linux-2.6-microblaze.git
13 years agocfg80211: enter psm when working as p2p_cli
Eliad Peller [Tue, 19 Jul 2011 09:57:13 +0000 (12:57 +0300)]
cfg80211: enter psm when working as p2p_cli

cfg80211_netdev_notifier_call() is configuring psm in case
of NL80211_IFTYPE_STATION interface type (on NETDEV_UP).
do the same for NL80211_IFTYPE_P2P_CLIENT interface type.

Signed-off-by: Eliad Peller <eliad@wizery.com>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: sync driver before TX
Johannes Berg [Tue, 19 Jul 2011 08:39:53 +0000 (10:39 +0200)]
mac80211: sync driver before TX

In P2P client mode, the GO (AP) to connect to might
have periods of time where it is not available due
to powersave. To allow the driver to sync with it
and send frames to the GO only when it is available
add a new callback tx_sync (and the corresponding
finish_tx_sync). These callbacks can sleep unlike
the actual TX.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: merge ath5k_hw and ath5k_softc
Pavel Roskin [Fri, 15 Jul 2011 00:21:55 +0000 (20:21 -0400)]
ath5k: merge ath5k_hw and ath5k_softc

Both ath5k_hw and ath5k_softc represent one instance of the hardware.
This duplication is historical and is not needed anymore.

Keep the name "ath5k_hw" for the merged structure and "ah" for the
variable pointing to it.  "ath5k_hw" is shorter than "ath5k_softc", more
descriptive and more widely used.

Put the combined structure to ath5k.h where the old ath5k_softc used to
be. Move some code from base.h to ath5k.h as needed.

Remove memory allocation for struct ath5k_hw and the corresponding error
handling.  Merge iobase and ah_iobase fields.

Signed-off-by: Pavel Roskin <proski@gnu.org>
Acked-by: Nick Kossifidis <mickflemm@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: bcma: implement full core reset
Rafał Miłecki [Sat, 16 Jul 2011 23:06:06 +0000 (01:06 +0200)]
b43: bcma: implement full core reset

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: allow enabling PLL
Rafał Miłecki [Sat, 16 Jul 2011 23:06:05 +0000 (01:06 +0200)]
bcma: allow enabling PLL

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: allow setting FAST clockmode for a core
Rafał Miłecki [Sat, 16 Jul 2011 23:06:04 +0000 (01:06 +0200)]
bcma: allow setting FAST clockmode for a core

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: trivial: add helpers for masking/setting
Rafał Miłecki [Sat, 16 Jul 2011 23:06:03 +0000 (01:06 +0200)]
bcma: trivial: add helpers for masking/setting

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: move define of BCMA_CLKCTLST register
Rafał Miłecki [Sat, 16 Jul 2011 16:43:36 +0000 (18:43 +0200)]
bcma: move define of BCMA_CLKCTLST register

Recent experiments have shown many cores share 0x1E0 register used for
clock management.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: correctly display longer chipsets ids
Rafał Miłecki [Mon, 18 Jul 2011 20:45:58 +0000 (22:45 +0200)]
b43: correctly display longer chipsets ids

Some of the newer Broadcom chipsets have longe names like BCM43224,
BCM43225, etc. However Broadcom decided to keep using u16 for storing
them. Use %X or %d depending on chip_id value to avoid BCMA8D8, etc.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Acked-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: implement scan supported rates
Johannes Berg [Mon, 18 Jul 2011 16:08:36 +0000 (18:08 +0200)]
mac80211: implement scan supported rates

Scanning currently uses the TX rate mask to
restrict the rate set, which is bogus. Make
it use the new set of rates from userspace.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: allow userspace to control supported rates in scan
Johannes Berg [Mon, 18 Jul 2011 16:08:35 +0000 (18:08 +0200)]
cfg80211: allow userspace to control supported rates in scan

Some P2P scans are not allowed to advertise
11b rates, but that is a rather special case
so instead of having that, allow userspace
to request the rate sets (per band) that are
advertised in scan probe request frames.

