ath9k: reset noiseimmunity level to default

After the chip reset, the noise immunity levels are restored with
history values. If the immunity levels are lower than the defaults,
lets start with the optimal values.

Signed-off-by: Rajkumar Manoharan <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

cfg80211: increse bss expire time

The background scan completion takes more time when the station is
having heavy uplink traffic. The scan state machine decides to fall
back to home channel on every off-channel visit when there are pending
frames in tx queue. bgscan completion took ~30sec on dual band US
regulatory card.

scan period = (20 active channels * probe timeout) +
              (12 passive channels * passive probe timeout) +
              (32 * timeout on home channel) +
              (32 * flush timeout)

Signed-off-by: Rajkumar Manoharan <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k: fix ibss beacon next tbtt

Sync-up ibss beacon timer with the beacon frame's timestamp. When the
node acts as joiner, it has to sync with the received beacon timestamp
instead of reading tsf from hw. As the hw tsf wont wont be update till
bssid is configured. This patch programs hw tsf with the received beacon
timestamp if beacon timers are yet to be configured.

Signed-off-by: Rajkumar Manoharan <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k: fix ibss fair beacon distribution for AR9462

Update AR9462 initval to fix unbalance beacon distribution
in Ad-Hoc network.

Signed-off-by: Rajkumar Manoharan <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

hostap: GFP_ATOMIC/GFP_KERNEL cleanup

The driver is allocating memory during initialization with GFP_ATOMIC
even though GFP_KERNEL is sufficient.  This patch fixes it.

Signed-off-by: Matt Renzelmann <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: Take into account TSF adjustment latency in Toffset setpoint

When testing mesh synchronization we observed a global TSF slowdown that
was dependent on the number of synchronized mesh stations.  This seems
to be caused by the TSF adjustment (read/write) latency.

Adding a small margin to the Toffset setpoint solved the problem.

Signed-off-by: Shinichi Hotori <[email protected]>
Signed-off-by: Yu Niiro <[email protected]>
Signed-off-by: Javier Cardona <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: Choose a new toffset setpoint if a big tsf jump is detected.

Signed-off-by: Javier Cardona <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211_hwsim: fixup for tsf setting

Last patch I sent failed to take into account the offset of each phy.

Signed-off-by: Javier Cardona <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: Set the correct values for hwmp (1) and airtimeLinkMetric (1)

Per sections 8.4.2.100.2 and 8.4.2.100.3 of Std 802.11-2012

Reported-by: Shinichi Hotori <[email protected]>
Signed-off-by: Javier Cardona <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k: optimize the hardware hang check

Since it's only called when beacons are stuck, move it to the SWBA handler
tasklet, to avoid doing redundant checks on every single interrupt.

Signed-off-by: Felix Fietkau <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k: do not register LEDs on AR913x

LED support is typically handled via system GPIO on these platforms.

Signed-off-by: Felix Fietkau <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k_hw: add support for 8 AP mode interfaces

Also tweak beacon response times for better stability with the shorter
timer intervals.

Signed-off-by: Felix Fietkau <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath5k: use compare_ether_addr on MAC addresses instead of memcmp

Following Felix's recent patchset as an example I have replaced memcmp with
compare_ether_addr.

"Because of the constant size and guaranteed 16 bit alignment, the inline
compare_ether_addr function is much cheaper than calling memcmp."

Signed-off-by: Jonathan Bither <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k: Gather and report IRQ sync_cause errors.

Report all defined sync_cause errors in debugfs
to aid with debugging.

Use a macro to print out the interrupts file contents
to decrease code duplication.

Signed-off-by: Ben Greear <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k: Use macro to decrease code when priting recv stats.

This hides some repetitive code, and will help if the
column widths ever need to change.

Signed-off-by: Ben Greear <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: check IEEE80211_HW_QUEUE_CONTROL in ieee80211_check_queues()

Commit 3a25a8c8 (mac80211: add improved HW queue control) introduced a
bug when running in AP mode without the IEEE80211_HW_QUEUE_CONTROL
flag set.  The ieee80211_check_queues() function always returns
-EINVAL, preventing AP mode from starting.  To fix this, check whether
this flag is set before checking if cab_queue is set properly.

Signed-off-by: Luciano Coelho <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

cfg80211: enforce lack of interface combinations

My grand plan to allow drivers to gradually move over
to advertising virtual interface combinations and only
enforce with drivers that do want it enforced doesn't
seem to be working out, only Christian ever added the
advertising (to carl9170), nobody else did.

