ath6kl: Use mutex to protect dma buffer in sync read write

Firmware crashes while starting Soft AP in 32 bit x86 platform.
The reason is that the single dma buffer (ar_sdio->dma_buffer)
is used in ath6kl_sdio_read_write_sync() for unaligned buffer
handling and this function is called in the multiple context
at the same time. So, finally hits dma buffer corruption and
firmware crash.

Mutex is used to protect dma buffer to avoid data corruption.
Spin lock can not used to fix this issue since mmc stack
read/write calls may for sleep.

Observed this issue with recently commited patch
"ath6kl: Claim sdio function only at appropriate places"
861dd058f495973c7ad2a44b8f68f3cc05733eab

kvalo: change name of mutex to more descriptive and add a comment
about what it protects

Signed-off-by: Raja Mani <rmani@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: unlock if ath6kl_cfg80211_connect() fails

There is an unlock missing on this error path.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Enable multiple vif support

The maximum number of supported virtual interfaces are 3.

Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Find ar->max_norm_iface in firmware IE parsing

Currently the max number of vifs which can be used for non-p2p
mode is determined in ath6kl_core_alloc(). But the maximum
supported vifs are parsed from firmware IE in ath6kl_fetch_fw_api2()
which would happen after ath6kl_core_alloc().

Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Remove modparam multi_norm_if_support

This modparam was introduced to enable non-p2p mode operation
on two virtual interfaces. It does not seem to be necessary to
have a separate module parameter to do that. Instead, this option
can be enabled when any one of the interfaces is not going to be
used for p2p (ath6kl_p2p).

Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: indicate probe-resp offload support

The ath6kl responds to probe-requests in HW while operating as an AP. It
supports offloading exclusions to support the WPS, WPS2, P2P and
802.11u protocols.

Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

Merge remote branch 'wireless-next/master' into ath6kl-next

mwifiex: add support for Marvell sd8797 device

This patch supports Marvell chipset 88W8797 (Avastar) with
SDIO interface.

The corresponding firmware image file is located at:
"mrvl/sd8797_uapsta.bin"

Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Tristan Xu <xurf@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kiran Divekar <dkiran@marvell.com>
Signed-off-by: Frank Huang <frankh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

p54spi: Remove FIXME in op_stop

Don't use the interruptible variant of mutex_lock(). It doesn't really
need to be interruptible. This avoids nasty error handling.

Signed-off-by: Michael Buesch <m@bues.ch>
Acked-by: Christian Lamparter <chunkeey@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: call ieee80211_recalc_idle() after sending packets

Some drivers (e.g. ath9k) assume that it's safe to go into low-power mode
immediately after the idle state changes. To support that, mac80211 even
calls drv_flush() before that happens.
In some instances, mac80211 sent a packet right after recalculating the
idle state, this patch fixes that.

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

wireless: use strlcpy routine in .get_drvinfo

Convert various seemingly still compiled wireless drivers' .get_drvinfo
routines to use the preferred strlcpy() routine.

Signed-off-by: Rick Jones <rick.jones2@hp.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

minstrel_ht: Remove unused function parameters

Remove unused function parameters in the following functions:
minstrel_calc_rate_ewma()
minstrel_ht_calc_tp()
minstrel_aggr_check()
minstrel_ht_set_rate()

Signed-off-by: Patrick Kelle <patrick.kelle81@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: memory leak in mesh_queue_preq()

We recently introduced a return here, but we need to call kfree
first.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: use kfree_skb() instead of kfree()

sk_buff structs should be freed using kfree_skb().

This was introduced recently in 029458212 "mac80211: Save probe
response data for bss".

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: Get rid of search loop for rate group index

Finding the group index for a specific rate is done by looping through
all groups and returning if the correct one is found. This code is
called for each tx'ed frame and thus it makes sense to reduce its
runtime.

Do this by calculating the group index by this formula based on the SGI
and HT40 flags as well as the stream number:

idx = (HT40 * 2 * MINSTREL_MAX_STREAMS) +
      (SGI * MINSTREL_MAX_STREAMS) +
      (streams - 1)

Hence, the groups are ordered by th HT40 flag first, then by the SGI
flag and afterwards by the number of used streams.

