platform/kernel/linux-stable.git
13 years agomac80211: Fix locking bug on mesh path table access
Javier Cardona [Tue, 3 May 2011 23:57:16 +0000 (16:57 -0700)]
mac80211: Fix locking bug on mesh path table access

The mesh and mpp path tables are accessed from softirq and workqueue
context so non-irq locking cannot be used.  Or at least that's what
PROVE_RCU seems to tell us here:

[  431.240946] =================================
[  431.241061] [ INFO: inconsistent lock state ]
[  431.241061] 2.6.39-rc3-wl+ #354
[  431.241061] ---------------------------------
[  431.241061] inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage.
[  431.241061] kworker/u:1/1423 [HC0[0]:SC0[0]:HE1:SE1] takes:
[  431.241061]  (&(&newtbl->hashwlock[i])->rlock){+.?...}, at:
[<c14671bf>] mesh_path_add+0x167/0x257

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Check size of a new mesh path table for changes since allocation.
Javier Cardona [Tue, 3 May 2011 23:57:15 +0000 (16:57 -0700)]
mac80211: Check size of a new mesh path table for changes since allocation.

Not sure if I'm chasing a ghost here, seems like the
mesh_path->size_order needs to be inside an RCU-read section to prevent
that value from changing between table allocation and copying.  We have
observed crashes that might be caused by this.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Self-protected management frames are not robust
Thomas Pedersen [Tue, 3 May 2011 23:57:14 +0000 (16:57 -0700)]
mac80211: Self-protected management frames are not robust

They may contain encrypted information elements (as AMPE frames do)
but they are not encrypted.

Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211: allow setting MFP flag for a meshif
Thomas Pedersen [Tue, 3 May 2011 23:57:13 +0000 (16:57 -0700)]
nl80211: allow setting MFP flag for a meshif

Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211: allow installing keys for a meshif
Thomas Pedersen [Tue, 3 May 2011 23:57:12 +0000 (16:57 -0700)]
nl80211: allow installing keys for a meshif

Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211: Let userspace drive the peer link management states.
Javier Cardona [Tue, 3 May 2011 23:57:11 +0000 (16:57 -0700)]
nl80211: Let userspace drive the peer link management states.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: Use capability info to detect mesh beacons.
Javier Cardona [Wed, 4 May 2011 17:24:56 +0000 (10:24 -0700)]
cfg80211: Use capability info to detect mesh beacons.

Mesh beacons no longer use all-zeroes BSSID.  Beacon frames for MBSS,
infrastructure BSS, or IBSS are differentiated by the Capability
Information field in the Beacon frame.  A mesh STA sets the ESS and IBSS
subfields to 0 in transmitted Beacon or Probe Response management
frames.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoopen80211s: Stop using zero for address 3 in mesh plink mgmt frames
Javier Cardona [Tue, 3 May 2011 23:57:10 +0000 (16:57 -0700)]
open80211s: Stop using zero for address 3 in mesh plink mgmt frames

Previous versions of 11s draft used the all zeroes address.  Current
draft uses the same address as address 2.

Also, use the ANA-approved action category code for peer establishment frames.

Note: This breaks compatibility with previous mesh protocol instances.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Drop MESH_PLINK category and use new ANA-approved MESH_ACTION
Javier Cardona [Tue, 3 May 2011 23:57:09 +0000 (16:57 -0700)]
mac80211: Drop MESH_PLINK category and use new ANA-approved MESH_ACTION

Note: This breaks compatibility with previous mesh protocol instances.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Let userspace send action frames over mesh interfaces
Javier Cardona [Tue, 3 May 2011 23:57:08 +0000 (16:57 -0700)]
mac80211: Let userspace send action frames over mesh interfaces

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211: Introduce NL80211_MESH_SETUP_USERSPACE_AMPE
Javier Cardona [Tue, 3 May 2011 23:57:07 +0000 (16:57 -0700)]
nl80211: Introduce NL80211_MESH_SETUP_USERSPACE_AMPE

Introduce a new configuration option to support AMPE from userspace.

Prior to this series we only supported authentication in userspace: an
authentication daemon would authenticate peer candidates in userspace
and hand them over to the kernel.  From that point the mesh stack would
take over and establish a peer link (Mesh Peering Management).

These patches introduce support for Authenticated Mesh Peering Exchange
in userspace.  The userspace daemon implements the AMPE protocol and on
successfull completion create mesh peers and install encryption keys.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agolibertas: Use netdev_<level> or dev_<level> where possible
Joe Perches [Mon, 2 May 2011 23:49:15 +0000 (16:49 -0700)]
libertas: Use netdev_<level> or dev_<level> where possible

Using the more descriptive logging styles gives a bit
more information about the device being operated on.

Makes the object trivially smaller too.

$ size drivers/net/wireless/libertas/built-in.o.*
 187730    2973   38488  229191   37f47 drivers/net/wireless/libertas/built-in.o.new
 188195    2973   38488  229656   38118 drivers/net/wireless/libertas/built-in.o.old

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agolibertas: Convert lbs_pr_<level> to pr_<level>
Joe Perches [Mon, 2 May 2011 23:49:14 +0000 (16:49 -0700)]
libertas: Convert lbs_pr_<level> to pr_<level>

Use the standard pr_<level> functions eases grep a bit.

Added a few missing terminating newlines to messages.
Coalesced long formats.

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agossb: update list of devices supporting multiple 80211 cores
Rafał Miłecki [Sun, 8 May 2011 18:30:31 +0000 (20:30 +0200)]
ssb: update list of devices supporting multiple 80211 cores

Some of the BCM43xx chips contain cores that are attached to the SSB, but are
inactive as they do not connect to the external environment. These must not be
registered. Several of these types are handled in driver ssb; however, the
specific case of an inactive 802.11 cores is now treated in b43 and b43legacy.
Although the current setup works, this minor change will place all such
workarounds in ssb, and simplify the code in drivers b43 and b43legacy.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Larry Finger <Larry.Finger@lwfinger.net>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomwifiex: remove unnecessary struct mwifiex_opt_sleep_confirm_buffer
Amitkumar Karwar [Tue, 10 May 2011 02:00:18 +0000 (19:00 -0700)]
mwifiex: remove unnecessary struct mwifiex_opt_sleep_confirm_buffer

The structure definition is
struct mwifiex_opt_sleep_confirm_buffer {
       u8 hdr[4];
       struct mwifiex_opt_sleep_confirm ps_cfm_sleep;
} __packed;

For sleep_confirm command we already reserve 4 bytes (using skb_reserve())
for an interface header. It will be filled later by interface specific code.
We don't need "hdr[4]" element in above structure. So we can use
"struct mwifiex_opt_sleep_confirm" directly instead of
"struct mwifiex_opt_sleep_confirm_buffer".

