Current rt2x00 drivers may result in a "ieee80211_tx_status: headroom too
small" error message when a frame needs to be properly aligned before
transmitting it.
This is because the space needed to ensure proper alignment isn't
requested from mac80211.
Fix this by adding sufficient amount of alignment space to the amount
of headroom requested for TX frames.

Reported-by: David Ellingsworth <david@identd.dyndns.org>
Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <ivdoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: fix skb buffering issue
Johannes Berg [Tue, 5 Jan 2010 17:00:58 +0000 (18:00 +0100)]
mac80211: fix skb buffering issue

Since I removed the master netdev, we've been
keeping internal queues only, and even before
that we never told the networking stack above
the virtual interfaces about congestion. This
means that packets are queued in mac80211 and
the upper layers never know, possibly leading
to memory exhaustion and other problems.

This patch makes all interfaces multiqueue and
uses ndo_select_queue to put the packets into
queues per AC. Additionally, when the driver
stops a queue, we now stop all corresponding
queues for the virtual interfaces as well.

The injection case will use VO by default for
non-data frames, and BE for data frames, but
downgrade any data frames according to ACM. It
needs to be fleshed out in the future to allow
chosing the queue/AC in radiotap.

Reported-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Cc: stable@kernel.org [2.6.32]
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agort2x00: Fix LED configuration setting for rt2800.
Gertjan van Wingerde [Wed, 30 Dec 2009 10:36:33 +0000 (11:36 +0100)]
rt2x00: Fix LED configuration setting for rt2800.

rt2800_blink_set uses an illegal value to set the LED_CFG_G_LED_MODE
field of the LED_CFG register. This field is only 2 bits large, so
should be initialized with value that fits. Use default value from
the vendor driver.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agort2x00: use correct headroom for transmission
Pavel Roskin [Wed, 30 Dec 2009 10:36:29 +0000 (11:36 +0100)]
rt2x00: use correct headroom for transmission

Use rt2x00dev->ops->extra_tx_headroom, not rt2x00dev->hw->extra_tx_headroom
in the tx code, as the later may include other headroom not to be used in
the chipset driver.

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agocfg80211: fix syntax error on user regulatory hints
Luis R. Rodriguez [Mon, 4 Jan 2010 16:37:39 +0000 (11:37 -0500)]
cfg80211: fix syntax error on user regulatory hints

This fixes a syntax error when setting up the user regulatory
hint. This change yields the same exact binary object though
so it ends up just being a syntax typo fix, fortunately.

Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath5k: Fix eeprom checksum check for custom sized eeproms
Luis R. Rodriguez [Mon, 4 Jan 2010 15:40:39 +0000 (10:40 -0500)]
ath5k: Fix eeprom checksum check for custom sized eeproms

Commit 8bf3d79bc401ca417ccf9fc076d3295d1a71dbf5 enabled EEPROM
checksum checks to avoid bogus bug reports but failed to address
updating the code to consider devices with custom EEPROM sizes.
Devices with custom sized EEPROMs have the upper limit size stuffed
in the EEPROM. Use this as the upper limit instead of the static
default size. In case of a checksum error also provide back the
max size and whether or not this was the default size or a custom
one. If the EEPROM is busted we add a failsafe check to ensure
we don't loop forever or try to read bogus areas of hardware.

This closes bug 14874

http://bugzilla.kernel.org/show_bug.cgi?id=14874

Cc: stable@kernel.org
Cc: David Quan <david.quan@atheros.com>
Cc: Stephen Beahm <stephenbeahm@comcast.net>
Reported-by: Joshua Covington <joshuacov@googlemail.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agozd1211rw: Fix multicast filtering.
Benoit Papillault [Sun, 3 Jan 2010 09:20:01 +0000 (10:20 +0100)]
zd1211rw: Fix multicast filtering.

If multicast parameter (as returned by zd_op_prepare_multicast) has
changed, no bit in changed_flags is set. To handle this situation, we do
not return if changed_flags is 0. If we do so, we will have some issue
with IPv6 which uses multicast for link layer address resolution.