This should reduce the runtime of minstrel_ht_get_group_idx
considerable.

Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: Check rate->idx before rate->count

The drivers are not required to fill in rate->count if rate->idx is set
to -1. Hence, we should first check rate->idx before accessing
rate->count.

Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: remove crypto special case for auth frames

The shared key authentication frame that needs to be
encrypted (the third one in the shared key handshake)
is directly encrypted in ieee80211_send_auth and the
IEEE80211_TX_INTFL_DONT_ENCRYPT is set. All others
are not encrypted, so the only way to get to this is
erroneously on no-monitor AP side.

Remove the special case for authentication frames to
fix the AP shared key side when operating without
cooked monitor interfaces -- with cooked monitor the
IEEE80211_TX_INTFL_DONT_ENCRYPT also gets set, so we
never get here -- an AP never encrypts auth frames.

Without this patch, an AP operating in WEP mode with
my no-monitor patches would erroneously encrypt all
authentication frames, instead of none.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmsmac: Use current logging styles

Add and use pr_fmt and pr_<level>
Remove useless double parentheses from macros.
Remove function names from format strings, add to pr_debug use.
Coalesce formats.
Remove uncompileable undeclared variable in a DMA_NONE use.

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

NFC: Fix indentation in nci.h file

Fix indentation in nci.h file.

Signed-off-by: Ilan Elias <ilane@ti.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k_hw: set btcoex weights for AR9462

Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k_hw: Cleanup btcoex wlan weights

Remove all wlan weight macros and group it together for better
understanding & readability. It makes the code reusable for
AR9462 wlan weights.

Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: Remove enabling btcoex from stomp type change

This patch removes btcoex_enable from stomp type change and let
it be called from callee functions that makes the code can be
reusable for MCI changes.

Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Merge branch 'for-linville' of git://github.com/kvalo/ath6kl

ath6kl: Fix AP mode operation after interface down/up

When operating AP interface is brough down the mode is reset to
STA. This STA will be reconfigured into AP mode when the interface
is brought up again. This sequence will be successful only when
change_virtual_intf() returns with no error, but there is a
check in this callback which does the type change only when
that interface is active. This callback does nothing more
than saving the new interface type to vif and wdev, so the
sanity check for interface state and wmi state is not necessary.
This makes the AP interface functional again after interface down/up.

Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Clear WPS ctrl flag if zero length IE is received from cfg80211

Connect control flag CONNECT_WPS_FLAG has to be cleared
by default even if the driver receives zero length IE
from cfg80211.

Otherwise this flag would be always set after WPS exchange
which would lead wpa_supplicant to fail to connect with
the received WPS credentials. This issue is observed only
in OPEN security.

kvalo: use cfg80211 instead of CFG in the commit log

Signed-off-by: Raja Mani <rmani@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: add firmware IE for maximum number of vifs

Not all firmwares support multiple vifs and we need to read the limit from
the firmware image.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: make maximum number of vifs runtime configurable

Needed when detecting how many vifs firmware supports.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: add firmware filename info to struct ath6kl_hw

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: add ar6004 firmwares to sdio module

When adding ar6004 SDIO support I forgot to add corresponding
MODULE_FIRMWARE() definitions to sdio.c.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: use hardware version names consistently

Part of ath6kl uses "REV3" style of naming hardware versions and elsewhere
"hw 2.1.1" is used instead for the same version. This is confusing, use
the latter term everywhere.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: add name field to struct ath6kl_hw

To make it easier to print name for each hardware type. Also move the hw
info print to ath6kl_init_hw_start() which is more logical place for it.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: add firmware IE for board data address

Board data address can change between firmwares so we need to read that
from the firmware image.

Also fix debug log for the patch address to print the address in hex.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: add board address to struct ath6kl_hw

This is to make it configurable by firmware IEs. Also determine if we need
to write or read the board address to the chip by checking if board address
is set or not.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: move hw version related to parameters to struct

It's easier to handle the values when they are defined in a struct.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: remove hw version related parameter defines

Having separate defines, in a different file, makes it difficult to read
the actual values. As we are just setting named fields in a struct
the defines don't make any sense anymore.