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
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 agobcma: add Broadcom specific AMBA bus driver
Rafał Miłecki [Mon, 9 May 2011 16:56:46 +0000 (18:56 +0200)]
bcma: add Broadcom specific AMBA bus driver

Broadcom has released cards based on a new AMBA-based bus type. From a
programming point of view, this new bus type differs from AMBA and does
not use AMBA common registers. It also differs enough from SSB. We
decided that a new bus driver is needed to keep the code clean.

In its current form, the driver detects devices present on the bus and
registers them in the system. It allows registering BCMA drivers for
specified bus devices and provides them basic operations. The bus driver
itself includes two important bus managing drivers: ChipCommon core
driver and PCI(c) core driver. They are early used to allow correct
initialization.

Currently code is limited to supporting buses on PCI(e) devices, however
the driver is designed to be used also on other hosts. The host
abstraction layer is implemented and already used for PCI(e).

Support for PCI(e) hosts is working and seems to be stable (access to
80211 core was tested successfully on a few devices). We can still
optimize it by using some fixed windows, but this can be done later
without affecting any external code. Windows are just ranges in MMIO
used for accessing cores on the bus.

Cc: Greg KH <greg@kroah.com>
Cc: Michael Büsch <mb@bu3sch.de>
Cc: Larry Finger <Larry.Finger@lwfinger.net>
Cc: George Kashperko <george@znau.edu.ua>
Cc: Arend van Spriel <arend@broadcom.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Andy Botting <andy@andybotting.com>
Cc: linuxdriverproject <devel@linuxdriverproject.org>
Cc: linux-kernel@vger.kernel.org <linux-kernel@vger.kernel.org>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: don't drop frames where skb->len < 24 in ieee80211_scan_rx()
Luciano Coelho [Mon, 9 May 2011 16:15:04 +0000 (19:15 +0300)]
mac80211: don't drop frames where skb->len < 24 in ieee80211_scan_rx()

This seems to be a leftover from the old days, when we didn't support
any frames that didn't contain the full ieee802.11 header.  This is
not the case anymore.  It does not cause problems now, because they
are only dropped during scan.  But when scheduled scans get merged,
this would become a problem because we would drop all small frames
while scheduled scan is running.

To fix this, return RX_CONTINUE instead of RX_DROP_MONITOR.

Cc: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Failed to set default beacon rssi in AP/IBSS mode
Rajkumar Manoharan [Mon, 9 May 2011 13:41:29 +0000 (19:11 +0530)]
ath9k: Failed to set default beacon rssi in AP/IBSS mode

This beacon rssi will be used to set noisefloor during ani reset.

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Fix rssi update in ad-hoc mode
Rajkumar Manoharan [Mon, 9 May 2011 13:41:28 +0000 (19:11 +0530)]
ath9k: Fix rssi update in ad-hoc mode

The average beacon rssi which will be used by ani is not updated
in adhoc mode.

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: Change DCU backoff thresh for AR9340
Rajkumar Manoharan [Mon, 9 May 2011 13:41:27 +0000 (19:11 +0530)]
ath9k_hw: Change DCU backoff thresh for AR9340

By changing DCU backoff threshold for AR9340 to 1, helps to
reduce rx overrurns seen while running bidirectional traffic.

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: Corrected xpabiaslevel register settings for AR9340
Rajkumar Manoharan [Mon, 9 May 2011 13:41:26 +0000 (19:11 +0530)]
ath9k_hw: Corrected xpabiaslevel register settings for AR9340

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192cu: Fix memset/memcpy using sizeof(ptr) not sizeof(*ptr)
Joe Perches [Mon, 9 May 2011 19:51:57 +0000 (12:51 -0700)]
rtlwifi: rtl8192cu: Fix memset/memcpy using sizeof(ptr) not sizeof(*ptr)

Found via coccinelle script

@@
type T;
T* ptr;
expression E1;
@@
* memset(E1, 0, sizeof(ptr));

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: remove get_channel_noise function
Mohammed Shafi Shajakhan [Mon, 9 May 2011 04:59:01 +0000 (10:29 +0530)]
ath9k_hw: remove get_channel_noise function

currently ath9k_hw_getchan_noise is not used anywhere

Signed-off-by: Mohammed Shafi Shajakhan <mshajakhan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: trivial: include ssb word in ssb specific functions
Rafał Miłecki [Sun, 8 May 2011 22:21:18 +0000 (00:21 +0200)]
b43: trivial: include ssb word in ssb specific functions

This can be helpful when we decide to add support for other buses.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomwifiex: fix null derefs, mem leaks and trivia
Christoph Fritz [Sun, 8 May 2011 20:50:09 +0000 (22:50 +0200)]
mwifiex: fix null derefs, mem leaks and trivia

This patch:
 - adds kfree() where necessary
 - prevents potential null dereferences
 - makes use of kfree_skb()
 - replaces -1 for failed kzallocs with -ENOMEM

Signed-off-by: Christoph Fritz <chf.fritz@googlemail.com>
Reviewed-by: Kiran Divekar <dkiran@marvell.com>
Tested-by: Amitkumar Karwar <akarwar@marvell.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43legacy: drop ssb-duplicated workaround for dangling cores
Rafał Miłecki [Sun, 8 May 2011 18:30:33 +0000 (20:30 +0200)]
b43legacy: drop ssb-duplicated workaround for dangling cores

Remove the code to detect inactive 802.11 cores, as that function is now done
in ssb.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Larry Finger <Larry.Finger@lwfinger.net>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: drop ssb-duplicated workaround for dangling cores
Rafał Miłecki [Sun, 8 May 2011 18:30:32 +0000 (20:30 +0200)]
b43: drop ssb-duplicated workaround for dangling cores

Remove the code to detect inactive 802.11 cores, as that function is now done
in ssb.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Larry Finger <Larry.Finger@lwfinger.net>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Fix rmmod hang of rt2800pci
Gertjan van Wingerde [Sat, 7 May 2011 15:27:46 +0000 (17:27 +0200)]
rt2x00: Fix rmmod hang of rt2800pci

txstatus_timer should only be deleted for USB devices, as it is only
initialized for USB devices.

