Christian Lamparter [Thu, 25 Aug 2011 21:47:35 +0000 (23:47 +0200)]
p54: Use do_div for 64-bit division to fix 32-bit kernels
Use the do_div macro for 64-bit division. Otherwise, the module will
reference __udivdi3 under 32-bit kernels, which is not allowed in
kernel space.
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Florian Fainelli [Thu, 25 Aug 2011 19:33:48 +0000 (21:33 +0200)]
ath9k: add missing AR9340 in ath_mac_bb_names
AR9340 is not listed in ath_mac_bb_names, which leads to such a message:
ieee80211 phy0: Atheros AR???? Rev:0 mem=0xb8100000, irq=2
Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Stanislaw Gruszka [Thu, 25 Aug 2011 15:07:24 +0000 (17:07 +0200)]
cfg80211: document wiphy->registered
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
John W. Linville [Wed, 24 Aug 2011 19:12:41 +0000 (15:12 -0400)]
mac80211: refactor skb copy to failq in mesh_path_move_to_queue
This seems a bit less awkward...
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Peter Huewe [Tue, 7 Jun 2011 20:36:14 +0000 (22:36 +0200)]
net/mac80211/debugfs: Convert to kstrou8_from_user
This patch replaces the code for getting an number from a
userspace buffer by a simple call to kstrou8_from_user.
This makes it easier to read and less error prone.
Since the old buffer was only 10 bytes long and the value is masked by a
nibble-mask anyway, we don't need to use kstrtoul but rather kstrtou8.
Kernel Version: v3.0-rc2
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
John W. Linville [Mon, 29 Aug 2011 18:27:52 +0000 (14:27 -0400)]
Merge branch 'for-linville' of git://git./linux/kernel/git/luca/wl12xx
Luis R. Rodriguez [Tue, 23 Aug 2011 20:37:07 +0000 (13:37 -0700)]
ath9k: add AR9580 support
This has been tested in STA and AP mode by Florian.
Cc: David Quan <dquan@qca.qualcomm.com>
Cc: Kathy Giori <kgiori@qca.qualcomm.com>
Cc: Senthil Balasubramanian <senthilb@qca.qualcomm.com>
Tested-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Christian Lamparter [Thu, 25 Aug 2011 13:00:54 +0000 (15:00 +0200)]
carl9170: Use do_div for 64-bit division to fix 32-bit kernels
Use the do_div macro for 64-bit division. Otherwise, the module will
reference __udivdi3 under 32-bit kernels, which is not allowed in
kernel space.
drivers/built-in.o: In function `carl9170_collect_tally':
cmd.c:191: undefined reference to `__udivdi3'
cmd.c:192: undefined reference to `__udivdi3'
cmd.c:193: undefined reference to `__udivdi3'
Reported-by: Kalle Valo <kvalo@adurom.com>
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Tested-by: Kalle Valo <kvalo@adurom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Luis R. Rodriguez [Wed, 24 Aug 2011 22:36:08 +0000 (15:36 -0700)]
ath9k_hw: add AR9580 support
Here are the AR9580 1.0 initvals checksums using the
Atheros initvals-tools [1]. This is useful for when
we udate the initvals again with other values. It ensures
that we match the same initvals used internally. The
tool is documented on the wiki [2].