Since it's needed in two places now, factor
out some common code parsing a rate array.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: HT-PHY: fix masks in radio ctl
Rafał Miłecki [Mon, 18 Jul 2011 00:13:23 +0000 (02:13 +0200)]
b43: HT-PHY: fix masks in radio ctl

Old masks were causing ugly, delayed lock ups.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: bcma: read info about supported bands
Rafał Miłecki [Mon, 18 Jul 2011 00:01:30 +0000 (02:01 +0200)]
b43: bcma: read info about supported bands

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: bcma: define 80211 core specific IO status bits
Rafał Miłecki [Mon, 18 Jul 2011 00:01:29 +0000 (02:01 +0200)]
b43: bcma: define 80211 core specific IO status bits

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: define IO status register
Rafał Miłecki [Mon, 18 Jul 2011 00:01:28 +0000 (02:01 +0200)]
bcma: define IO status register

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoieee80211: add few wmm tspec values
Kalle Valo [Sun, 17 Jul 2011 09:13:56 +0000 (12:13 +0300)]
ieee80211: add few wmm tspec values

These are needed by ath6kl for parsing tspec status from an IE.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: handle alternative SPROM location
Rafał Miłecki [Sun, 17 Jul 2011 09:00:59 +0000 (11:00 +0200)]
bcma: handle alternative SPROM location

Some cards do not use additional 0x30 offset for SPROM location. We do
not know the real condition for it yet, make it BCM4331 specific for
now.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: HT-PHY: fix typo in 0x2059 radio init
Rafał Miłecki [Sun, 17 Jul 2011 08:30:34 +0000 (10:30 +0200)]
b43: HT-PHY: fix typo in 0x2059 radio init

Following operation was incorrectly translated:
 radio_read(0x0011) -> 0xffff
radio_write(0x0011) <- 0xfff7

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: HT-PHY: find channel entry with regs data
Rafał Miłecki [Sun, 17 Jul 2011 08:30:33 +0000 (10:30 +0200)]
b43: HT-PHY: find channel entry with regs data

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: HT-PHY: switch to channel after enabling radio
Rafał Miłecki [Sun, 17 Jul 2011 08:30:32 +0000 (10:30 +0200)]
b43: HT-PHY: switch to channel after enabling radio

Also change the default channel to 11. This is the first channel closed
driver switches to.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: use agent R/W ops for BCMA_IOCTL
Rafał Miłecki [Sat, 16 Jul 2011 16:27:55 +0000 (18:27 +0200)]
b43: use agent R/W ops for BCMA_IOCTL

BCMA_IOCTL is register in agent (AKA wrapper) core, we need to use
special R/W ops for it.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Fix some smatch warnings
Rajkumar Manoharan [Sun, 17 Jul 2011 06:08:49 +0000 (11:38 +0530)]
ath9k: Fix some smatch warnings

drivers/net/wireless/ath/ath9k/hif_usb.c +135 hif_usb_mgmt_cb(6) warn:
variable dereferenced before check 'cmd'
drivers/net/wireless/ath/ath9k/btcoex.c +77 ath9k_hw_init_btcoex_hw(38)
error: buffer overflow 'ah->hw_gen_timers.gen_timer_index' 32 <=
2009813776

Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoMAINTAINERS: change maintainer of the wl1251 driver
Luciano Coelho [Sun, 17 Jul 2011 07:06:34 +0000 (10:06 +0300)]
MAINTAINERS: change maintainer of the wl1251 driver

Kalle Valo has asked me to take over the maintenance of the wl1251
driver.  Thanks Kalle for all his work on the wl1251 driver, I'll try
to keep up with his good work from now on.

At the same time, update the Web URL to something more specific.
There's not much in that page, but we will try to update it soon.

Signed-off-by: Luciano Coelho <coelho@ti.com>
Acked-by: Kalle Valo <kvalo@adurom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Fix sparse warnings
Rajkumar Manoharan [Sun, 17 Jul 2011 06:13:02 +0000 (11:43 +0530)]
ath9k: Fix sparse warnings

drivers/net/wireless/ath/ath9k/init.c:199:21: warning: context imbalance
in 'ath9k_reg_rmw' - different lock contexts for basic block
drivers/net/wireless/ath/ath9k/xmit.c:1175:31: warning: context
imbalance in 'ath_drain_txq_list' - unexpected unlock
drivers/net/wireless/ath/ath9k/xmit.c:2047:23: warning: context
imbalance in 'ath_tx_process_buffer' - unexpected unlock
drivers/net/wireless/ath/ath9k/ar9003_eeprom.c:3041:24: warning: cast to
restricted __le32

Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: bus: drop inline from SSB functions
Rafał Miłecki [Sat, 16 Jul 2011 23:15:18 +0000 (01:15 +0200)]
b43: bus: drop inline from SSB functions

We use all that functions by pointers only. This forces compiler to
create additional duplicated functions that are not inline.
Noticed by Michael in similar bcma code.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocarl9170: set beacon xmit power to the max
Christian Lamparter [Sat, 16 Jul 2011 15:30:19 +0000 (17:30 +0200)]
carl9170: set beacon xmit power to the max

Harshal Chhaya discovered during network tests, that
several of his clients dropped-off the network. The
captured packets shows that the beacons sent by the
AP are at a much lower power than the other data
packets.