Reported-by: Andreas Hartmann <andihartmann@01019freenet.de>
Signed-off-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 agop54pci: fix -Wunused-but-set-variable warnings
Christian Lamparter [Sat, 7 May 2011 09:16:08 +0000 (11:16 +0200)]
p54pci: fix -Wunused-but-set-variable warnings

p54pci.c: In function ‘p54p_tx’:
p54pci.c:334:6: warning: variable ‘device_idx’ set but not used

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocarl9170: fix -Wunused-but-set-variable warnings
Christian Lamparter [Sat, 7 May 2011 09:13:37 +0000 (11:13 +0200)]
carl9170: fix -Wunused-but-set-variable warnings

tx.c: In function ‘carl9170_tx_accounting_free’:
tx.c:159:28: warning: variable ‘txinfo’ set but not used
tx.c: In function ‘carl9170_tx_status_process_ampdu’:
tx.c:383:27: warning: variable ‘ar_info’ set but not used
tx.c: In function ‘__carl9170_tx_process_status’:
tx.c:626:27: warning: variable ‘arinfo’ set but not used
tx.c: In function ‘carl9170_tx_ampdu_queue’:
tx.c:1324:15: warning: variable ‘max’ set but not used

Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Remove need to disable ASPM
Chaoming Li [Fri, 6 May 2011 20:32:02 +0000 (15:32 -0500)]
rtlwifi: rtl8192se: Remove need to disable ASPM

When this driver was initially submitted, the system would crash unless
ASPM was disabled. This problem has been fixed.

This patch also adds a printk that outputs the name of the firmware
file that is used.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: Move 2 large arrays off stack
Larry Finger [Fri, 6 May 2011 18:56:18 +0000 (13:56 -0500)]
rtlwifi: Move 2 large arrays off stack

In driver rtlwifi, efuse_read() places two relatively large arrays on the
stack - a 1D u8 array of size 128, and a 2D array of u16 with 128 * 4 elements.
With driver rtl8192de, the sizes will be 256 and 256 * 4 respectively. As that
will make the 2D array be 2048 bytes, I have changed the code to use kmalloc to
allocate the space.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlagn: fix iwl_is_any_associated
Johannes Berg [Fri, 6 May 2011 18:11:20 +0000 (11:11 -0700)]
iwlagn: fix iwl_is_any_associated

The function iwl_is_any_associated() was intended
to check both contexts, but due to an oversight
it only checks the BSS context. This leads to a
problem with scanning since the passive dwell
time isn't restricted appropriately and a scan
that includes passive channels will never finish
if only the PAN context is associated since the
default dwell time of 120ms won't fit into the
normal 100 TU DTIM interval.

Fix the function by using for_each_context() and
also reorganise the other functions a bit to take
advantage of each other making the code easier to
read.

Cc: stable@kernel.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: process TSF out of range before RX
Rajkumar Manoharan [Fri, 6 May 2011 12:57:47 +0000 (18:27 +0530)]
ath9k: process TSF out of range before RX

Processing TSF out of range before RX helps to update beacon
timers so early in the succeeding rx process.

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: avoid enabling interrupts while processing rx
Rajkumar Manoharan [Fri, 6 May 2011 12:57:46 +0000 (18:27 +0530)]
ath9k: avoid enabling interrupts while processing rx

The assumsion is that while processing ath9k tasklet,
interrupts were already disabled and it will be enabled
at the completion of ath9k tasklet. But whenever TSFOOR is raised,
the driver configures the beacon timers after having received a
beacon frame from the AP which inturn enables the interrupts.

Cc: stable@kernel.org
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Fix build error when CONFIG_PM is not defined
Larry Finger [Thu, 5 May 2011 23:48:47 +0000 (18:48 -0500)]
mac80211: Fix build error when CONFIG_PM is not defined

When mac80211 is built without CONFIG_PM being defined, the following errors
are output:

net/mac80211/main.c: In function ‘ieee80211_register_hw’:
net/mac80211/main.c:700: error: ‘const struct ieee80211_ops’ has no member named ‘suspend’
net/mac80211/main.c:700: error: ‘const struct ieee80211_ops’ has no member named ‘resume’
make[2]: *** [net/mac80211/main.o] Error 1
make[1]: *** [net/mac80211] Error 2
make[1]: *** Waiting for unfinished jobs....
make: *** [net] Error 2

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlagn: led stay solid on when no traffic
Wey-Yi Guy [Sat, 30 Apr 2011 15:38:16 +0000 (08:38 -0700)]
iwlagn: led stay solid on when no traffic

commit 5ed540aecc2aae92d5c97b9a9306a5bf88ad5574 change the led behavior
for iwlwifi driver; the side effect cause led blink all the time.

Modify the led blink table to fix this problem

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agowireless: Fix warnings due to -Wunused-but-set-variable
Rajkumar Manoharan [Mon, 25 Apr 2011 10:26:16 +0000 (15:56 +0530)]
wireless: Fix warnings due to -Wunused-but-set-variable

These warnings are exposed by gcc 4.6.
net/wireless/reg.c: In function 'freq_reg_info_regd':
net/wireless/reg.c:675:38: warning: variable 'pr' set but not used
[-Wunused-but-set-variable]
net/wireless/lib80211_crypt_wep.c: In function 'lib80211_wep_build_iv':
net/wireless/lib80211_crypt_wep.c:99:12: warning: variable 'len' set but
not used [-Wunused-but-set-variable]

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoMerge branch 'wireless-next-2.6' of git://git.kernel.org/pub/scm/linux/kernel/git...
John W. Linville [Tue, 10 May 2011 19:49:11 +0000 (15:49 -0400)]
Merge branch 'wireless-next-2.6' of git://git./linux/kernel/git/iwlwifi/iwlwifi-2.6

13 years agoiwlagn: check DMA mapping errors
Johannes Berg [Thu, 28 Apr 2011 14:27:10 +0000 (07:27 -0700)]
iwlagn: check DMA mapping errors

DMA mappings can fail, but the current code
doesn't check for that. Add checking, which
requires some restructuring for proper error
paths.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: remove bytecount indirection
Johannes Berg [Thu, 28 Apr 2011 14:27:09 +0000 (07:27 -0700)]
iwlagn: remove bytecount indirection

All AGN devices need the bytecount table, so
remove the indirection and make the functions
static again.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: dont update bytecount table for command queue
Johannes Berg [Thu, 28 Apr 2011 14:27:08 +0000 (07:27 -0700)]
iwlagn: dont update bytecount table for command queue

The device doesn't use the bytecount table for the
command queue, only for aggregation queues to make
aggregation decisions. So don't update it for the
command queue (and we even updated it with wrong
values).