There are no functional changes, only moving of constants.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Fix bug in setting dot11_auth_mode in AP mode

OPEN_AUTH is passed as dot11_auth_mode by default, this would
affect the AP mode when configured with shared authentication
type. Assign appropriate auth type to fix this from driver.
A patch in wpa_supplicant (wpa_supplicant: Set configured auth_algs)
is also needed to fix this.

Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Fix bug in setting default key index for tx in AP mode

vif->def_txkey_index is set to key_index in ath6kl_cfg80211_add_key().
If the interface is configured with multiple static wep keys,
vif->def_txkey_index would be holding the index of the last key configured,
not the default tx key index. Remove this unnecessary default key index setting
in ath6kl_cfg80211_add_key() to configure the right key index in WEP thereby
make it work when multiple wep keys are configured.

Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: enable USB support

Now two modules are built, ath6kl_sdio.ko and ath6kl_usb.ko. But the USB
module isn't fully functional yet as HTC layer is missing support and
that's why it's marked as experimental for now.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: disable HTC for USB devices

As HTC layer doesn't support USB devices return an error if that happens.
USB support will be added to HTC in the future, this is just a temporary
solution.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: add USB support

Add USB support for ar6004. Currently only firmware can be booted,
no commands can be sent to firmware yet as HTC layer doesn't work
with USB yet.

Based on patches by Kevin Fang.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: add hif_type

In some rare cases core code needs to know what hif type is used. Add
a field to struct ath6kl to denote that. Hopefully this is just a
temporary solution.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: AR6004 SDIO support

Add support for AR6004 SDIO. Tested scan, association (open mode) and
ping.

kvalo: change commit log a bit, drop board address changes

Signed-off-by: Naveen Gangadharan <ngangadh@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: firmware boot fixes for ar6004

These have changed a bit since last time ar6004 code was commited.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: update ar6004 definitions

Add also hw 1.1.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: move diag commands to hif driver

This is preparation for USB support which will have different diag
commands.

Based on code by Kevin Fang.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: change bmi sizes being configurable by HIF

SDIO and USB have different maximum sizes for BMI commands so make that
configurable.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: move bmi calls to hif driver

In preparation for USB support which has it's own method for bmi.

Based on code by Kevin Fang.

Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Fix target minimum length requirement for WMI_SEND_PROBE_RESPONSE_CMDID

The firmware expects the minimum length of
WMI_SEND_PROBE_RESPONSE_CMDID to be 13. However, when the device is a
P2P GO and it needs to send a probe response to a non-P2P client,
there are no P2P IEs to be added, and therefore the length of the WMI
command is 12. This command gets rejected by the firmware. To fix
this, add an extra byte to satisfy the minimum length requirement.

Signed-off-by: Aarthi Thiruvengadam <athiruve@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Fix cfg80211 warning while starting IBSS mode

When the connect event is received from the target in IBSS mode,
cfg80211_ibss_joined() is called without informing BSS info to
cfg80211 layer which internally hits the below WARN_ON message.