$ ./initvals -f ar9580-1p0
0x00000000e912711f ar9580_1p0_modes_fast_clock
0x000000004a488fc7 ar9580_1p0_radio_postamble
0x00000000f3888b02 ar9580_1p0_baseband_core
0x0000000003f783bb ar9580_1p0_mac_postamble
0x0000000094be244a ar9580_1p0_low_ob_db_tx_gain_table
0x0000000094be244a ar9580_1p0_high_power_tx_gain_table
0x0000000090be244a ar9580_1p0_lowest_ob_db_tx_gain_table
0x00000000ed9eaac6 ar9580_1p0_baseband_core_txfir_coeff_japan_2484
0x00000000c4d66d1b ar9580_1p0_mac_core
0x00000000e8e9043a ar9580_1p0_mixed_ob_db_tx_gain_table
0x000000003521a300 ar9580_1p0_wo_xlna_rx_gain_table
0x00000000301fc841 ar9580_1p0_soc_postamble
0x00000000a9a06b3a ar9580_1p0_high_ob_db_tx_gain_table
0x00000000a15ccf1b ar9580_1p0_soc_preamble
0x0000000029495000 ar9580_1p0_rx_gain_table
0x0000000037ac0ee8 ar9580_1p0_radio_core
0x00000000603a1b80 ar9580_1p0_baseband_postamble
0x000000003d8b4396 ar9580_1p0_pcie_phy_clkreq_enable_L1
0x00000000398b4396 ar9580_1p0_pcie_phy_clkreq_disable_L1
0x00000000397b4396 ar9580_1p0_pcie_phy_pll_on_clkreq
[1] git://git.kernel.org/pub/scm/linux/kernel/git/mcgrof/initvals-tool.git
[2] http://wireless.kernel.org/en/users/Drivers/ath9k_hw/initvals-tool
Cc: David Quan <dquan@qca.qualcomm.com>
Cc: Kathy Giori <kgiori@qca.qualcomm.com>
Cc: Senthil Balasubramanian <senthilb@qca.qualcomm.com>
Tested-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Mohammed Shafi Shajakhan [Wed, 24 Aug 2011 19:31:22 +0000 (01:01 +0530)]
ath9k_htc: Add get_stats call back
currently this call back is used only in debugfs of mac80211
Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
John W. Linville [Wed, 24 Aug 2011 19:05:14 +0000 (15:05 -0400)]
b43: correct warning for uninitialized variable 'macstat'
CC [M] drivers/net/wireless/b43/pio.o
drivers/net/wireless/b43/pio.c: In function ‘pio_rx_frame’:
drivers/net/wireless/b43/pio.c:614:6: warning: ‘macstat’ may be used uninitialized in this function
Signed-off-by: John W. Linville <linville@tuxdriver.com>
John W. Linville [Wed, 24 Aug 2011 18:08:41 +0000 (14:08 -0400)]
ath9k: remove replicated null check in ath_pci_aspm_init
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Wed, 24 Aug 2011 09:52:35 +0000 (11:52 +0200)]
b43: HT-PHY: read clip state
We don't know yet when to restore it, implement just reading. We found
out what for are that PHY ops by comparing HT with N code.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Wed, 24 Aug 2011 09:52:34 +0000 (11:52 +0200)]
b43: HT-PHY: use separated function for forcing RF sequence
Comparison of the HT and N code has shown similarities in the ops
performed after b43_mac_phy_clock_set. That way we understood what is
happening in the HT-PHY code.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Luis R. Rodriguez [Tue, 23 Aug 2011 22:07:31 +0000 (15:07 -0700)]
MAINTANERS: update Qualcomm Atheros addresses
Qualcomm ate up Atheros, all of the old e-mail addresses
no longer work and e-mails sent to it will bounce. Update
the addresses to the new shiny Qualcomm Atheros (QCA) ones.
Cc: stable@kernel.org
Cc: netdev@vger.kernel.org
Cc: jouni@qca.qualcomm.com
Cc: yangjie@qca.qualcomm.com
Cc: vthiagar@qca.qualcomm.com
Cc: senthilb@qca.qualcomm.com
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Luis R. Rodriguez [Tue, 23 Aug 2011 22:03:34 +0000 (15:03 -0700)]
wireless: relicense regulatory header to ISC
I will suck out stuff to userspace to start the regulatory
revampamp, this work will be permissively licensed.
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Eliad Peller [Tue, 23 Aug 2011 11:37:47 +0000 (14:37 +0300)]
mac80211: add uapsd_queues and max_sp params fields
Add uapsd_queues and max_sp fields to ieee80211_sta.
These fields might be needed by low-level drivers in
order to configure the AP.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Eliad Peller [Tue, 23 Aug 2011 11:37:46 +0000 (14:37 +0300)]
nl80211/cfg80211: add STA WME parameters
Add new NL80211_ATTR_STA_WME nested attribute that contains
wme params needed by the low-level driver (uapsd_queues and
max_sp).
Add these params to the station_parameters struct as well.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Arik Nemtsov [Tue, 23 Aug 2011 07:21:27 +0000 (10:21 +0300)]
mac80211: allow action frames with unknown BSSID in GO mode
When operating as a P2P GO, we receive some P2P action frames where the
BSSID is set to the peer MAC address. Specifically, this occurs for
invitation responses. These are valid action frames and they should be
passed up.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Tue, 23 Aug 2011 07:02:57 +0000 (12:32 +0530)]
ath9k: Send legacy rated frames as unaggregated
Currently the aggregation is formed till the aggregation limit
is reached and the rate lookup is done for the first frame alone.