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: remove unused variable
Johannes Berg [Thu, 28 Apr 2011 14:27:07 +0000 (07:27 -0700)]
iwlagn: remove unused variable

The variable 'len' here is set but never used.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: remove frame pre-allocation
Johannes Berg [Fri, 29 Apr 2011 16:48:14 +0000 (09:48 -0700)]
iwlagn: remove frame pre-allocation

The frame pre-allocation is quite a bit of complex
code, all to avoid a single allocation. Remove it
and consolidate the beacon sending code.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlagn: remove get_hcmd_size indirection
Johannes Berg [Thu, 28 Apr 2011 14:27:05 +0000 (07:27 -0700)]
iwlagn: remove get_hcmd_size indirection

There's no need for this, all commands are the right size.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agoiwlwifi: support the svtool messages interactions through nl80211 test mode
Cindy H. Kao [Fri, 6 May 2011 17:40:15 +0000 (10:40 -0700)]
iwlwifi: support the svtool messages interactions through nl80211 test mode

This patch adds the feature to support the test mode operation through
the generic netlink channel NL80211_CMD_TESTMODE between intel
wireless device iwlwifi and the user space application svtool.

The main purpose is to create a transportation layer between the iwlwifi
device and the user space application so that the interaction between the
user space application svtool and the iwlwifi device in the kernel space is
in a way of generic netlink messaging.

The detail specific functions are:

1. The function iwl_testmode_cmd() is added to digest the svtool test command
   from the user space application. The svtool test commands are categorized  to
   three types : commands to be processed by the device ucode, commands to access
   the registers, and commands to be processed at the driver level(such as reload
   the ucode). iwl_testmode_cmd() dispatches the commands the corresponding handlers
   and reply to user space regarding the command execution status. Extra data is
   returned to the user space application if there's any.

2. The function iwl_testmode_ucode_rx_pkt() is added to multicast all the spontaneous
   messages from the iwlwifi device to the user space. Regardless the message types,
   whenever there is a valid spontaneous message received by the iwlwifi ISR,
   iwl_testmode_ucode_rx_pkt() is invoked to multicast the message content to user
   space. The message content is not attacked and the message parsing is left to
   the user space application.

Implementation guidelines:

1. The generic netlink messaging for iwliwif test mode is through  NL80211_CMD_TESTMODE
   channel, therefore, the codes need to follow the regulations set by cfg80211.ko
   to get the actual device instance ieee80211_ops via cfg80211.ko, so that the iwlwifi
   device is indicated with ieee80211_ops and can be actually accessed.

   Therefore, a callback iwl_testmode_cmd() is added to the structure
   iwlagn_hw_ops in iwl-agn.c.

2. It intends to utilize those low level device access APIs from iwlwifi device driver
   (ie. iwlagn.ko) rather than creating it's own set of device access functions.
   For example, iwl_send_cmd(), iwl_read32(), iwl_write8(), and iwl_write32() are reused.

3. The main functions are maintained in new files instead of spreading all over the
   existing iwlwifi driver files.

   The new files added are :

   drivers/net/wireless/iwlwifi/iwl-sv-open.c
        - to handle the user space test mode application command
          and reply the respective command status to the user space application.
        - to multicast the spontaneous messages from device to user space.

   drivers/net/wireless/iwlwifi/iwl-testmode.h
        - the commonly referenced definitions for the TLVs used in
          the generic netlink messages

Signed-off-by: Cindy H. Kao <cindy.h.kao@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
13 years agomwl8k: Do not ask mac80211 to generate IV for crypto keys
Yogesh Ashok Powar [Thu, 5 May 2011 11:00:48 +0000 (16:30 +0530)]
mwl8k: Do not ask mac80211 to generate IV for crypto keys

Since firmware is capable of generating IV's for all crypto
suits (TKIP, CCMP and WEP), do not ask mac80211 to generate
IV when HW crypto is being used. Instead only reserve
appropriate space in tx skb's in the driver, so that the
firmware can write IV's values.

Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43legacy: drop invalid IMCFGLO workaround
Rafał Miłecki [Thu, 5 May 2011 10:46:04 +0000 (12:46 +0200)]
b43legacy: drop invalid IMCFGLO workaround

We were performing it on wrong core, it was outdated and is already
implemented in ssb.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Tested-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agob43: drop invalid IMCFGLO workaround
Rafał Miłecki [Thu, 5 May 2011 10:45:52 +0000 (12:45 +0200)]
b43: drop invalid IMCFGLO workaround

We were performing it on wrong core, it was outdated and is already
implemented in ssb.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: fix power for the HT40 duplicate frames
Luis R. Rodriguez [Wed, 4 May 2011 21:01:26 +0000 (14:01 -0700)]
ath9k_hw: fix power for the HT40 duplicate frames

With AR9003 at about ~ 10 feet from an AP that uses RTS / CTS you
will be able to associate but not not get data through given that
the power for the rates used was set too low. This increases the
power and permits data connectivity at longer distances from
access points when connected with HT40. Without this you will not
get any data through when associated to APs configured in HT40
at about more than 10 feet away.

Cc: stable@kernel.org
Cc: Fiona Cain <fcain@atheros.com>
Cc: Zhen Xie <Zhen.Xie@Atheros.com>
Cc: Kathy Giori <kathy.giori@atheros.com>
Cc: Neha Choksi <neha.choksi@atheros.com>
Cc: Wayne Daniel <wayne.daniel@atheros.com>
Cc: Gaurav Jauhar <gaurav.jauhar@atheros.com>
Cc: Samira Naraghi <samira.naraghi@atheros.com>
CC: Ashok Chennupati <ashok.chennupati@atheros.com>
Cc: Lance Zimmerman <lance.zimmerman@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Fix transfer speed regression for USB hardware
Ivo van Doorn [Wed, 4 May 2011 19:42:05 +0000 (21:42 +0200)]
rt2x00: Fix transfer speed regression for USB hardware

Patch:
  rt2x00: Make rt2x00_queue_entry_for_each more flexible
  commit: 10e11568ca8b8a15f7478f6a4ceebabcbdba1018

introduced a severe regression on the throughput
for USB hardware. It turns out that the exiting of
the rt2x00queue_for_each_entry() was done too early.