The reason for this mishap: The driver never updated
the beacon phy register, so all beacons were always
sent at the lowest power.

Reference: http://marc.info/?l=linux-wireless&m=131067225105801
Reported-by: Harshal Chhaya <harshal@gmail.com>
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocarl9170: move beacon_update into tx.c
Christian Lamparter [Sat, 16 Jul 2011 15:25:14 +0000 (17:25 +0200)]
carl9170: move beacon_update into tx.c

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocarl9170 firmware: update firmware headers
Christian Lamparter [Sat, 16 Jul 2011 15:21:01 +0000 (17:21 +0200)]
carl9170 firmware: update firmware headers

 * reserves feature bit for CCA counters

 * extends hardware register file definitions

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: use ath_opmode_to_string()
Pavel Roskin [Fri, 15 Jul 2011 23:06:46 +0000 (19:06 -0400)]
ath9k: use ath_opmode_to_string()

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: remove defines in reg.h that exist in ../reg.h
Pavel Roskin [Fri, 15 Jul 2011 23:06:40 +0000 (19:06 -0400)]
ath9k: remove defines in reg.h that exist in ../reg.h

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: use get_unaligned_{b16, le16, le32} where possible
Pavel Roskin [Fri, 15 Jul 2011 23:06:33 +0000 (19:06 -0400)]
ath9k: use get_unaligned_{b16, le16, le32} where possible

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath: use get_unaligned_le{16,32} in ath_hw_keysetmac()
Pavel Roskin [Fri, 15 Jul 2011 23:06:27 +0000 (19:06 -0400)]
ath: use get_unaligned_le{16,32} in ath_hw_keysetmac()

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: extract SPROM rev 9 the same way as rev 8
Rafał Miłecki [Thu, 14 Jul 2011 19:49:21 +0000 (21:49 +0200)]
bcma: extract SPROM rev 9 the same way as rev 8

SPROM rev 9 was discovered on 14e4:4331, it seems to have very similar
layout to rev 8 one. Use the same extracting function until we find some
differences.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: cc: set GPIOTIMER register
Rafał Miłecki [Thu, 14 Jul 2011 19:49:19 +0000 (21:49 +0200)]
bcma: cc: set GPIOTIMER register

We use value supplied via SPROM, or default 10:90 as fallback.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agossb: SPROM: add LED duty cycle fields
Rafał Miłecki [Thu, 14 Jul 2011 19:49:18 +0000 (21:49 +0200)]
ssb: SPROM: add LED duty cycle fields

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: reconfigure tx on device reconfiguration
Eliad Peller [Thu, 14 Jul 2011 17:29:42 +0000 (20:29 +0300)]
mac80211: reconfigure tx on device reconfiguration

Add tx_conf array to save the current tx queues
configuration, and reconfig it on resume (ieee80211_reconfig).

On resume, the driver is being reconfigured. Without
reconfiguring the tx queues as well, the driver might
configure the device to use wrong ac params (e.g. ps-poll
instead of uapsd).

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlagn: add tx start API to transport layer
Emmanuel Grumbach [Thu, 7 Jul 2011 12:50:10 +0000 (15:50 +0300)]
iwlagn: add tx start API to transport layer

tx start will start the tx queues: basically configure the SCD
Remove the IWLAGN prefix to SCD defines on the way.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.guy@intel.com>
13 years agoiwlagn: move all the ICT related functions to iwl-trans-rx-pcie.c
Emmanuel Grumbach [Mon, 11 Jul 2011 14:44:57 +0000 (07:44 -0700)]
iwlagn: move all the ICT related functions to iwl-trans-rx-pcie.c

Since the ICT is transport related, move all its functions to the transport
layer.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: move iwlagn_stop_device to transport layer
Emmanuel Grumbach [Thu, 7 Jul 2011 11:37:26 +0000 (14:37 +0300)]
iwlagn: move iwlagn_stop_device to transport layer

Since iwlagn_stop_device was the only caller to the rx_stop / tx_stop,
these two don't need to be API any more.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.guy@intel.com>
13 years agoiwlagn: move tx transport functions to iwl-trans-tx-pcie.c
Emmanuel Grumbach [Mon, 11 Jul 2011 14:39:46 +0000 (07:39 -0700)]
iwlagn: move tx transport functions to iwl-trans-tx-pcie.c

There are still a few functions here and there that should be
put in the transport layer. Mainly the functions that are related to the reclaim flow.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: move rx transport functions to iwl-trans-rx-pcie.c
Emmanuel Grumbach [Mon, 11 Jul 2011 14:35:34 +0000 (07:35 -0700)]
iwlagn: move rx transport functions to iwl-trans-rx-pcie.c

Also create a new file: iwl-trans-int-pcie.h which will include
the non static functions that are shared among the current pcie transport layer.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: remove indirection for iwlagn_hw_valid_rtc_data_addr
Fry, Donald H [Thu, 7 Jul 2011 22:24:07 +0000 (15:24 -0700)]
iwlagn: remove indirection for iwlagn_hw_valid_rtc_data_addr

Not needed since the driver split.

Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: calibration bitmap
Wey-Yi Guy [Thu, 7 Jul 2011 19:24:46 +0000 (12:24 -0700)]
iwlagn: calibration bitmap

Define bitmap for calibration

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: set default of uCode ownership to driver
Hsu, Kenny [Thu, 7 Jul 2011 10:10:27 +0000 (03:10 -0700)]
iwlagn: set default of uCode ownership to driver

The driver should take the ownership of the uCode as default setting
for later operations after interface up.

Signed-off-by: Kenny Hsu <kenny.hsu@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: simplify TX flags assignments
Johannes Berg [Thu, 7 Jul 2011 12:11:51 +0000 (05:11 -0700)]
iwlagn: simplify TX flags assignments

The first assignment of TX_CMD_FLG_SEQ_CTL_MSK for
ack-expected mgmt frames is overwritten later in
the function, so it's useless. Also, probe response
frames, BACK request and others there are mutually
exclusive so can be moved into an else branch.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: add comment to tx and get_tx_cmd in iwl_trans_ops
Emmanuel Grumbach [Thu, 7 Jul 2011 16:25:51 +0000 (09:25 -0700)]
iwlagn: add comment to tx and get_tx_cmd in iwl_trans_ops

Those comments were missed in a previous commit.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: comments for iwl_cfg
Wey-Yi Guy [Wed, 6 Jul 2011 23:28:48 +0000 (16:28 -0700)]
iwlagn: comments for iwl_cfg

Modify the comments for iwl_cfg, no functional changes

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: another double indirect removed
Wey-Yi Guy [Wed, 6 Jul 2011 23:28:47 +0000 (16:28 -0700)]
iwlagn: another double indirect removed

Another clean up work after driver split

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: remove dual-indirect call to simply the code
Wey-Yi Guy [Thu, 7 Jul 2011 15:33:01 +0000 (08:33 -0700)]
iwlagn: remove dual-indirect call to simply the code

After driver split, no need to make the code so complex

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: remove un-necessary file
Wey-Yi Guy [Thu, 7 Jul 2011 15:27:41 +0000 (08:27 -0700)]
iwlagn: remove un-necessary file

Most of the functions in iwl-agn-hcmd are move to other files, no point to
keep the file anymore.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: move the Rx dispatching to the upper layer
Emmanuel Grumbach [Thu, 7 Jul 2011 14:59:02 +0000 (07:59 -0700)]
iwlagn: move the Rx dispatching to the upper layer

The upper layer receives a pointer to an iwl_rx_mem_buffer. I would prefer the
upper layer to receive a pointer to an iwl_rx_packet, but this is impossible
since the Rx path needs to add the address of the page to the skb.
I may find a solution later.

All the pre_rx_handler and notification code has been moved to the upper layer.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: move sync_irq to transport layer
Emmanuel Grumbach [Mon, 4 Jul 2011 06:06:44 +0000 (09:06 +0300)]
iwlagn: move sync_irq to transport layer

Since all the irq / tasklet is now handled in the transport layer, it should
give an API to ensure that all the irq / tasklet have finished running. This
will allow the upper layer to release all its resources.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: move the tasklet / irq to the transport layer
Emmanuel Grumbach [Mon, 4 Jul 2011 05:58:19 +0000 (08:58 +0300)]
iwlagn: move the tasklet / irq to the transport layer

PCIe doesn't provide any ISR registration API, whereas other buses do.
Hence, we need to move the tasklet and irq to the transport layer to allow this
flexibility.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: move Tx datapath to transport layer
Emmanuel Grumbach [Sun, 3 Jul 2011 08:22:15 +0000 (11:22 +0300)]
iwlagn: move Tx datapath to transport layer

Split the Tx datapath in two parts:
* the first deals with the Tx cmd composition
* the second attaches the skb + Tx cmd to the queues

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agomac80211: be more careful in suspend/resume
Johannes Berg [Thu, 14 Jul 2011 14:48:54 +0000 (16:48 +0200)]
mac80211: be more careful in suspend/resume

When suspending with all netdevs down, the device
is stopped but we still call a number of driver
callbacks that the driver might not expect. The
same happens during resume, we might call a few
callbacks without starting the driver. Fix this
by checking open_count around more things and
exiting quickly if it is 0.