But there can be a legacy rated frames in tid queue. This patch
limits the subframe addition based on presence of legacy rate and
sends the legacy rated frames as unaggregated one.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Guy Eilam [Wed, 17 Aug 2011 12:18:15 +0000 (15:18 +0300)]
mac80211: fix race condition between assoc_done and first EAP packet
When associating to an AP, the station might miss the first EAP
packet that the AP sends due to a race condition between the association
success procedure and the rx flow in mac80211.
In such cases, the packet might fall in ieee80211_rx_h_check due to
the fact that the relevant rx->sta wasn't allocated yet.
Allocation of the relevant station info struct before actually
sending the association request and setting it with a new
dummy_sta flag solve this problem.
The station will accept only EAP packets from the AP while it
is in the pre-association/dummy state.
This dummy station entry is not seen by normal sta_info_get()
calls, only by sta_info_get_bss_rx().
The driver is not notified for the first insertion of the
dummy station. The driver is notified only after the association
is complete and the dummy flag is removed from the station entry.
That way, all the rest of the code flow should be untouched by
this change.
Signed-off-by: Guy Eilam <guy@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Guy Eilam [Wed, 17 Aug 2011 12:18:14 +0000 (15:18 +0300)]
mac80211: refactor sta_info_insert_rcu to 3 main stages
Divided the sta_info_insert_rcu function to 3 mini-functions:
sta_info_insert_check - the initial checks done when inserting
a new station
sta_info_insert_ibss - the function that handles the station
addition for IBSS interfaces
sta_info_insert_non_ibss - the function that handles the station
addition in other cases
The outer API was not changed.
The refactoring was done for better usage of the different
stages in the station addition in new scenarios added
in the next commit.
Signed-off-by: Guy Eilam <guy@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Tue, 16 Aug 2011 10:14:07 +0000 (12:14 +0200)]
b43: drop Kconfig option of forcing PIO mode
We have module param called use_pio which is much easier to use.
Cc: Larry Finger <larry.finger@lwfinger.net>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Thomas Pedersen [Thu, 25 Aug 2011 17:36:14 +0000 (10:36 -0700)]
mac80211: mesh gate fixes
Since a v1 of the mesh gate series was accidentally applied, this patch
contains the changes in v2.
These are:
- automatically make mesh gate a root node.
- use TU_TO_EXP_TIME macro.
- initialize timer instead of checking for NULL timer function.
- cleanups.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Amitkumar Karwar [Tue, 23 Aug 2011 00:26:37 +0000 (17:26 -0700)]
libertas: update readme file
Since all wext specific code is removed, currently there is no
way to configure deep sleep mode. This patch removes deep sleep
configuration information in readme file.
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Larry Finger [Mon, 22 Aug 2011 21:50:18 +0000 (16:50 -0500)]
rtlwifi: rtl8192de: Convert to use the new rate-mapping routine in rtlwifi
This patch also deletes the now unused parts of rtl8192de/def.h.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Chaoming Li <chaoming_li@realsil.com.cn>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Larry Finger [Mon, 22 Aug 2011 21:50:17 +0000 (16:50 -0500)]
rtlwifi: rtl8192se: Convert to use the new rate-mapping routine in rtlwifi
This patch also deletes the now unused parts of rtl8192se/def.h.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Chaoming Li <chaoming_li@realsil.com.cn>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Larry Finger [Mon, 22 Aug 2011 21:50:16 +0000 (16:50 -0500)]
rtlwifi: rtl8192cu: Convert to use the new rate-mapping routine in rtlwifi
This patch also removes the now unused code from rtl8192ce/def.h.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Chaoming Li <chaoming_li@realsil.com.cn>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Larry Finger [Mon, 22 Aug 2011 21:50:15 +0000 (16:50 -0500)]
rtlwifi: rtl8192ce: Convert to use the new rate-mapping routine in rtlwifi
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Chaoming Li <chaoming_li@realsil.com.cn>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Larry Finger [Mon, 22 Aug 2011 21:50:14 +0000 (16:50 -0500)]
rtlwifi: Install updated rate-mapping routine
In preparation for fixing the rate-mapping situation, place a driver-agnostic
version in rtlwifi. This one contains the updated rate incormation.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Chaoming Li <chaoming_li@realsil.com.cn>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Julia Lawall [Mon, 22 Aug 2011 14:16:14 +0000 (16:16 +0200)]
drivers/net/wireless/mwifiex/scan.c: test the just-initialized value
Test the just-initialized value rather than some other one.