The exact cause for this regression is unknown,
but by disabling the premature exiting of the loop
seems to resolve the issue.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Reported-by: Yasushi SHOJI <yashi@atmark-techno.com>
Reported-by: Balint Viragh <bviragh@dension.com>
Tested-by: Balint Viragh <bviragh@dension.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Initial support for RT5370 USB devices.
Gertjan van Wingerde [Wed, 4 May 2011 19:41:36 +0000 (21:41 +0200)]
rt2x00: Initial support for RT5370 USB devices.

Add necessary RF chipset define and basic support for these devices.

Tested-by: Juan Carlos Garza <juancarlosgarza@gmail.com>
Signed-off-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 agoath9k_hw: do noise floor calibration only on required chains
Rajkumar Manoharan [Wed, 4 May 2011 14:07:17 +0000 (19:37 +0530)]
ath9k_hw: do noise floor calibration only on required chains

At present the noise floor calibration is processed in supported
control and extension chains rather than required chains.
Unnccesarily doing nfcal in all supported chains leads to
invalid nf readings on extn chains and these invalid values
got updated into history buffer. While loading those values
from history buffer is moving the chip to deaf state.

This issue was observed in AR9002/AR9003 chips while doing
associate/dissociate in HT40 mode and interface up/down
in iterative manner. After some iterations, the chip was moved
to deaf state. Somehow the pci devices are recovered by poll work
after chip reset. Raading the nf values in all supported extension chains
when the hw is not yet configured in HT40 mode results invalid values.

Cc: stable@kernel.org
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: add basic support for WoWLAN
Johannes Berg [Wed, 4 May 2011 13:37:29 +0000 (15:37 +0200)]
mac80211: add basic support for WoWLAN

This adds basic support for the new WoWLAN
configuration in mac80211. The behaviour is
completely offloaded to the driver though,
with two new callbacks (suspend/resume).

Options for the driver include a complete
reconfiguration after wakeup, and exposing
all the triggers it wants to support.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211/cfg80211: WoWLAN support
Johannes Berg [Wed, 4 May 2011 13:37:28 +0000 (15:37 +0200)]
nl80211/cfg80211: WoWLAN support

This is based on (but now quite far from) the
original work from Luis and Eliad. Add support
for configuring WoWLAN triggers, and getting
the configuration out again. Changes from the
original patchset are too numerous to list,
but one important change needs highlighting:
the suspend() callback is passed NULL for the
trigger configuration if userspace has not
configured WoWLAN at all.

Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocarl9170: fix allmulticast mode
Nicolas Cavallari [Wed, 4 May 2011 13:26:52 +0000 (15:26 +0200)]
carl9170: fix allmulticast mode

Currently, the driver only disable multicast filtering when the
FIF_ALLMULTI driver flag has been just set (ie,
if changed_flags& FIF_ALLMULTI and *new_flags& FIF_ALLMULTI) or else
it will reenable multicast filtering.

But next time, this condition will be false and multicast filtering
will be reenabled, even through FIF_ALLMULTI is still set.

This mean that allmulticast only works for less than two minutes in
ad-hoc mode.  This patch fixes that to disable multicast filtering
as long as FIF_ALLMULTI is set.

Signed-off-by: Nicolas Cavallari <cavallar@lri.fr>
Acked-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomwl8k: Fix broken WEP
Yogesh Ashok Powar [Wed, 4 May 2011 11:52:16 +0000 (17:22 +0530)]
mwl8k: Fix broken WEP

The WEP key length was being set to 0 erroneously which broke WEP support.
Fix the same by setting the key length appropriately.

Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agocfg80211: Remove unused wiphy flag
Jouni Malinen [Wed, 4 May 2011 05:50:15 +0000 (22:50 -0700)]
cfg80211: Remove unused wiphy flag

The only user of WIPHY_FLAG_SUPPORTS_SEPARATE_DEFAULT_KEYS was removed
and consequently, this flag can be removed, too. In addition, a single
capability flag was not enough to indicate this capability clearly since
the device behavior may be different based on which operating mode is
being used.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agonl80211: Fix set_key regression with some drivers
Jouni Malinen [Wed, 4 May 2011 05:45:16 +0000 (22:45 -0700)]
nl80211: Fix set_key regression with some drivers

Commit dbd2fd656f2060abfd3a16257f8b51ec60f6d2ed added a mechanism for
user space to indicate whether a default key is being configured for
only unicast or only multicast frames instead of all frames. This
commit added a driver capability flag for indicating whether separate
default keys are supported and validation of the set_key command based
on that capability.

However, this single capability flag is not enough to cover possible
difference based on mode (AP/IBSS/STA) and the way this change was
introduced resulted in a regression with drivers that do not indicate
the new capability (i.e.., more or less any non-mac80211 driver using
cfg80211) when using a recent wpa_supplicant snapshot.

Fix the regression by removing the new check which is not strictly
speaking needed. The new separate default key functionality is needed
only for RSN IBSS which has a separate capability indication.

Cc: stable@kernel.org
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomwifiex: remove unnecessary variable initialization
Yogesh Ashok Powar [Wed, 4 May 2011 03:11:46 +0000 (20:11 -0700)]
mwifiex: remove unnecessary variable initialization

Skip initialization of local variables with some default values
if the values are not going to be used further down the code path.

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 agomwifiex: fix simultaneous assoc and scan issue
Amitkumar Karwar [Wed, 4 May 2011 03:11:45 +0000 (20:11 -0700)]
mwifiex: fix simultaneous assoc and scan issue

When scan and assoc (infra/ibss) commands are simultaneously
given in two terminals, association response is erroneously
served while serving the scan response.

mwifiex_cfg80211_results() is the common routine for sending
ioctl (scan, assoc etc.) results to cfg80211 stack. In above
scenario even if the common routine is called for scan ioctl
context, it also tries to send information about assoc ioctl to
cfg80211 because "priv->assoc_request/priv->ibss_join_request"
flag is on at that time.

Fix the issue by updating request variable after assoc handling
and modifying the variable check in mwifiex_cfg80211_results.