Also, while at this I noticed that the coverage
class isn't reprogrammed after resume, so add
that.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocarl9170: fix formatting issues found by checkpatch
Pavel Roskin [Thu, 14 Jul 2011 01:38:18 +0000 (21:38 -0400)]
carl9170: fix formatting issues found by checkpatch

Signed-off-by: Pavel Roskin <proski@gnu.org>
Acked-By: christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoMAINTAINERS: add entry for Marvell mwifiex wireless driver
Bing Zhao [Wed, 13 Jul 2011 20:11:58 +0000 (13:11 -0700)]
MAINTAINERS: add entry for Marvell mwifiex wireless driver

Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: let key iteration get keys in install order
Johannes Berg [Wed, 13 Jul 2011 17:50:53 +0000 (19:50 +0200)]
mac80211: let key iteration get keys in install order

ieee80211_iter_keys() currently returns keys in
the backward order they were installed in, which
is a bit confusing. Add them to the tail of the
key list to make sure iterations go in the same
order that keys were originally installed in.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: allow driver access to TKIP RX P1K
Johannes Berg [Wed, 13 Jul 2011 17:50:34 +0000 (19:50 +0200)]
mac80211: allow driver access to TKIP RX P1K

When the driver wants to pre-program the TKIP
RX phase 1 key, it needs to be able to obtain
it for the peer's TA. Add API to allow it to
generate it.

The generation uses a dummy on-stack context
since it doesn't know the RX queue.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: improve reliability of MIC error detection
Felix Fietkau [Wed, 13 Jul 2011 15:35:05 +0000 (23:35 +0800)]
ath9k: improve reliability of MIC error detection

For unicast the hardware sometimes reports MIC errors even though the
frame that it received actually contains a valid MIC - on some chips this
can happen frequently enough to trigger TKIP countermeasures.
Fix this issue by not reporting MIC errors for unicast frames with a
valid key, letting mac80211 validate the MIC instead.

Additionally, strip the MIC for all frames that the hardware considers
valid to avoid wasting CPU cycles re-validating it.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoorinoco: minor fixes for problems found by checkpatch.pl
Pavel Roskin [Wed, 13 Jul 2011 15:19:57 +0000 (11:19 -0400)]
orinoco: minor fixes for problems found by checkpatch.pl

Eliminate spaces before tabs.  Eliminate typedefs.  Add spaces around
operators.

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomwl8k: Fixing sta dereference when ieee80211_tx_info->control.sta is NULL
Yogesh Ashok Powar [Wed, 13 Jul 2011 12:12:06 +0000 (17:42 +0530)]
mwl8k: Fixing sta dereference when ieee80211_tx_info->control.sta is NULL

Following oops was seen on SMP machine

>BUG: unable to handle kernel NULL pointer dereference at 00000012
>IP: [<f8c56691>] mwl8k_tx+0x20e/0x561 [mwl8k]
>*pde = 00000000
>Oops: 0000 [#1] SMP
>Modules linked in: mwl8k mac80211 cfg80211 [last unloaded: cfg80211]

As ieee80211_tx_info->control.sta may be NULL during ->tx call, avoiding sta
dereference in such scenario with the following patch.

Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211/cfg80211: add max_sched_scan_ie_len in the hw description
Luciano Coelho [Wed, 13 Jul 2011 11:57:29 +0000 (14:57 +0300)]
nl80211/cfg80211: add max_sched_scan_ie_len in the hw description

Some chips may support different lengths of user-supplied IEs with a
single scheduled scan command than with a single normal scan command.

To support this, this patch creates a separate hardware description
element that describes the maximum size of user-supplied information
element data supported in scheduled scans.

Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211/cfg80211: add max_sched_scan_ssids in the hw description
Luciano Coelho [Wed, 13 Jul 2011 11:57:28 +0000 (14:57 +0300)]
nl80211/cfg80211: add max_sched_scan_ssids in the hw description

Some chips can scan more SSIDs with a single scheduled scan command
than with a single normal scan command (eg. wl12xx chips).

To support this, this patch creates a separate hardware description
element that describes the amount of SSIDs supported in scheduled
scans.

Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211: advertise GTK rekey support, new triggers
Johannes Berg [Wed, 13 Jul 2011 08:48:55 +0000 (10:48 +0200)]
nl80211: advertise GTK rekey support, new triggers

Since we now have the necessary API in place to support
GTK rekeying, applications will need to know whether it
is supported by a device. Add a pseudo-trigger that is
used only to advertise that capability. Also, add some
new triggers that match what iwlagn devices can do.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: allow driver to disconnect after resume
Johannes Berg [Tue, 12 Jul 2011 10:30:59 +0000 (12:30 +0200)]
mac80211: allow driver to disconnect after resume

In WoWLAN, devices may use crypto keys for TX/RX
and could also implement GTK rekeying. If the
driver isn't able to retrieve replay counters and
similar information from the device upon resume,
or if the device isn't responsive due to platform
issues, it isn't safe to keep the connection up
as GTK rekey messages from during the sleep time
could be replayed against it.