The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)
// <smpl>
@r@
identifier x,y,f!={PTR_ERR,ERR_PTR,ERR_CAST};
statement S;
@@
x = f(...);
(
if (\(x == NULL\|IS_ERR(x)\)) S
|
*if (\(y == NULL\|IS_ERR(y)\))
{ ... when != x
return ...; }
)
// </smpl>
Signed-off-by: Julia Lawall <julia@diku.dk>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Michael Büsch [Sun, 21 Aug 2011 15:24:47 +0000 (17:24 +0200)]
b43/legacy: Remove firmware IDs
This removes the "FWxx" ID strings from the b43 and b43legacy
drivers. They were once used to match a specific driver revision
to a set of firmware files. However, this is hardly useful today.
Additionally, the IDs are not updated and maintained properly, so
they might mislead users.
Signed-off-by: Michael Buesch <m@bues.ch>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 20 Aug 2011 12:04:19 +0000 (17:34 +0530)]
ath9k_hw: Disable Walsh spatial spreading for 2 chains
The Walsh bit is disabled for regulatory consideration.
FCC limit for walsh enable is lower than that for walsh disable. So
disabling walsh bit will not limit tx power/affect tx power even in
cases where we are not FCC limited (most client cards). If the tx
power is not FCC limited, then enabling/disabling walsh bit will
not affect Avg. EVM/overall performance in any visible manner.
Cc: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 20 Aug 2011 11:52:10 +0000 (17:22 +0530)]
ath9k: Change rate control to use legacy rate as last MRR
In congested network, having all rate reties at MCS rates
is failing to transmit the frame offenly. By the time reaching
the success rate set, the application gets timed out. One such
scenario is that authentication time out during 4-Way handshake.
This patch uses a legacy rate as last retry sequnce for
unaggregated frames or if the first selected rate's PER is ~80%
of max limit. And also observed from the tx status that the frame
was trasmitted successfully by using legacy rates.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 20 Aug 2011 11:52:09 +0000 (17:22 +0530)]
ath9k_hw: Fix descriptor status of TxOpExceeded
Cc: stable@kernel.org
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Mohammed Shafi Shajakhan [Sat, 20 Aug 2011 11:51:42 +0000 (17:21 +0530)]
ath9k: Add support for get_stats callback
this useful for debugging and to keep track of success/failure of
frames such as ACK, RTS and FCS error count in a noisy environment
Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Christian Lamparter [Fri, 19 Aug 2011 23:53:59 +0000 (01:53 +0200)]
p54: improve site survey
The firmware keeps track of channel usage. This data can
be used by the automatic channel selection to find the best
channel.
Survey data from wlan4
frequency: 5200 MHz [in use]
noise: -91 dBm
channel active time: 811909 ms
channel busy time: 63395 ms
channel transmit time: 59636 ms
Survey data from wlan4
frequency: 5210 MHz
noise: -91 dBm
channel active time: 121 ms
channel busy time: 119 ms
channel transmit time: 0 ms
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Aloisio Almeida Jr [Fri, 19 Aug 2011 19:09:49 +0000 (16:09 -0300)]
MAINTAINERS: Add NFC subsystem entry
Update MAINTAINERS with NFC subsystem and drivers entry.
Signed-off-by: Aloisio Almeida Jr <aloisio.almeida@openbossa.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Samuel Ortiz [Fri, 19 Aug 2011 13:47:11 +0000 (15:47 +0200)]
NFC: Reserve tx head and tail room
We can have the NFC core layer allocating the tx head and tail
room for the drivers and avoid 1 or more SKBs copy on write on
the Tx path.
Signed-off-by: Samuel Ortiz <sameo@linux.intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Dan Carpenter [Tue, 23 Aug 2011 19:15:35 +0000 (22:15 +0300)]
bcma: signedness bug in bcma_get_next_core()
The u32 would never be less than zero so the error handling would
break. I changed it to s32 to match how bcma_erom_get_mst_port() is
declared.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Bill Jordan [Fri, 19 Aug 2011 15:10:22 +0000 (11:10 -0400)]
ath9k: fix MGMT packets when using TKIP
Prevent 8 bytes from being truncated from MGMT packets
when using TKIP.
Signed-off-by: Bill Jordan <bjordan@rajant.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Tue, 16 Aug 2011 19:44:21 +0000 (21:44 +0200)]
b43: warn when forcing PIO mode
We have resolved all the known issues with DMA mode, however some users
(or distros) are still forcing PIO mode by config files. Without
debugging enabled it's not noticable at all. Add the warning for them.