Signed-off-by: Benoit Papillault <benoit.papillault@free.fr>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoiwl: off by one bug
Dan Carpenter [Sun, 3 Jan 2010 09:19:35 +0000 (11:19 +0200)]
iwl: off by one bug

tid is used as an array offset.
agg = &priv->stations[sta_id].tid[tid].agg;
iwl4965_tx_status_reply_tx(priv, agg, tx_resp, txq_id, index);

It should be limitted to MAX_TID_COUNT - 1;
        struct iwl_tid_data tid[MAX_TID_COUNT];

regards,
dan carpenter

Signed-off-by: Dan Carpenter <error27@gmail.com>
CC: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomwl8k: fix configure_filter() memory leak on error
Lennert Buytenhek [Sat, 2 Jan 2010 09:31:42 +0000 (10:31 +0100)]
mwl8k: fix configure_filter() memory leak on error

If there was an error acquiring the firmware lock in
mwl8k_configure_filter(), we would end up leaking the multicast
command packet prepared by mwl8k_prepare_multicast().

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: fix ito64
Ming Lei [Tue, 29 Dec 2009 14:57:28 +0000 (22:57 +0800)]
ath9k: fix ito64

The unit of sizeof() is byte instead of bit, so fix it.
The patch can fix debug output of some dma_addr_t variables.

Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoath9k: Fix Kconfig depends for ATH9K_DEBUGFS
Dominik Geyer [Tue, 29 Dec 2009 07:27:57 +0000 (08:27 +0100)]
ath9k: Fix Kconfig depends for ATH9K_DEBUGFS

Add missing DEBUG_FS dependency for ATH9K_DEBUGFS in ath9k's Kconfig.

Signed-off-by: Dominik D. Geyer <dominik.geyer@gmx.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: fix typo added by "mac80211: fix propagation of failed..."
John W. Linville [Mon, 4 Jan 2010 20:58:03 +0000 (15:58 -0500)]
mac80211: fix typo added by "mac80211: fix propagation of failed..."

'Typo: it's "Hardware", not "Harware". Hmm, sometimes it's hairware :-)"'
-- Holger Schurig

Reported-by: Holger Schurig <holgerschurig@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agowl1271: fix timeout in wl1271_top_reg_read
Juha Leppanen [Mon, 4 Jan 2010 20:52:50 +0000 (15:52 -0500)]
wl1271: fix timeout in wl1271_top_reg_read

I noticed a timeout bug in

/drivers/net/wireless/wl12xx/wl1271_spi.c

In the current code you cannot tell why you exited
the "poll for data ready" do-while loop if exiting
was done after the last possible loop.

Then timeout==0 regardless of (val & OCP_READY_MASK) or
!(val & OCP_READY_MASK), leading to possible false timeout...

Simple correction could be decreasing timeout after checking
for !(val & OCP_READY_MASK), not before

(Manually converted from email to an actual patch by me. -- JWL)

Reported-by: "Juha Leppanen" <juha_motorsportcom@luukku.com>
Signed-off-by: "Juha Leppanen" <juha_motorsportcom@luukku.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-2.6
John W. Linville [Wed, 30 Dec 2009 20:25:08 +0000 (15:25 -0500)]
Merge git://git./linux/kernel/git/linville/wireless-2.6

Conflicts:
drivers/net/wireless/libertas/scan.c

14 years agoRevert "b43: Enforce DMA descriptor memory constraints"
John W. Linville [Tue, 29 Dec 2009 19:07:42 +0000 (14:07 -0500)]
Revert "b43: Enforce DMA descriptor memory constraints"

This reverts commit 9bd568a50c446433038dec2a5186c5c57c3dbd23.

That commit is shown to cause allocation failures during initialization
on some machines.

http://bugzilla.kernel.org/show_bug.cgi?id=14844

Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: annotate sleeping driver ops
Kalle Valo [Wed, 23 Dec 2009 12:15:47 +0000 (13:15 +0100)]
mac80211: annotate sleeping driver ops

To make it easier to notice cases of calling sleeping ops in atomic context,
annotate driver-ops.h with appropiate might_sleep() calls. At the same time,
also document in mac80211.h the op functions with missing contexts.

mac80211 doesn't seem to use get_tx_stats anywhere currently. Just to be on
the safe side, I documented it to be atomic, but hopefully the op can be
removed in the future.