The only protection against that is disconnecting
from the AP. Modifying mac80211 to do that while
it is resuming would be very complex and invasive
in the case that the driver requires a reconfig,
so do it after it has resumed completely. In that
case, however, packets might be replayed since it
can then only happen after TX/RX are up again, so
mark keys for interfaces that need to disconnect
as "tainted" and drop all packets that are sent
or received with those keys.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: remove a redundant check
Mohammed Shafi Shajakhan [Tue, 12 Jul 2011 09:49:04 +0000 (15:19 +0530)]
mac80211: remove a redundant check

is_valid_ether_addr itself checks for is_zero_ether_addr

Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomwifiex: add cfg80211 handler for set_bitrate_mask
Yogesh Ashok Powar [Tue, 12 Jul 2011 03:04:44 +0000 (20:04 -0700)]
mwifiex: add cfg80211 handler for set_bitrate_mask

Currently, setting only legacy bitrates on 2.4GHz band
are supported. Mode 802.11b/g/bg is enabled based on
bitrates selection. If only CCK bitrates selected then
802.11b mode is enabled. If only OFDM bitrates are
selected then 802.11g mode is enabled. For both: CCK
and OFDM rates 802.11bg mixed mode is enabled.

Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: do not call ieee80211_stop_queue for queues not managed by mac80211
Felix Fietkau [Tue, 12 Jul 2011 01:02:05 +0000 (09:02 +0800)]
ath5k: do not call ieee80211_stop_queue for queues not managed by mac80211

Instead of using ieee80211_stop_queue, check the configured tx queue
limit before calling ieee80211_get_buffered_bc.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: disable 32KHz sleep clock operation by default
Felix Fietkau [Tue, 12 Jul 2011 01:02:04 +0000 (09:02 +0800)]
ath5k: disable 32KHz sleep clock operation by default

While 32 KHz sleep clock might provide some power saving benefits,
it is also a major source of stability issues, on OpenWrt it produced
some reproducible data bus errors on register accesses on several
different MIPS platforms.

All the Atheros drivers that I can find do not enable this feature,
so it makes sense to leave it disabled in ath5k as well.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Acked-by: Nick Kossifidis <mickflemm@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: fix reference clock usec duration setting restore
Felix Fietkau [Tue, 12 Jul 2011 01:02:03 +0000 (09:02 +0800)]
ath5k: fix reference clock usec duration setting restore

enabling the sleep clock alters the AR5K_USEC_32 field, but disabling
it didn't restore it.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Acked-by: Nick Kossifidis <mickflemm@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: delay full calibration after reset
Felix Fietkau [Tue, 12 Jul 2011 01:02:02 +0000 (09:02 +0800)]
ath5k: delay full calibration after reset

During scans the full calibration usually does not make much sense,
PAPD probing and IQ calibration should be deferred until there is
enough time to complete them. Adding 100 ms to the initial full
calibration delay should be enough to do this.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Acked-by: Nick Kossifidis <mickflemm@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: initialize common->clockrate
Felix Fietkau [Tue, 12 Jul 2011 01:02:01 +0000 (09:02 +0800)]
ath5k: initialize common->clockrate

Fixes a division by zero when setting distance before activating the
device for the first time.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Acked-by: Nick Kossifidis <mickflemm@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: fix reference clock frequency for spur mitigation on AR2413
Felix Fietkau [Tue, 12 Jul 2011 01:02:00 +0000 (09:02 +0800)]
ath5k: fix reference clock frequency for spur mitigation on AR2413

AR2413 uses the same reference clock as AR5413

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: apply the synth voltage tweak only on AR5112 rev 2
Felix Fietkau [Tue, 12 Jul 2011 01:01:59 +0000 (09:01 +0800)]
ath5k: apply the synth voltage tweak only on AR5112 rev 2

Might fix some stability issues on newer chips

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Acked-by: Nick Kossifidis <mickflemm@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: add missing checks for rfgain probe
Felix Fietkau [Tue, 12 Jul 2011 01:01:58 +0000 (09:01 +0800)]
ath5k: add missing checks for rfgain probe

rfgain probe is only necessary for OFDM operation on AR5111 and AR5112.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: Fix incorrect key_miss handling
Senthil Balasubramanian [Mon, 11 Jul 2011 18:32:56 +0000 (00:02 +0530)]
ath9k_hw: Fix incorrect key_miss handling

Decryping frames on key_miss handling shouldn't be done for Michael
MIC failed frames as h/w would have already decrypted such frames
successfully anyway.