Cc: Gregory Bellier <gregory.bellier@gmail.com>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Alex Hacker [Tue, 16 Aug 2011 10:41:40 +0000 (16:41 +0600)]
ath9k_hw: fix EIFS value to microseconds
The EIFS value read from AR_D_GBL_IFS_EIFS register in core clocks and then
written back as microsecond value.
Signed-off-by: Alex Hacker <hacker@epn.ru>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Mon, 15 Aug 2011 16:50:56 +0000 (18:50 +0200)]
b43: LCN-PHY: init 0x2064 radio
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Mon, 15 Aug 2011 16:50:55 +0000 (18:50 +0200)]
b43: LCN-PHY: basic PHY init
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Christian Lamparter [Mon, 15 Aug 2011 18:09:54 +0000 (20:09 +0200)]
carl9170: export HW random number generator
All AR9170 hardware have a 16-Bit random number generator.
The documentation claims the values are suitable for
"security keys".
The "throughput" is around 320Kibit/s. It's slow, but it
does work without introducing any special offload
firmware commands.
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Christian Lamparter [Mon, 15 Aug 2011 17:50:48 +0000 (19:50 +0200)]
carl9170: improve site survey
The firmware keeps track of channel usage. This data can
be used by the automatic channel selection to find the
*best* channel.
Survey data from wlan22
frequency: 2412 MHz [in use]
noise: -86 dBm
channel active time:
3339608 ms
channel busy time: 270982 ms
channel transmit time: 121515 ms
Survey data from wlan22
frequency: 2417 MHz
noise: -86 dBm
channel active time: 70 ms
channel busy time: 2 ms
channel transmit time: 1 ms
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Christian Lamparter [Mon, 15 Aug 2011 17:39:51 +0000 (19:39 +0200)]
carl9170: import updated firmware headers
Import new headers from our firmware branch:
git://git.kernel.org/pub/scm/linux/kernel/git/chr/carl9170fw.git
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Christian Lamparter [Mon, 15 Aug 2011 16:45:54 +0000 (18:45 +0200)]
carl9170: fix timekeeping for HW_COUNTER firmwares
AR9170_PWR_REG_PLL_ADDAC is used to set the main clock
divisor which affects the AHB/CPU speed. Because this
would interfere with the firmware internal timekeeping,
the function has to be moved into the firmware.
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Dan Carpenter [Mon, 15 Aug 2011 11:25:35 +0000 (14:25 +0300)]
libertas: handle mesh networks in lbs_iface_active()
There was an extra semicolon so the if condition wasn't used. We
checked "priv->dev" twice instead of "priv->mesh_dev".
Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Sun, 14 Aug 2011 23:23:09 +0000 (01:23 +0200)]
b43: LCN-PHY: add init tables
They were taken from MMIO dump with few RegExps and vim.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Sun, 14 Aug 2011 21:27:30 +0000 (23:27 +0200)]
b43: LCN-PHY: switch analog
Analog is switched on right after reading PHY version:
read16 0xfaafc3e0 -> 0xa801
phy_read(0x043b) -> 0x0000
phy_write(0x043b) <- 0x0000
Switched off after after killing radio:
>>> Switch Radio(OFF) end
phy_read(0x043c) -> 0x0000
phy_write(0x043c) <- 0x0007
phy_read(0x043b) -> 0x0000
phy_write(0x043b) <- 0x0007
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Sun, 14 Aug 2011 21:27:29 +0000 (23:27 +0200)]
b43: LCN-PHY: implement disabling radio
wl reads radio version, then disables it. That's how we found it in MMIO
dump:
radio_read(0x0000) -> 0x0031 <-- RADIO READ WITHOUT 0x200 SET!
radio_read(0x0001) -> 0x0064 <-- RADIO READ WITHOUT 0x200 SET!
radio_read(0x0002) -> 0x0020 <-- RADIO READ WITHOUT 0x200 SET!