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kiran Divekar <dkiran@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Modify Kconfig and Makefile routines for new driver
Chaoming Li [Tue, 3 May 2011 14:49:36 +0000 (09:49 -0500)]
rtlwifi: rtl8192se: Modify Kconfig and Makefile routines for new driver

Modify rtlwifi routines for rtl8192se and set up Kconfig
and Makefile for new driver.

This patch also disables ASPM for the RTL8192SE to prevent some strange
crashes on LF's system.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge TX and RX routines
Chaoming Li [Tue, 3 May 2011 14:49:26 +0000 (09:49 -0500)]
rtlwifi: rtl8192se: Merge TX and RX routines

Merge routines trx.c and trx.h for RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge table routines
Chaoming Li [Tue, 3 May 2011 14:49:16 +0000 (09:49 -0500)]
rtlwifi: rtl8192se: Merge table routines

Merge routines table.c and table.h for RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge main (sw) routines
Chaoming Li [Tue, 3 May 2011 14:49:06 +0000 (09:49 -0500)]
rtlwifi: rtl8192se: Merge main (sw) routines

Merge routines sw.c and sw.h for RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge rf routines
Chaoming Li [Tue, 3 May 2011 14:48:56 +0000 (09:48 -0500)]
rtlwifi: rtl8192se: Merge rf routines

Merge routines rf.c and rf.h for RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge register definitions
Chaoming Li [Tue, 3 May 2011 14:48:45 +0000 (09:48 -0500)]
rtlwifi: rtl8192se: Merge register definitions

Merge routines reg.h for RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge phy routines
Chaoming Li [Tue, 3 May 2011 14:48:35 +0000 (09:48 -0500)]
rtlwifi: rtl8192se: Merge phy routines

Merge routines phy.c and phy.h for RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge led routines
Chaoming Li [Tue, 3 May 2011 14:48:25 +0000 (09:48 -0500)]
rtlwifi: rtl8192se: Merge led routines

Merge routines led.c and led.h for RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge hardware routines
Chaoming Li [Tue, 3 May 2011 14:48:15 +0000 (09:48 -0500)]
rtlwifi: rtl8192se: Merge hardware routines

Merge routines hw.c and hw.h for RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge firmware routines
Chaoming Li [Tue, 3 May 2011 14:48:05 +0000 (09:48 -0500)]
rtlwifi: rtl8192se: Merge firmware routines

Merge routines fw.c and fw.h for RTL8192SE. In addition, make changes
to rtlwifi/wifi.h to support RTL8192SE.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge dynamic management routines
Chaoming Li [Tue, 3 May 2011 14:47:55 +0000 (09:47 -0500)]
rtlwifi: rtl8192se: Merge dynamic management routines

Merge routines dm.c and dm.h for RTL8192SE.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: rtl8192se: Merge def.h
Chaoming Li [Tue, 3 May 2011 14:47:45 +0000 (09:47 -0500)]
rtlwifi: rtl8192se: Merge def.h

Introduce routine def.h for rtl8192se.

Signed-off-by: Chaoming_Li <chaoming_li@realsil.com.cn>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: Postpond ps timer if tx is stopped by others
Rajkumar Manoharan [Tue, 3 May 2011 11:33:59 +0000 (17:03 +0530)]
mac80211: Postpond ps timer if tx is stopped by others

Whenever the driver's queue depth reaches the max, the queues are
stopped by the driver till the driver can accept the frames.
At the mean time dynamic_ps_timer can be expired due to not
receiving packet from upper layer which could restart the transmission
at the end of ps work. Due to the mismatch with driver state,
mac80211 is unneccesarity buffering all the frames till the driver
wakes up the queue.

Check whether there is no transmit or the tx queues were stopped by some
reasons. If any of the queue was stopped, the postpond ps timer and
do not restart netif_tx.

Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: set TID of internal mgmt packets to 7
Arik Nemtsov [Tue, 3 May 2011 08:53:07 +0000 (11:53 +0300)]
mac80211: set TID of internal mgmt packets to 7

The queue mapping of internal mgmt packets is set to VO. Set the TID
value to match the queue mapping. Otherwise drivers that only look at
the TID might get confused.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k_hw: remove aggregation protection mode
Mohammed Shafi Shajakhan [Tue, 3 May 2011 07:44:06 +0000 (13:14 +0530)]
ath9k_hw: remove aggregation protection mode

when aggregation protection mode is enabled the hardware needs
to send RTS/CTS for each HT frame. Currently its disabled so
remove the unused call backs.

Signed-off-by: Mohammed Shafi Shajakhan <mshajakhan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agolibertas: remove tx_timeout handler
Daniel Drake [Mon, 2 May 2011 20:38:16 +0000 (21:38 +0100)]
libertas: remove tx_timeout handler

As described at http://marc.info/?l=linux-netdev&m=130428493104730&w=2
libertas frequently generates spurious tx timeouts, because the tx
queue is brought down for extended periods during scanning. The net
layer takes a look and incorrectly assumes the queue has been down for
several seconds, and generates a tx_timeout.

One way to fix this is to bump the trans_start counter while scanning
so that the network layer knows that the device is still alive, but
I think the tx_timeout handler is implemented wrongly here and not of
any real use, so I vote to remove it.

As explained at http://marc.info/?l=linux-wireless&m=130430311115755&w=2
the watchdog is primarily meant to deal with lockup on the hardware TX
path (detected by the tx queue being stopped for an extended period of
time), but this is unlikely to happen with libertas. In this case, the tx
queue is stopped only while waiting for lbs_thread to send the queued frame
to the driver, and lbs_thread wakes up the queue immediately after, even
if the frame could not be sent correctly.

So, the only hardware-related possibility that this catches is if
hw_host_to_card hangs - this is something I have never seen. And if it
were to happen, nothing done by lbs_tx_timeout would actually wake up
lbs_thread any quicker than otherwise.

Removing this oddly-behaving spuriously-firing tx_timeout handler should
fix an occasional kernel crash during resume
(http://dev.laptop.org/ticket/10748)

Signed-off-by: Daniel Drake <dsd@laptop.org>
Acked-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: use wake_queue to restart trasmit
Rajkumar Manoharan [Fri, 29 Apr 2011 09:18:45 +0000 (14:48 +0530)]
mac80211: use wake_queue to restart trasmit

netif_tx_start_all_queues is used to allow the upper layer
to transmit frames but it does not restart transmission.
To restart the trasmission use netif_tx_wake_all_queues.
Not doing so, sometimes stalls the transmission and the
application has to be restarted to proceed further.