Also leaving CRC and PHY error(where the frame is going to be dropped
anyway), we are left to prcoess Decrypt error for which s/w decrypt is
selected anway and so having key_miss as a separate check doesn't serve
anything. So making key_miss handling mutually exlusive with other RX
status handling makes much more sense.

This patch addresses an issue with STA not reporting MIC failure events
resulting in STA being disconnected immediately.

Cc: stable@kernel.org
Signed-off-by: Senthil Balasubramanian <senthilb@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlagn: Add missing comma between constant string array
Joe Perches [Sat, 9 Jul 2011 06:20:25 +0000 (23:20 -0700)]
iwlagn: Add missing comma between constant string array

Multiple quoted strings are concatenated without comma separators.

Make the array const while there.

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlegacy: Add missing comma between constant string array
Joe Perches [Sat, 9 Jul 2011 06:20:24 +0000 (23:20 -0700)]
iwlegacy: Add missing comma between constant string array

Multiple quoted strings are concatenated without comma separators.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlagn: fix warning in testmode attribute table
Wey-Yi Guy [Mon, 11 Jul 2011 17:47:39 +0000 (10:47 -0700)]
iwlagn: fix warning in testmode attribute table

Fix the compile warning cause by [IWL_TM_ATTR_MAX - 1]

Reported-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: remove an unnecessary function declaration
Mohammed Shafi Shajakhan [Mon, 11 Jul 2011 14:44:14 +0000 (20:14 +0530)]
ath9k: remove an unnecessary function declaration

Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2800usb: Add new device ID for Belkin
Eduardo Bacchi Kienetz [Sun, 10 Jul 2011 15:05:14 +0000 (17:05 +0200)]
rt2800usb: Add new device ID for Belkin

Belkin's Connect N150 Wireless USB Adapter, model F7D1101 version 2, uses ID 0x945b.
Chipset info: rt: 3390, rf: 000b, rev: 3213.
I have just bought one, which started to work perfectly after the ID was added through this patch.

Signed-off-by: Eduardo Bacchi Kienetz <eduardo@kienetz.com>
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: Add missing breaks in switch/case
Joe Perches [Sun, 10 Jul 2011 09:28:26 +0000 (02:28 -0700)]
ath5k: Add missing breaks in switch/case

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192cu: Fix duplicate if test
Larry Finger [Sat, 9 Jul 2011 18:15:58 +0000 (13:15 -0500)]
rtlwifi: rtl8192cu: Fix duplicate if test

A typo causes routine rtl92cu_phy_rf6052_set_cck_txpower() to test the
same condition twice. The problem was found using cppcheck-1.49, and the
proper fix was verified against the pre-mac80211 version of the code.

Reported-by: David Binderman <dcb314@hotmail.com>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: stable <stable@kernel.org> [back to 2.6.39]
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agolibertas: fix handling of command timeout, completion and interruption
Daniel Drake [Sat, 9 Jul 2011 14:41:25 +0000 (15:41 +0100)]
libertas: fix handling of command timeout, completion and interruption

When commands time out, corruption ensues. As lbs_complete_command()
is called without locking, the command node is mistakenly freed twice.
Also fixed up locking here in a few other places.

The nature of command timeout may be that the card didn't even
acknowledge receipt of the request. Detect this case and reset dnld_sent
so that other commands don't hang forever.

When cmdnodes are moved between the free list and the pending list,
their list heads should be reinitialized. Fixed this.

Sometimes commands are completed without actually submitting them or
removing them from cmdpendingq. We must remember to remove them from
cmdpendingq in these cases, so handle this in lbs_complete_command().

Harmless signals generated during suspend/resume were interrupting
lbs_cmd. Convert to an uninterruptible sleep to avoid this.

lbs_thread must be woken up every time there is some new work to do.
I found that when 2 commands are queued, ther completion of the first
command would not wake up lbs_thread to submit the second. Poke lbs_thread
at the end of lbs_complete_command() to fix this.