read32 0xfaafc120 -> 0x04000400
phy_read(0x044d) -> 0x0000
phy_write(0x044d) <- 0x0000
phy_read(0x044c) -> 0x1fff
phy_write(0x044c) <- 0x1fff
phy_read(0x04b7) -> 0x0000
phy_write(0x04b7) <- 0x0000
phy_read(0x04b1) -> 0x0000
phy_write(0x04b1) <- 0x0000
phy_read(0x04b0) -> 0x7dff
phy_write(0x04b0) <- 0x7dff
phy_read(0x04fa) -> 0x0000
phy_write(0x04fa) <- 0x0000
phy_read(0x04f9) -> 0x007f
phy_write(0x04f9) <- 0x007f
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Sun, 14 Aug 2011 21:27:28 +0000 (23:27 +0200)]
b43: LCN-PHY: add very basic PHY ops
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Sun, 14 Aug 2011 18:16:37 +0000 (20:16 +0200)]
b43: fix DMA on some bugged hardware
Some hardware with 64-bit DMA uses lower address word for setting
routing (translation) bit. Add workaround for such boards.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Sun, 14 Aug 2011 17:39:40 +0000 (19:39 +0200)]
ssb: fix DMA translation for some specific boards
Signed-off-by: Michael Buesch <m@bues.ch>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Felix Fietkau [Sat, 13 Aug 2011 16:13:49 +0000 (10:13 -0600)]
ath9k: remove a bogus WARN_ON
On embedded hardware it's normal to not have a PCI device for the PCI
bridge that the wifi card is attached to. pdev->bus->self will be
NULL in that case. In that case, simply return without emitting an
useless kernel stack trace.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:18 +0000 (10:28 +0530)]
ath9k: Dump modal noisefloor calibration history
Debugfs file location:
<debugfs_mnt>/ieee80211/phy#/ath9k/dump_nfcal
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:17 +0000 (10:28 +0530)]
ath9k: Fix invalid noisefloor reading due to channel update
While switching b/w HT20/40, the current channel's nf values
are updated into history buffer. Since the current channel's
channel type, channel flag got updated before reading
nf value from hw. This channel type mismatch is causing invalid
readings when hw is on ht20 but getnf tries to read on extn chains.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:16 +0000 (10:28 +0530)]
ath9k: Fix noisefloor history update for extn chains
Before doing hw reset the current channel's noisefloor readings
are updated into history buffer. The extension chain's readings
are considered only if the current channel was configured in HT40.
While moving from HT40 to HT20, the extn chain's readings are
skipped though the current channel is in ht40. This patch updates
extn chain reading based on channel flag.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:15 +0000 (10:28 +0530)]
ath9k: Move ath9k_init_crypto to common
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:14 +0000 (10:28 +0530)]
ath9k: Remove SC_OP_ENABLE_APM
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:13 +0000 (10:28 +0530)]
ath9k_hw: Remove unnecessary chainmask configuration
The chainmasks were already configured at process_ini
before doing init calibration.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:12 +0000 (10:28 +0530)]
ath9k: qinfo never be NULL in setuptxqueue
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:11 +0000 (10:28 +0530)]
ath9k: Remove unused argument tsf from ath9k_hw_rxprocdesc
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:10 +0000 (10:28 +0530)]
ath9k: Re-enable RXOEL interrupt after processing rx buffers
Once RXEOL was disabled, it never be enabled again. This patch
re-enables rxeol at the end of rx tasklet.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:09 +0000 (10:28 +0530)]
ath9k: Fix rx overrun interrupt storm
Whenever RXEOL is received, both RXORN and RXEOL got cleared
to avoid rx overrun interrupt storm. This was handled only for
edma chips. The same scenario was also observered with AR9280,
doing frequent channel type switch b/w HT20/40 with bidi traffic
that is causing failure to stop rx dma. This patch clears
the RXEOL & RXORN interrupts for all chips.
ath: DMA failed to stop in 10 ms AR_CR=0x00000024 AR_DIAG_SW=0x42000020
DMADBG_7=0x000062c0
ath: Could not stop RX, we could be confusing the DMA engine when we
start RX up
------------[ cut here ]------------
WARNING: at drivers/net/wireless/ath/ath9k/recv.c:532
ath_stoprecv+0x110/0x120 [ath9k]()
Call Trace:
[<
ffffffff8104a55a>] warn_slowpath_common+0x7a/0xb0
[<
ffffffff8104a5a5>] warn_slowpath_null+0x15/0x20
[<
ffffffffa0560380>] ath_stoprecv+0x110/0x120 [ath9k]
[<
ffffffffa055e6fa>] ath_reset+0x6a/0x200 [ath9k]
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rajkumar Manoharan [Sat, 13 Aug 2011 04:58:08 +0000 (10:28 +0530)]
ath9k_hw: Fix exceed transmission burst-time of 5GHz
The WAR which adds extra delimiters when using RTS/CTS
with aggregation and non-enterprise AR9003 chips.
This extra padding is done after doing all the 4ms limit
checks and hence the total aggregate sizes are exceeding
the allowed duration. This patch limits the aggregate
sizes appropriately after including these extra delimiters.