This issue was originally found while sending udp traffic
in higer bandwidth in open environment without bgscan.

Cc: stable@kernel.org
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Fix drain txq failure in flush
Rajkumar Manoharan [Thu, 28 Apr 2011 10:01:57 +0000 (15:31 +0530)]
ath9k: Fix drain txq failure in flush

While draining the txq in flush, the buffers can be
added into the tx queue by tx_tasklet which leads to
unneccesary chip reset.

This issue was originially found with AR9382 and
running heavy uplink udp traffic with higher bandwidth
and doing frequent bgscan.

Cc: stable@kernel.org
Signed-off-by: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: Rajkumar Manoharan <rmanoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoMerge branch 'for-linville' of git://git.kernel.org/pub/scm/linux/kernel/git/luca...
John W. Linville [Thu, 5 May 2011 17:56:27 +0000 (13:56 -0400)]
Merge branch 'for-linville' of git://git./linux/kernel/git/luca/wl12xx

13 years agomac80211: Fix a warning due to skipping tailroom reservation for IV
Mohammed Shafi Shajakhan [Thu, 5 May 2011 05:29:24 +0000 (10:59 +0530)]
mac80211: Fix a warning due to skipping tailroom reservation for IV

The devices that require IV generation in software need tailroom
reservation  for ICVs used in TKIP or WEP encryptions.
Currently, decision to skip the tailroom reservation in the tx
path was taken only on whether driver wants MMIC to be generated
in software or not. Following patch appends IV generation check for
such decisions and fixes the following warning.

WARNING: at net/mac80211/wep.c:101 ieee80211_wep_add_iv+0x56/0xf3()
Hardware name: 64756D6
Modules linked in: ath9k ath9k_common ath9k_hw
Pid: 0, comm: swapper Tainted: G        W   2.6.39-rc5-wl
 Call Trace:
[<c102fd29>] warn_slowpath_common+0x65/0x7a
[<c1465c4e>] ? ieee80211_wep_add_iv+0x56/0xf3
[<c102fd4d>] warn_slowpath_null+0xf/0x13
[<c1465c4e>] ieee80211_wep_add_iv+0x56/0xf3
[<c1466007>] ieee80211_crypto_wep_encrypt+0x63/0x88
[<c1478bf3>] ieee80211_tx_h_encrypt+0x2f/0x63
[<c1478cba>] invoke_tx_handlers+0x93/0xe1
[<c1478eda>] ieee80211_tx+0x4b/0x6d
[<c147907c>] ieee80211_xmit+0x180/0x188
[<c147779d>] ? ieee80211_skb_resize+0x95/0xd9
[<c1479edf>] ieee80211_subif_start_xmit+0x64f/0x668
[<c13956fc>] dev_hard_start_xmit+0x368/0x48c
[<c13a8bd6>] sch_direct_xmit+0x4d/0x101
[<c1395ae1>] dev_queue_xmit+0x2c1/0x43f
[<c13a74a2>] ? eth_header+0x1e/0x90
[<c13a7400>] ? eth_type_trans+0x91/0xc2
[<c13a7484>] ? eth_rebuild_header+0x53/0x53
[<c139f079>] neigh_resolve_output+0x223/0x27e
[<c13c6b23>] ip_finish_output2+0x1d4/0x1fe
[<c13c6bc6>] ip_finish_output+0x79/0x7d
[<c13c6cbe>] T.1075+0x43/0x48
[<c13c6e6e>] ip_output+0x75/0x7b
[<c13c4970>] dst_output+0xc/0xe
[<c13c62c9>] ip_local_out+0x17/0x1a
[<c13c67bb>] ip_queue_xmit+0x2aa/0x2f8
[<c138b742>] ? sk_setup_caps+0x21/0x92
[<c13d95ea>] ? __tcp_v4_send_check+0x7e/0xb7
[<c13d5d2e>] tcp_transmit_skb+0x6a1/0x6d7
[<c13d533b>] ? tcp_established_options+0x20/0x8b
[<c13d6f28>] tcp_retransmit_skb+0x43a/0x527
[<c13d8d6d>] tcp_retransmit_timer+0x32e/0x45d
[<c13d8f23>] tcp_write_timer+0x87/0x16c
[<c103a030>] run_timer_softirq+0x156/0x1f9
[<c13d8e9c>] ? tcp_retransmit_timer+0x45d/0x45d
[<c1034d65>] __do_softirq+0x97/0x14a
[<c1034cce>] ? irq_enter+0x4d/0x4d

Cc: Yogesh Powar <yogeshp@marvell.com>
Reported-by: Fabio Rossi <rossi.f@inwind.it>
Tested-by: Fabio Rossi <rossi.f@inwind.it>
Signed-off-by: Mohammed Shafi Shajakhan <mshajakhan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomwifiex: fix missing tsf_val TLV
Bing Zhao [Mon, 2 May 2011 18:00:45 +0000 (11:00 -0700)]
mwifiex: fix missing tsf_val TLV

In mwifiex_cmd_append_tsf_tlv(), two tsf_val TLVs should be
filled in the buffer and then sent to firmware.

The missing first TLV for tsf_val is added back in this patch.

Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: Fix typo in pci.c
Larry Finger [Mon, 2 May 2011 03:30:54 +0000 (22:30 -0500)]
rtlwifi: Fix typo in pci.c

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Add autowake support for USB hardware
Ivo van Doorn [Sat, 30 Apr 2011 15:18:18 +0000 (17:18 +0200)]
rt2x00: Add autowake support for USB hardware

The USB drivers don't support automatically waking up when in powersaving mode,
add a work object which will wakeup the device in time to receive the next beacon.
Based on that beacon, we either go back into powersaving mode, or we remain awake
to receive the buffered frames for our station.

Some part of the code, especially rt2x00lib_find_ie and rt2x00lib_rxdone_check_ps
are inspired on the code from carl9170.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Introduce capability flag for Bluetooth co-existence.
Gertjan van Wingerde [Sat, 30 Apr 2011 15:15:37 +0000 (17:15 +0200)]
rt2x00: Introduce capability flag for Bluetooth co-existence.

Use flag instead of re-reading the eeprom every time.