Begin enforcing combinations in cfg80211 so that users
can rely on the information reported about a device.

Cc: "Luis R. Rodriguez" <[email protected]>
Cc: Jouni Malinen <[email protected]>
Cc: Vasanthakumar Thiagarajan <[email protected]>
Cc: Senthil Balasubramanian <[email protected]>
Cc: Kalle Valo <[email protected]>
Cc: Jiri Slaby <[email protected]>
Cc: Nick Kossifidis <[email protected]>
Cc: Bob Copeland <[email protected]>
Cc: Bing Zhao <[email protected]>
Cc: Lennert Buytenhek <[email protected]>
Cc: Ivo van Doorn <[email protected]>
Cc: Gertjan van Wingerde <[email protected]>
Cc: Helmut Schaa <[email protected]>
Cc: Luciano Coelho <[email protected]>
Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

Merge branch 'wireless-next' of git://git./linux/kernel/git/iwlwifi/iwlwifi

brcm80211: smac: only provide valid regulatory hint

The driver provides a regulatory hint to cfg80211 as obtained from the
SPROM. Mostly, this will be a two-letter ISO country code. However, it
may obtain special country code similar to the world regulatory domain
as used in cfg80211. This patch avoids setting these special codes as
the hint is lost to cfg80211.

Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Reviewed-by: Brett Rudley <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

brcm80211: smac: do not use US as fallback regulatory hint

The brcmsmac driver provides the country code from sprom as a regulatory
hint to cfg80211. When brcmsmac does not find a country code entry in
the sprom it passes 'US' as regulatory hint. Better approach is to rely
on the world regulatory domain in cfg80211/crda.

Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Reviewed-by: Brett Rudley <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

brcm80211: fmac: clean up chip id table

Remove unsupported chip ID and rearrange the list in alphabetical
order

Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Reviewed-by: Arend van Spriel <[email protected]>
Signed-off-by: Franky Lin <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

brcm80211: fmac: postpone dongle RF enabling.

BRCMF_C_UP is the command that asks the firmware to enable RF of
dongle. Some firmware initialization steps must be performed during
RF is down. Postpone BRCMF_C_UP firing until brcmf_netdev_open get
called to ensure firmware have enough time to finish
initialization.

Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Reviewed-by: Arend van Spriel <[email protected]>
Signed-off-by: Franky Lin <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

brcm80211: fmac: add frame header extension support

A header extension is introduced in the received frame to provide
extra space for dongle information. This won't affect current
supported chipset since the data_offset is 0. But it's necessary for
adding support for future chipset.

Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Reviewed-by: Arend van Spriel <[email protected]>
Signed-off-by: Franky Lin <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

brcm80211: fmac: register primary net device with device mac address

The primary net device was registered with a primary mac address and
upon IFUP it was set to match the actual mac address from the device.
This patch changes that and moves the brcmf_add_if() call to the common
part of the driver.

Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Reviewed-by: Franky (Zhenhui) Lin <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

brcm80211: fmac: remove primary mac address handling from brcmf_net_attach

The mac address for the primary interface was handled different from
the other interfaces. The code has been restructured such that the
brcmf_net_attach function treats the interfaces equal.

Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Reviewed-by: Franky (Zhenhui) Lin <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

brcm80211: fmac: make brcmf_net_attach() static

The function brcmf_net_attach() is only used within dhd_linux.c so
it does not need to be extern. This patch makes the function static.

Reviewed-by: Pieter-Paul Giesberts <[email protected]>
Reviewed-by: Franky (Zhenhui) Lin <[email protected]>
Signed-off-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: remove ieee80211_rx_bss_get

its not used where, while we directly obtain ieee80211_bss's
pointer in ibss.c by calling cfg80211_get_bss

Signed-off-by: Mohammed Shafi Shajakhan <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k_htc: Add Panasonic N5HBZ0000055 device id

Reported-by: Ryan Roper <[email protected]>
Signed-off-by: Sujith Manoharan <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: remove hw.conf.channel usage where possible

Removes hw.conf.channel usage from the following functions:
 * ieee80211_mandatory_rates
 * ieee80211_sta_get_rates
 * ieee80211_frame_duration
 * ieee80211_rts_duration
 * ieee80211_ctstoself_duration

This is in preparation for multi-channel operation.