WARNING: at net/wireless/ibss.c:33 __cfg80211_ibss_joined+0x153/0x180 [cfg80211]()
[..]
[ 4916.845878] Call Trace:
[ 4916.845889]  [<c10427d2>] warn_slowpath_common+0x72/0xa0
[ 4916.845905]  [<f8bccc63>] ? __cfg80211_ibss_joined+0x153/0x180 [cfg80211]
[ 4916.845918]  [<f8bccc63>] ? __cfg80211_ibss_joined+0x153/0x180 [cfg80211]
[ 4916.845923]  [<c1042822>] warn_slowpath_null+0x22/0x30
[ 4916.845934]  [<f8bccc63>] __cfg80211_ibss_joined+0x153/0x180 [cfg80211]
[ 4916.845941]  [<c1025108>] ? default_spin_lock_flags+0x8/0x10
[ 4916.845952]  [<f8bb7fcd>] cfg80211_process_rdev_events+0x19d/0x220 [cfg80211]
[ 4916.845962]  [<f8bb669b>] cfg80211_event_work+0x2b/0x50 [cfg80211]
[ 4916.845968]  [<c105aae6>] process_one_work+0x116/0x3c0
[ 4916.845977]  [<f8bb6670>] ? cfg80211_get_dev_from_info+0x40/0x40 [cfg80211]
[ 4916.845982]  [<c105cdf0>] worker_thread+0x140/0x3b0
[ 4916.845986]  [<c105ccb0>] ? manage_workers+0x1f0/0x1f0
[ 4916.845991]  [<c1060c64>] kthread+0x74/0x80
[ 4916.845995]  [<c1060bf0>] ? kthread_worker_fn+0x160/0x160
[ 4916.846001]  [<c14e7bbe>] kernel_thread_helper+0x6/0x10
[ 4916.846005] ---[ end trace 769254924e409367 ]---

This patch make sures that BSS info is delivered via cfg80211_inform_bss()
to cfg80211 in advance before intimating IBSS status to cfg80211.

In addition to this, one debug message is also added to know
ad-hoc mode status (creator/joiner).

kvalo: change subject

Signed-off-by: Raja Mani <rmani@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Fix packet drop when ath6kl_cookie runs out

"ath6kl: Maintain virtual interface in a list" mistakenly stops the
netq only when the mode is ibss. This causes packet drops in sta mode
when the available cookies (buffer abstraction in ath6kl and also used
for tx throttling) runs out for the highest priority traffic. This patch
just fixes this regression though the original code may still need fixes
which can be addressed in separate patches.

Reported-by: Kalle Valo <kvalo@qca.qualcomm.com>
Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Fix accessing wrong skb->data in ath6kl_tx_complete()

When buffer alignmnet is applied, the data pointer of skb taken from
cookie will no longer point to the first byte of the actual data.
But the skb->data pointer is used in ath6kl_tx_complete() to get
the index of the virtual interface which will not give the correct
interface index and sometimes may give the following WARN_ON() message.
Use packet->buf instead of skb->data to fix this.

WARNING: at drivers/net/wireless/ath/ath6kl/wmi.c:88 ath6kl_get_vif_by_index+0x5b/0x60 [ath6kl]()
Hardware name: 2842K3U
Modules linked in: ath6kl mmc_block cfg80211 binfmt_misc ppdev nfs nfsd lockd nfs_acl auth_rpcgss sunrpc exportfs snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_intel
+snd_hda_codec snd_hwdep snd_pcm_oss snd_mixer_oss snd_pcm snd_seq_dummy thinkpad_acpi snd_seq_oss snd_seq_midi snd_rawmidi joydev fbcon tileblit font bitblit softcursor
+snd_seq_midi_event snd_seq snd_timer snd_seq_device i915 uvcvideo drm_kms_helper drm psmouse serio_raw snd i2c_algo_bit sdhci_pci videodev intel_agp soundcore intel_gtt jmb38x_ms
+memstick sdhci snd_page_alloc nvram lp parport agpgart video ahci r8169 mii libahci [last unloaded: ath6kl]
Pid: 15482, comm: kworker/u:1 Tainted: G        W   3.1.0-rc10-wl+ #2
Call Trace:
 [<c0144d72>] warn_slowpath_common+0x72/0xa0
 [<fb7c94fb>] ? ath6kl_get_vif_by_index+0x5b/0x60 [ath6kl]
 [<fb7c94fb>] ? ath6kl_get_vif_by_index+0x5b/0x60 [ath6kl]
 [<c0144dc2>] warn_slowpath_null+0x22/0x30
 [<fb7c94fb>] ath6kl_get_vif_by_index+0x5b/0x60 [ath6kl]
 [<fb7c7028>] ath6kl_tx_complete+0x128/0x4d0 [ath6kl]
 [<c04df920>] ? mmc_request_done+0x80/0x80
 [<fb7b9e2e>] htc_tx_complete+0x5e/0x70 [ath6kl]
 [<c05e4cf6>] ? _raw_spin_unlock_bh+0x16/0x20
 [<fb7ce588>] ? ath6kl_sdio_scatter_req_add+0x48/0x60 [ath6kl]
 [<fb7b9f42>] htc_async_tx_scat_complete+0xb2/0x120 [ath6kl]
 [<fb7ce9e7>] ath6kl_sdio_scat_rw+0x87/0x370 [ath6kl]
 [<c0101e12>] ? __switch_to+0xd2/0x190
 [<c01397b5>] ? finish_task_switch+0x45/0xd0
 [<c05e272e>] ? __schedule+0x3ae/0x8b0
 [<fb7cf00a>] ath6kl_sdio_write_async_work+0x4a/0xf0 [ath6kl]
 [<c015d266>] process_one_work+0x116/0x3c0
 [<fb7cefc0>] ? ath6kl_sdio_read_write_sync+0xb0/0xb0 [ath6kl]
 [<c015f5b0>] worker_thread+0x140/0x3b0
 [<c015f470>] ? manage_workers+0x1f0/0x1f0
 [<c0163424>] kthread+0x74/0x80
 [<c01633b0>] ? kthread_worker_fn+0x160/0x160
 [<c05ebdc6>] kernel_thread_helper+0x6/0x10