Signed-off-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 agort2x00: Streamline rt2800 eeprom initialisations.
Gertjan van Wingerde [Sat, 30 Apr 2011 15:15:13 +0000 (17:15 +0200)]
rt2x00: Streamline rt2800 eeprom initialisations.

In rt2800lib.c the rt2800_init_eeprom function the same eeprom
words were read multiple times, due to inefficient ordering of the
eeprom checks.

Reorder the checks so that each EEPROM word only has to be read once.

Signed-off-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 agort2x00: Fix optimize register access for rt2800pci
Ivo van Doorn [Sat, 30 Apr 2011 15:14:49 +0000 (17:14 +0200)]
rt2x00: Fix optimize register access for rt2800pci

The patch
rt2x00: Optimize register access in rt2800pci
from Helmut Schaa missed one register call, namely
the rt2800_register_multiwrite which should be changed
to rt2x00pci_register_multiwrite.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Acked-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Optimize TX_STA_FIFO register reading
Ivo van Doorn [Sat, 30 Apr 2011 15:14:23 +0000 (17:14 +0200)]
rt2x00: Optimize TX_STA_FIFO register reading

Add recycling functionality to rt2x00usb_register_read_async.
When the callback function returns true, resubmit the urb to
read the register again.

This optimizes the rt2800usb driver when multiple TX status reports
are pending in the register, because now we don't need to allocate
the rt2x00_async_read_data and urb structure each time.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Reduce tx status reading timeout
Ivo van Doorn [Sat, 30 Apr 2011 15:13:46 +0000 (17:13 +0200)]
rt2x00: Reduce tx status reading timeout

When no TX status was available, the default timeout
of 20ms is a bit high. The frame is highly likely already
send out, so the TX status should be available within
only a few milliseconds.

Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agomac80211: consolidate MIC failure report handling
Christian Lamparter [Sat, 30 Apr 2011 13:24:30 +0000 (15:24 +0200)]
mac80211: consolidate MIC failure report handling

Currently, mac80211 handles MIC failures differently
depending on whenever they are detected by the stack's
own software crypto or when are handed down from the
driver.

This patch tries to unify both by moving the special
branch out of mac80211 rx hotpath and into into the
software crypto part. This has the advantage that we
can run a few more sanity checks on the data and verify
if the key type was TKIP. This is very handy because
several devices generate false postive MIC failure
reports. Like carl9170, ath9k and wl12xx:
<http://www.spinics.net/lists/linux-wireless/msg68494.html>
"mac80211: report MIC failure for truncated packets in AP mode"

Cc: Luciano Coelho <coelho@ti.com>
Cc: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agossb: cc: prepare clockmode support for cores rev 10+
Rafał Miłecki [Wed, 27 Apr 2011 15:40:11 +0000 (17:40 +0200)]
ssb: cc: prepare clockmode support for cores rev 10+

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoRevert "wl12xx: support FW TX inactivity triggers"
Luciano Coelho [Mon, 2 May 2011 09:37:33 +0000 (12:37 +0300)]
Revert "wl12xx: support FW TX inactivity triggers"

This reverts commit 47684808fd89d6809c0886e06f8ac324252499d8.

Conflicts:

drivers/net/wireless/wl12xx/conf.h
drivers/net/wireless/wl12xx/main.c

13 years agowl12xx: export driver state to debugfs
Arik Nemtsov [Mon, 18 Apr 2011 11:15:29 +0000 (14:15 +0300)]
wl12xx: export driver state to debugfs

By reading the "driver_state" debugfs value we get all the important
state information from the wl12xx driver. This helps testing and
debugging, particularly in situations where the driver seems "stuck".

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
13 years agowl12xx: fix race condition during recovery in AP mode
Arik Nemtsov [Mon, 18 Apr 2011 11:15:28 +0000 (14:15 +0300)]
wl12xx: fix race condition during recovery in AP mode

When operating as AP, the TX queues are not stopped when we start
recovery. mac80211 is notified only after the fact. When there is
pending TX, it will be queued even after the FW is down. This leads to
situations where the TX queues are stopped (because of the TX-watermark
mechanism), and are never woken up when we return from recovery.

Fix this by explicitly stopping the TX queues when before initiating
recovery.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
13 years agowl12xx: add debugfs entry for starting recovery
Arik Nemtsov [Mon, 18 Apr 2011 11:15:27 +0000 (14:15 +0300)]
wl12xx: add debugfs entry for starting recovery

This entry is useful for debugging the driver state machine during
recovery.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
13 years agowl12xx: AP-mode - reconfigure templates after basic rates change
Arik Nemtsov [Mon, 18 Apr 2011 11:15:26 +0000 (14:15 +0300)]
wl12xx: AP-mode - reconfigure templates after basic rates change

When there's a change in the basic rates of the AP, reconfigure relevant
templates with the new rates.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
13 years agowl12xx: AP-mode - overhaul rate policy configuration
Arik Nemtsov [Mon, 18 Apr 2011 11:15:25 +0000 (14:15 +0300)]
wl12xx: AP-mode - overhaul rate policy configuration

Use the minimal rate configured in the basic rates set as the AP
broadcast and multicast rate. The minimal rate is used to ensure weak
links can still communicate.

When the basic rates contains at least one OFDM rate, configure all
unicast TX rates to OFDM only.

Unify rate configuration on initialization and on change notification
into a single function.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
13 years agowl12xx: print firmware program counter during recovery
Arik Nemtsov [Mon, 18 Apr 2011 11:15:24 +0000 (14:15 +0300)]
wl12xx: print firmware program counter during recovery

When performing recovery, print the firmware version and program
counter (by reading the SCR_PAD4 register). The value of the firmware
program counter during assert can be useful for debugging.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
13 years agowl12xx: schedule recovery on command timeout
Arik Nemtsov [Mon, 18 Apr 2011 11:15:23 +0000 (14:15 +0300)]
wl12xx: schedule recovery on command timeout

We use a long timeout (2 seconds) when sending commands to the FW.
When a command times out, it means the FW is stuck, and we should
commence recovery.

This should make recovery times shorter as we'll recover on the first
timeout indication.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
13 years agowl12xx: AP-mode - disable beacon filtering on start up
Arik Nemtsov [Mon, 18 Apr 2011 11:15:22 +0000 (14:15 +0300)]
wl12xx: AP-mode - disable beacon filtering on start up

New AP-mode FWs filter external beacons by default. Disable this
filtering on start up so we can properly configure ERP protection.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>