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Tue, 9 Aug 2011 23:45:11 +0000 (16:45 -0700)]
{nl,cfg,mac}80211: let userspace make meshif mesh gate
Allow userspace to set NL80211_MESHCONF_GATE_ANNOUNCEMENTS attribute,
which will advertise this mesh node as being a mesh gate.
NL80211_HWMP_ROOTMODE must be set or this will do nothing.
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Tue, 9 Aug 2011 23:45:10 +0000 (16:45 -0700)]
{nl,cfg,mac}80211: let userspace set RANN interval
Allow userspace to set Root Announcement Interval for our mesh
interface. Also, RANN interval is now in proper units of TUs.
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Tue, 9 Aug 2011 23:45:09 +0000 (16:45 -0700)]
{nl,mac}80211: add missing root mode meshconf entries
This fix allows userspace to mark a meshif as a root node.
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Tue, 9 Aug 2011 23:45:08 +0000 (16:45 -0700)]
mac80211: mesh gate implementation
In this implementation, a mesh gate is a root node with a certain bit
set in its RANN flags. The mpath to this root node is marked as a path
to a gate, and added to our list of known gates for this if_mesh. Once a
path discovery process fails, we forward the unresolved frames to a
known gate. Thanks to Luis Rodriguez for refactoring and bug fix help.
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Tue, 9 Aug 2011 23:45:07 +0000 (16:45 -0700)]
mac80211: fix mesh path flushing
Previously, mpaths were never flushed since the mpath is not active once
we call this function.
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Tue, 9 Aug 2011 23:45:06 +0000 (16:45 -0700)]
mac80211: mesh locking fixes
mesh_queue_preq is invoked invoked from both user (work queue) and
softirq (timer) context, so the _bh version of spinlock needs to be
used. Also, the mpath->state_lock should be softirq safe as well.
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Tue, 9 Aug 2011 23:45:05 +0000 (16:45 -0700)]
mac80211: fix mpath timer NULL function
If we have an mpath whose timer has not been initialized, don't try to
delete it.
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Javier Cardona [Tue, 9 Aug 2011 23:45:04 +0000 (16:45 -0700)]
mac80211: improve mpath debugging
make hwmp_dbg print the relevant sdata->name by default and improve
formatting. Also add mpath_dbg macro for debugging of mesh path
operations.
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Tue, 23 Aug 2011 17:09:30 +0000 (19:09 +0200)]
b43: HT-PHY: allow writing longer tables with a single call
Sometimes we need to write table which is 2-10 elements long. It's
easier to create such a function instead of defining array every time.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Sat, 13 Aug 2011 15:54:04 +0000 (17:54 +0200)]
b43: HT-PHY: init: add missing small-tables writes
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Stefan Assmann [Sat, 13 Aug 2011 10:12:36 +0000 (12:12 +0200)]
rtlwifi: add module parameter to set global debug level
No need to recompile the module anymore to set the debug level.
Signed-off-by: Stefan Assmann <sassmann@kpanic.de>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Fri, 12 Aug 2011 23:41:12 +0000 (01:41 +0200)]
b43: HT-PHY: init: add some AFE (Analog Frontend) operation
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Fri, 12 Aug 2011 23:41:11 +0000 (01:41 +0200)]
b43: HT-PHY: init: add missing PHY mask/set ops
MMIO hacks were used to trick ndis&wl. For example following:
phy_read(0x0280) -> 0xffff
phy_write(0x0280) <- 0xff3e
***
phy_read(0x0280) -> 0x0000
phy_write(0x0280) <- 0x003e
was translated to mask 0xff00 and set 0x3e.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Fri, 12 Aug 2011 13:27:34 +0000 (15:27 +0200)]
b43: HT-PHY: init: init BPHY and upload 0x1a table
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Fri, 12 Aug 2011 11:13:46 +0000 (13:13 +0200)]
b43: HT-PHY: init: copy tables and reset CCA
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Fri, 12 Aug 2011 11:13:45 +0000 (13:13 +0200)]
b43: HT-PHY: init: implement few simple PHY writes
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Fri, 12 Aug 2011 11:13:44 +0000 (13:13 +0200)]
b43: HT-PHY: init: zero EXTG registers
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Nishant Sarmukadam [Fri, 12 Aug 2011 06:50:21 +0000 (12:20 +0530)]
mwl8k: Traffic to clients gets affected when one client leaves a cyrpto bss
When a client disassociates from a crypto enabled bss, data traffic to
other clients connected to the bss is stalled. This was due to a boolean
variable used to keep track if HW crypto is enabled i.e. if set key has
been called to add a key. This flag was being reset every time delete
key was called e.g when a station leaves the bss. Once the flag is
reset, rx status flags were not being set for connected clients which
disrupts traffic to these clients. Fix this issue by not resetting the
flag since we do not need to reset this flag during the life time of the
bss.