Signed-off-by: Michal Kazior <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k_hw: Remove BTCOEX initvals

The MAX_TXPWR table for BTCOEX is not needed for AR9462.
Programming these values to the HW results in undesirable
behavior - for example, large number of delimiter/data
underruns are seen in AES-CCMP mode. Also, registers like
AR_PCU_MISC_MODE2 return 0xdeadbeef after the BTCOEX_MAX
power table is programmed to the HW, and frames being transmitted
end up being looped back to the RX engine, an example being beacon
frames in IBSS mode.

Remove this table for now - this fixes CCMP performance and general
IBSS usage.

Signed-off-by: Sujith Manoharan <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: fix an issue in ieee80211_tx_info count field management

I noticed a possible issue in the status count field management of the
ieee80211_tx_info data structure. In particular, when the AGGR
processing is employed,
status.rates[].count is set just for the first frame and not for
others belonging to the same burst, leading to wrong statistic data in
the mac80211 debug file system.

Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

cfg80211: Add channel information to NL80211_CMD_GET_INTERFACE

If the current channel is known, add frequency and channel type to
NL80211_CMD_GET_INTERFACE.

Signed-off-by: Pontus Fuchs <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

rt2800: zero registers of unused TX rings

This is needed if we take over after drivers which use those.

Signed-off-by: Jakub Kicinski <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Reviewed-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

rt2800: initialize queues before giving up due to DMA error

Don't immediately abort .start if DMA is busy before we
initialize the queues. Some drivers do not deinitialize
queues properly and we would fail to take over after them.

This behaviour is consistent with legacy driver.

Signed-off-by: Jakub Kicinski <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Reviewed-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

rt2800: add disabling of DMA before loading firmware

Legacy driver disables DMA before loading firmware.

Signed-off-by: Jakub Kicinski <[email protected]>
Reviewed-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

rt2800: introduce wpdma_disable function

Introduce wpdma_disable function to simplify the code.

Signed-off-by: Jakub Kicinski <[email protected]>
Reviewed-by: Stanislaw Gruszka <[email protected]>
Acked-by: Gertjan van Wingerde <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: protect ->scanning by mutex in ieee80211_work_work()

Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: do not scan and monitor connection in parallel

Before we send probes in connection monitoring we check if scan is not
pending. But we do that check without locking. Fix that and also do not
start scan if connection monitoring is in progress.

Signed-off-by: Stanislaw Gruszka <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

Merge branch 'for-linville' of git://git./linux/kernel/git/luca/wl12xx

iwlwifi: change the default P2P support to "Yes"

Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: more generic name for bluetooth command

Instead of hardcode 6000 and 2000, use more generic name

Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move scan related declarations out of iwl-core.h

Move these as part of iwl-core.h cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move utility functions out of iwl-core.h

Move these functions to iwl-agn.h as part of
iwl-core.h cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_bcast_addr to iwl-agn-sta.c

Move this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_force_rf_reset to iwl-agn-rx.c

Move this function as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_send_statistics_request to iwl-agn.c

Move this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_print_rx_config_cmd to iwl-agn-rxon.c

Move this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_send_bt_config and mark it static

Move this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_set_tx_power and make it static

This function is used only in iwl-agn-rxon.c,
move it there and mark it static.

Move this function as part of iwl-core.c
cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwlagn_fw_error to iwl-agn.c

Move this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_chswitch_done to iwl-mac80211.c

Move this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_set_rate to iwl-agn-rxon.c

Move this as part of  iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_connection_init_rx_config to iwl-agn-rxon.c

Move this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_set_flags_for_band to iwl-agn-rxon.c

Moving this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_set_rxon_channel to iwl-agn-rxon.c

Moving this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_set_rxon_ht to iwl-agn-rxon.c

Moving this as part of iwl-core.c cleanup.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_is_ht40_tx_allowed out of iwl-core.c

This is really something determined by
station parameters, so move it to
iwl-agn-sta.c.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_set_hw_rfkill_state and mark it static

iwl_set_hw_rfkill_state is used only in
iwl-agn.c. Move it there and mark it static.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwl_free_skb and mark it static

iwl_free_skb is used only in iwl-agn.c, move
it there and mark it static.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move channel switch related functions

With the creation of iwl-agn-devices.c,
iwl-core.c can be cleaned up a bit more by
moving beacon time related functions from
iwl-core.c to iwl-agn-devices.c

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: calculate active legacy rates per station

Not all stations are guaranteed to have the same
active (available) legacy rates, so calculate them
on rate control init instead of hard-coding them
based on our own available rates. I have no idea
why that was done here before.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: make iwl_sta_fill_lq static

It's only used in a single place.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>

iwlwifi: move iwlagn_hw_valid_rtc_data_addr prototype

Since this function is now used only by
op_mode, remove it from iwl-shared.h and move
it to iwl-agn.h.