Compile-tested only.

Signed-off-by: Kalle Valo <kalle.valo@iki.fi>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: remove requeue from work
Johannes Berg [Wed, 23 Dec 2009 12:15:46 +0000 (13:15 +0100)]
mac80211: remove requeue from work

There's no need to be requeueing the work struct
since we check for the scan after removing items
due to possible timeouts.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: remove struct ieee80211_if_init_conf
Johannes Berg [Wed, 23 Dec 2009 12:15:45 +0000 (13:15 +0100)]
mac80211: remove struct ieee80211_if_init_conf

All its members (vif, mac_addr, type) are now available
in the vif struct directly, so we can pass that instead
of the conf struct. I generated this patch (except the
mac80211 and header file changes) with this semantic
patch:

@@
identifier conf, fn, hw;
type tp;
@@
tp fn(struct ieee80211_hw *hw,
-struct ieee80211_if_init_conf *conf)
+struct ieee80211_vif *vif)
{
<...
(
-conf->type
+vif->type
|
-conf->mac_addr
+vif->addr
|
-conf->vif
+vif
)
...>
}

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211/cfg80211: add station events
Johannes Berg [Wed, 23 Dec 2009 12:15:44 +0000 (13:15 +0100)]
mac80211/cfg80211: add station events

When, for instance, a new IBSS peer is found, userspace
wants to be notified. Add events for all new stations
that mac80211 learns about.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: make off-channel work generic
Johannes Berg [Wed, 23 Dec 2009 12:15:43 +0000 (13:15 +0100)]
mac80211: make off-channel work generic

This changes mac80211 to allow being off-channel for
any type of work, not just the 'remain-on-channel'
work. This also helps fast transition to a BSS on a
different channel.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: support remain-on-channel command
Johannes Berg [Wed, 23 Dec 2009 12:15:42 +0000 (13:15 +0100)]
mac80211: support remain-on-channel command

This implements the new remain-on-channel cfg80211
command in mac80211, extending the work interface.

Also change the work purge code to be able to clean
up events properly (pretending they timed out.)

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agocfg80211: add remain-on-channel command
Jouni Malinen [Wed, 23 Dec 2009 12:15:41 +0000 (13:15 +0100)]
cfg80211: add remain-on-channel command

Add new commands for requesting the driver to remain awake
on a specified channel for the specified amount of time
(and another command to cancel such an operation). This
can be used to implement userspace-controlled off-channel
operations, like Public Action frame exchange on another
channel than the operation channel.

The off-channel operation should behave similarly to scan,
i.e. the local station (if associated) moves into power
save mode to request the AP to buffer frames for it and
then moves to the other channel to allow the off-channel
operation to be completed. The duration parameter can be
used to request enough time to receive a response from
the target station.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: Generalize off-channel operation helpers from scan code
Jouni Malinen [Wed, 23 Dec 2009 12:15:40 +0000 (13:15 +0100)]
mac80211: Generalize off-channel operation helpers from scan code

The off-channel operations for going into power save mode (station
mode) or stop beaconing (AP/IBSS) are not limited to scanning. Move
these into a separate file and allow them to be used for other
purposes, too.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: proper bss private data handling
Johannes Berg [Wed, 23 Dec 2009 12:15:39 +0000 (13:15 +0100)]
mac80211: proper bss private data handling

cfg80211 offers private data for each BSS struct,
which mac80211 uses. However, mac80211 uses internal
and external (cfg80211) BSS pointers interchangeably
and has a hack to put the cfg80211 bss struct into
the private struct.

Remove this hack, properly converting between the
pointers wherever necessary.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: split up and insert custom IEs correctly
Johannes Berg [Wed, 23 Dec 2009 12:15:38 +0000 (13:15 +0100)]
mac80211: split up and insert custom IEs correctly

Currently, we insert all user-specified IEs before the HT
IE for association, and after the HT IE for probe requests.
For association, that's correct only if the user-specified
IEs are RSN only, incorrect in all other cases including
WPA. Change this to split apart the user-specified IEs in
two places for association: before the HT IE (e.g. RSN),
after the HT IE (generally empty right now I think?) and
after WMM (all other vendor-specific IEs). For probes,
split the IEs in different places to be correct according
to the spec.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: refactor association
Johannes Berg [Wed, 23 Dec 2009 12:15:37 +0000 (13:15 +0100)]
mac80211: refactor association

Refactor the code to reserve an skb of the right size
(instead of hoping 200 bytes are enough forever), and
also put HT IE generation into an own function.