Signed-off-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: allocate only one RX queue
Johannes Berg [Sat, 9 Jul 2011 13:39:16 +0000 (15:39 +0200)]
mac80211: allocate only one RX queue

We don't have multiple RX queues, so there's no use
in allocating multiple, use alloc_netdev_mqs() to
allocate multiple TX but only one RX queue.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: fix docbook
Johannes Berg [Sat, 9 Jul 2011 12:48:31 +0000 (14:48 +0200)]
mac80211: fix docbook

I changed the TKIP key functions, but forgot to
update the documentation includes, fix that.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: fix docbook
Johannes Berg [Sat, 9 Jul 2011 12:48:30 +0000 (14:48 +0200)]
cfg80211: fix docbook

Looks like I forgot to document the "gfp" parameter
to cfg80211_gtk_rekey_notify, add it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agobcma: add check if sprom is available before accessing it.
Hauke Mehrtens [Sat, 9 Jul 2011 11:22:03 +0000 (13:22 +0200)]
bcma: add check if sprom is available before accessing it.

The SoCs like the bcm4716 do not have a sprom on the bcma bus like a
pcie device. It stores the values in some partition on flash memory.
For ssb this informations are read out in the bcm47xx arch code,
something like that should also be implemented for bcma. Without this
patch bcma panics on SoCs.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agolibertas: fix errors in functions accessing module registers
Olivier Sobrie [Sat, 9 Jul 2011 10:57:05 +0000 (12:57 +0200)]
libertas: fix errors in functions accessing module registers

Problems located in the two functions lbs_set_reg() and lbs_get_reg():
- The offset field of struct cmd_ds_reg_access was not filled in
- The test on the return code of lbs_cmd_with_response() in function
  lbs_get_reg() was inverted

Signed-off-by: Olivier Sobrie <olivier@sobrie.be>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: fix typos, bad comment formatting and GHz in place of MHz
Pavel Roskin [Sat, 9 Jul 2011 04:17:51 +0000 (00:17 -0400)]
ath5k: fix typos, bad comment formatting and GHz in place of MHz

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath5k: validate mode in ath5k_ani_init() before trying to set it
Pavel Roskin [Sat, 9 Jul 2011 04:17:45 +0000 (00:17 -0400)]
ath5k: validate mode in ath5k_ani_init() before trying to set it

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: add a missing delay for an analog register
Felix Fietkau [Sat, 9 Jul 2011 04:12:54 +0000 (11:12 +0700)]
ath9k_hw: add a missing delay for an analog register

AR_AN_SYNTH9 is in the analog shift register range and thus needs to be
written using the ath9k_hw_analog_shift_rmw function.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: fix synth delay for half/quarter channels
Felix Fietkau [Sat, 9 Jul 2011 04:12:53 +0000 (11:12 +0700)]
ath9k_hw: fix synth delay for half/quarter channels

The radio needs twice / four times as much time to stabilize for half/quarter
channels.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: set 5 GHz half/quarter channels on AR9002 using fractional mode
Felix Fietkau [Sat, 9 Jul 2011 04:12:52 +0000 (11:12 +0700)]
ath9k_hw: set 5 GHz half/quarter channels on AR9002 using fractional mode

The radio seems to perform better that way

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: remove ar9287 v1.3+ specific hardcoded register hacks
Felix Fietkau [Sat, 9 Jul 2011 04:12:51 +0000 (11:12 +0700)]
ath9k_hw: remove ar9287 v1.3+ specific hardcoded register hacks

Now that the clock rate is initialized properly and SIFS, EIFS, USEC,
slot time and ACK timeout are properly calculated by the generic code,
the 'async FIFO' register hacks are no longer necessary.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: initialize more timing related registers for half/quarter channels
Felix Fietkau [Sat, 9 Jul 2011 04:12:50 +0000 (11:12 +0700)]
ath9k_hw: initialize more timing related registers for half/quarter channels

Initialize the the clock-to-TSF field of AR_USEC and the SIFS and EIFS time
registers based on the clock rate instead of relying on initvals.
With those changes, some of the hardcoded AR9287 1.3+ specific overrides
can be dropped.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: fix MAC clock rate for AR9287 v1.3+
Felix Fietkau [Sat, 9 Jul 2011 04:12:49 +0000 (11:12 +0700)]
ath9k_hw: fix MAC clock rate for AR9287 v1.3+

This chip uses the async FIFO feature and runs the MAC at 117 MHz

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: fix MAC clock for half/quarter rate
Felix Fietkau [Sat, 9 Jul 2011 04:12:48 +0000 (11:12 +0700)]
ath9k_hw: fix MAC clock for half/quarter rate

Enabling half/quarter rate makes the MAC run at half/quarter clock speed.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: always enable 5 GHz fast clock for AR9280
Felix Fietkau [Sat, 9 Jul 2011 04:12:47 +0000 (11:12 +0700)]
ath9k_hw: always enable 5 GHz fast clock for AR9280

All devices support this, but some disable it using an EEPROM flag
for some reason. Improves 5 GHz performance on those devices.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>