Signed-off-by: Meenakshi Venkataraman <[email protected]>
Signed-off-by-by: Wey-Yi Guy <[email protected]>

Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-next into wireless-next

iwlwifi: dynamically determine lib_ops

Having the pointer to lib_ops in the config
makes it impossible to split the driver into
different modules. Determine the ops based on
the device family enumeration to get rid of
the direct pointer.

Also move all the opmode specific code from
the iwl-[1256]000.c files into a new file
iwl-agn-devices.c so that the former only
have configuration data now.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

net/wireless: ipw2x00: add supported cipher suites to wiphy initialization

Driver doesn't report its supported cipher suites through cfg80211
interface. It still uses wext interface and probably will not work
through nl80211, but will at least correctly advertise supported
features.

Bug was reported by Omar Siam.
https://bugzilla.kernel.org/show_bug.cgi?id=43049

Signed-off-by: Stanislav Yakovlev <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

prism54: replace open-coded ARRAY_SIZE with macro

Signed-off-by: Jim Cromie <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

brcm80211: replace open-coded ARRAY_SIZE with the macro

Signed-off-by: Jim Cromie <[email protected]>
Acked-by: Arend van Spriel <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Fix LLCP link timeout typo

We were sending the LTO TLV as a version TLV instead of the actual link
timeout one.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Add MIUX to the local LLCP general bytes

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: pn533 Rx chaining support

When buffers on the receiption path exceed 262 bytes, the pn533 uses
a chaining mechanism where the initiator has to send NULL data frames
to fetch the remaining frames.
We do that from a workqueue context while holding the cmd lock. Once the
MI bit is gone, we aggregate the queued received skbs.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Convert pn533 from tasklet to workqueues

There is no need for soft IRQ contexts, and workqueues are more flexible.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Call llcp_add_header properly when sending LLCP DM or DISC

dsap and ssap were swapped when sending DN or DISC.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Fix LLCP TLV building routine

The if logic could lead to zero length TLVs.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: No need to apply twice the modulo op to LLCP's recv_n

recv_n is set properly when receiving an HDLC frame.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Dump LLCP frames

At KERN_DEBUG level.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Add Core support to generate tag lost event

Some HW/drivers get notifications when a tag moves out of the radio field.
This notification is now forwarded to user space through netlink.

Signed-off-by: Eric Lapuyade <[email protected]>
Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Changed target activated state logic

Signed-off-by: Eric Lapuyade <[email protected]>
Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Fix next target_idx type and rename for clarity

Signed-off-by: Eric Lapuyade <[email protected]>
Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: The core part should generate the target index

The target index can be used by userspace to uniquely identify a target
and thus should be kept unique, per NFC adapter. Moreover, some protocols
do not provide a logical index when discovering new targets, so we have to
generate one for them.
For NCI or pn533 to fetch their logical index, we added a logical_idx field
to the target structure.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Add HCI documentation

Signed-off-by: Eric Lapuyade <[email protected]>
Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: SHDLC implementation

Most NFC HCI chipsets actually use a simplified HDLC link layer to
carry HCI payloads.
This implementation registers itself as an HCI device on behalf of the
NFC driver.

Signed-off-by: Eric Lapuyade <[email protected]>
Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: HCI support

This is an implementation of ETSI TS 102 622 specification.
Many NFC chipsets use HCI as the host <-> target protocol on top of a
serial link like i2c.

Signed-off-by: Eric Lapuyade <[email protected]>
Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Export target lost function

NFC drivers will call this routine when they detect that a tag leaves the
RF field. This will eventually lead to the corresponding netlink event
to be sent.

Signed-off-by: Eric Lapuyade <[email protected]>
Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

NFC: Add a target lost netlink event

Some chips are capable of detecting when a tag is out of the field, so
they could send a netlink event about it to userspace.

Signed-off-by: Samuel Ortiz <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

ath9k: remove dead code

Clean up some orphaned code lines containing
* unused variables (not referenced / write-only)
* non-implemented function prototypes

Signed-off-by: Zefir Kurtisi <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mac80211: fix the assignment of PREQ's MAC address for Proactive RANN

Record the RANN sender's address only for RANNs that meet the acceptance
criteria (per sections 13.10.12.4.2).