Additionally, put the HT IE before the vendor-specific
WMM IE. This still leaves things not quite ordered
correctly, due to user-specified IEs, add a note about
that for now.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: rewrite a few work messages
Johannes Berg [Wed, 23 Dec 2009 12:15:36 +0000 (13:15 +0100)]
mac80211: rewrite a few work messages

The station we're authenticating/associating with
may not always be an AP in the sense that word is
mostly understood, so print only the MAC address
of the peer instead.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: generalise work handling
Johannes Berg [Wed, 23 Dec 2009 12:15:35 +0000 (13:15 +0100)]
mac80211: generalise work handling

In order to use auth/assoc for different purposes
other than MLME, it needs to be split up. For other
purposes, a generic work handling (potentially on
another channel) will be useful.

To achieve that, this patch moves much of the MLME
work handling out of mlme into a new work API. The
API can currently handle probing a specific AP,
authentication and association. The MLME previously
handled probe/authentication as one step and will
continue to do so, but they are separate in the new
work handling.

Work items are RCU-managed to be able to check for
existence of an item for a specific frame in the RX
path, but they can be re-used which the MLME right
now will do for its combined probe/auth step.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: generalise management work a bit
Johannes Berg [Wed, 23 Dec 2009 12:15:34 +0000 (13:15 +0100)]
mac80211: generalise management work a bit

As a first step of generalising management work,
this renames a few things and puts more information
directly into the struct so that auth/assoc need
not access the BSS pointer as often -- in fact it
can be removed from auth completely. Also since the
previous patch made sure a new work item is used
for association, we can make the different data a
union.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: let cfg80211 manage auth state
Johannes Berg [Wed, 23 Dec 2009 12:15:33 +0000 (13:15 +0100)]
mac80211: let cfg80211 manage auth state

mac80211 currently hangs on to the auth state by
keeping it on the work list. That can lead to
confusing behaviour like rejecting scans while
authenticated to any AP (but not yet associated.)
It also means that it needs to keep track of the
work struct while associated for when it gets
disassociated (or disassociates.)

Change this to free the work struct after the
authentication completed successfully and
allocate a new one for associating, thereby
letting cfg80211 manage the auth state. Another
change necessary for this is to tell cfg80211
about all unicast deauth frames sent to mac80211
since now it can no longer check the auth state,
but that check was racy anyway.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: introduce flush operation
Johannes Berg [Wed, 23 Dec 2009 12:15:32 +0000 (13:15 +0100)]
mac80211: introduce flush operation

We've long lacked a good confirmation that frames
have really gone out, e.g. before going off-channel
for a scan. Add a flush() operation that drivers
can implement to provide that confirmation, and use
it in a few places:
 * before scanning sends the nullfunc frames
 * after scanning sends the nullfunc frames, if any
 * when going idle, to send any pending frames

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agomac80211: add ieee80211_sdata_running
Johannes Berg [Wed, 23 Dec 2009 12:15:31 +0000 (13:15 +0100)]
mac80211: add ieee80211_sdata_running

Instead of always using netif_running(sdata->dev)
use ieee80211_sdata_running(sdata) now which is
just an inline containing netif_running() for now.

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
14 years agortl8187: remove priv->mode
John W. Linville [Tue, 22 Dec 2009 23:13:05 +0000 (18:13 -0500)]
rtl8187: remove priv->mode

It is checked in add_interface, but there it is easily replaced with
a check of priv->vif.  If that information should become necessary,
it is available in vif->type anyway.

It is also checked in led_turn_on and led_turn_off, where I made the
substitutions as described above.  Of course, these checks seem to
have been incorrect since the driver was using NL80211_IFTYPE_MONITOR
to indicate no interface rather than NL80211_IFTYPE_UNSPECIFIED.
Anyway, I think these checks may be extraneous...?

Signed-off-by: John W. Linville <linville@tuxdriver.com>