Signed-off-by: Nishant Sarmukadam <nishants@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Thomas Pedersen [Fri, 12 Aug 2011 02:35:15 +0000 (19:35 -0700)]
mac80211: update mesh path selection frame format
Make mesh path selection frames Mesh Action category, remove outdated
Mesh Path Selection category and defines, use updated reason codes, add
mesh_action_is_path_sel for readability, and update/correct path
selection IEs.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Thomas Pedersen [Fri, 12 Aug 2011 02:35:14 +0000 (19:35 -0700)]
ieee80211: add mesh action codes
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Thomas Pedersen [Sat, 13 Aug 2011 03:01:00 +0000 (20:01 -0700)]
mac80211: update mesh peering frame format
This patch updates the mesh peering frames to the format specified in
the recently ratified 802.11s standard. Several changes took place to
make this happen:
- Change RX path to handle new self-protected frames
- Add new Peering management IE
- Remove old Peer Link IE
- Remove old plink_action field in ieee80211_mgmt header
These changes by themselves would either break peering, or work by
coincidence, so squash them all into this patch.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Thomas Pedersen [Fri, 12 Aug 2011 02:35:12 +0000 (19:35 -0700)]
mac80211: update mesh peering frame codes
Have the mesh peering frames use the self-protected action and reason codes
specified in 802.11s and defined in ieee80211.h. Remove the local enums.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Thomas Pedersen [Fri, 12 Aug 2011 02:35:11 +0000 (19:35 -0700)]
ieee80211: introduce Self Protected Action codes
802.11s introduces a new action frame category, add action codes as well
as an entry in ieee80211_mgmt.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Thomas Pedersen [Fri, 12 Aug 2011 02:35:10 +0000 (19:35 -0700)]
mac80211: fix mesh beacon format
Correct ordering of IEs in the mesh beacon while removing unneeded IEs
from mesh peering frames. Set privacy bit in capability info if security
is enabled. Add utility functions to aid in construction
of IEs and reduce code duplication.
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Thu, 11 Aug 2011 22:03:26 +0000 (00:03 +0200)]
b43: make forcing clock common (HT-PHY also uses that)
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Thu, 11 Aug 2011 22:03:25 +0000 (00:03 +0200)]
b43: include HT-PHY in some common code
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Thu, 11 Aug 2011 21:46:44 +0000 (23:46 +0200)]
bcma: implement BCM4331 workaround for external PA lines
We need to disable ext. PA lines for reading SPROM. It's disabled by
default, but this patch allows using bcma after loading wl, which leaves
workaround enabled.
Cc: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Thu, 11 Aug 2011 15:16:27 +0000 (17:16 +0200)]
b43: support new RX header, noticed to be used in 598.314+ fw
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Thu, 11 Aug 2011 13:07:16 +0000 (15:07 +0200)]
b43: support new TX header, noticed to be used by 598.314+ fw
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Thu, 11 Aug 2011 13:07:15 +0000 (15:07 +0200)]
b43: use enum for firmware header format
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Rafał Miłecki [Thu, 11 Aug 2011 13:07:14 +0000 (15:07 +0200)]
b43: rename TX header formats
Replace "old" and "new" with number of the first firmware known to use
the given format.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Helmut Schaa [Thu, 11 Aug 2011 14:17:42 +0000 (16:17 +0200)]
mac80211: Tear down BA session on BAR tx failure
As described at [1] some STAs (i.e. Intel 5100 Windows) can end up
correctly BlockAcking incoming frames without delivering them to user
space if a AMPDU subframe got lost and we don't flush the receipients
reorder buffer with a BlockAckReq. This in turn results in stuck
connections.
According to 802.11n-2009 it is not necessary to send a BAR to flush
the recepients RX reorder buffer but we still do that to be polite.
However, assume the following frame exchange:
AP -> STA, AMPDU (failed)
AP -> STA, BAR (failed)
The client in question then ends up in the same situation and won't
deliver frames to userspace anymore since we weren't able to flush
its reorder buffer.
This is not a hypothetical situation but I was able to observe this
exact behavior during a stress test between a rt2800pci AP and a Intel
5100 Windows client.
In order to work around this issue just tear down the BA session as
soon as a BAR failed to be TX'ed.
[1] http://comments.gmane.org/gmane.linux.kernel.wireless.general/66867
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>