Signed-off-by: Chun-Yeow Yeoh <[email protected]>
Reviewed-by: Javier Cardona <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mwifiex: use asynchronous firmware loading

Make use of request_firmware_nowait instead of request_firmware
to load FW asynchronously. This fixes timeouts introduced with
recent udev changes.

Signed-off-by: Amitkumar Karwar <[email protected]>
Signed-off-by: Kiran Divekar <[email protected]>
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mwifiex: add support for WPS2.0

This patches enables setting association request and probe request
IE for station interface. WPS exchange between WPS2.0 AP and mwifiex
STA Enrollee/External Registrar completes successfully.

Tested with wpa_supplicant 1.0 and 2.0 devel.

Signed-off-by: Avinash Patil <[email protected]>
Signed-off-by: Kiran Divekar <[email protected]>
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mwifiex: display correct country information in debugfs "info"

Use "priv->country_code" string to display country information in
debugfs command "info" instead of "adapter->region_code".

"adapter->region_code" contains default region code got from FW while
initialization, whereas "priv->country_code" is updated in reg_notifier
handler whenever there is a change in regulatory domain.

Signed-off-by: Amitkumar Karwar <[email protected]>
Signed-off-by: Kiran Divekar <[email protected]>
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mwifiex: set default regulatory domain

Driver gets region code from FW during initialisation. This patch
makes use of it for settting default regulatory domain using
regulatory_hint() API.

Signed-off-by: Amitkumar Karwar <[email protected]>
Signed-off-by: Kiran Divekar <[email protected]>
Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

mwifiex: fix typo in RSSI_HIGH event handling

This is a copy-n-paste error introduced in recent patch
"mwifiex: add set_cqm_rssi_config handler support".

Signed-off-by: Bing Zhao <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

iwlwifi: introduce device family enum

This will later be used to dynamically bind
the configuration data for DVM and MVM.

For now, we can use it to get rid of the
additional_nic_config() hook.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

iwlwifi: create device configuration header file

The iwl-shared.h header file will be going away
soon. There isn't much left in it that we keep,
other than the device configuration declarations.
Move those out now to a new iwl-config.h header.
iwl-cfg.h seemed like a possible alternative but
those declarations will later live in the PCIe
transport code.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

iwlwifi: move eeprom into priv

The whole code around eeprom is distributed
across whole bunch of different files, most
of which belong to the to-be-DVM code. As a
result, it is currently very hard to split
out the EEPROM code to be generic. However,
it is also quite unlikely that the current
EEPROM code will be needed by the MVM code
as that has different mechanisms to query
the EEPROM (it does so through the uCode.)

So, at least temporarily, move everything
into priv. If it becomes necessary to use
the code from MVM, we will have to split it
out, but then it's also easier since we'll
know what pieces we need.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

iwlwifi: split force_reset debugfs file

Split the force_reset debugfs file into two
different files:
 * "rf_reset" triggers a reset of the RF when
   written to and exposes statistics on RF
   resets when read
 * fw_restart triggers a firmware restart when
   written to and lives in the transport

This cleans up all sources of firmware restart
to originate within the transport layer and
allows us to simplify some code.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

iwlwifi: remove ack_check module parameter

This defaults to false, and we don't recommend
to use it anywhere, so just remove it.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

iwlwifi: move hw_params into priv

The hw_params are mostly values that are
derived from the actual hardware config.
As such, while it is possible that MVM
will require similar ones, it makes more
sense -- at least for now -- to put them
into the DVM struct.

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

iwlwifi: move queue watchdog into transport

This removes one of the two sources of device
restarts in the upper layer -- those are a bit
inconvenient because normal restarts originate
in the transport. By moving the watchdog down
it can be treated the same.

Also rewrite the watchdog logic. Timers are
much more efficient when they never fire, so
instead firing a timer every 500ms set up a
timer for each TX queue and fire it only when
the queue is really stuck. This avoids the CPU
waking up when everything is working well.

While at it, remove the wd_disable config item
and replace it by simply setting wd_timeout to
IWL_WATCHHDOG_DISABLED (0).

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>

iwlwifi: remove unneeded struct declarations

Signed-off-by: Johannes Berg <[email protected]>
Signed-off-by: Wey-Yi Guy <[email protected]>
Signed-off-by: John W. Linville <[email protected]>