Reported-by: Aarthi Thiruvengadam <athiruve@qca.qualcomm.com>
Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Indicate WAPI IE from (Re)Association Request frame

This is needed to know whether the STA requests WAPI to be used and if
so, with what AKM and cipher.

Signed-off-by: Dai Shuibing <shuibing@qca.qualcomm.com>
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Allow SMS4 to be configured in AP mode

Signed-off-by: Dai Shuibing <shuibing@qca.qualcomm.com>
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

ath6kl: Add support for configuring SMS4 keys

Indicate support for WPI-SMS4 cipher and allow SMS4 keys to be
configured.

Signed-off-by: Dai Shuibing <shuibing@qca.qualcomm.com>
Signed-off-by: Jouni Malinen <jouni@qca.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>

mac80211: fix warning in ieee80211_probe_client

The warning is spurious -- if !sta we always exit without using the
unassigned qos variable, and if we do find the sta we assign it.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

libertas: release bss references and avoid warning from cfg80211_inform_bss

  CC [M]  drivers/net/wireless/libertas/cfg.o
drivers/net/wireless/libertas/cfg.c: In function ‘lbs_ret_scan’:
drivers/net/wireless/libertas/cfg.c:636:24: warning: ignoring return value of ‘cfg80211_inform_bss’, declared with attribute warn_unused_result
drivers/net/wireless/libertas/cfg.c: In function ‘lbs_join_post’:
drivers/net/wireless/libertas/cfg.c:1766:21: warning: ignoring return value of ‘cfg80211_inform_bss’, declared with attribute warn_unused_result

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

brcmsmac: fix warning in _initvars_srom_pci

  CC [M]  drivers/net/wireless/brcm80211/brcmsmac/srom.o
drivers/net/wireless/brcm80211/brcmsmac/srom.c: In function ‘_initvars_srom_pci’:
drivers/net/wireless/brcm80211/brcmsmac/srom.c:641:6: warning: ‘val’ may be used uninitialized in this function

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

brcm80211: fmac: release bss struct returned from cfg80211_inform_bss

Referenced struct returned by cfg80211_inform_bss must be released with
cfg80211_put_bss to avoid memory leak.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: fmac: remove PCMCIA core related code

all the devices supported or will be supported by brcmfmac
do not have a PCMCIA bus core. So remove the corresponding code.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: fmac: add function to free the glom skb queue

In several places in dhd_sdio.c a skb packet queue was being emptied
and the packets freed. This warrants to have a function in place to
do this.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Alwin Beukers <alwin@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: util: remove pointer traversal from brcmu_pkt_buf_free_skb

The function brcmu_pkt_buf_free_skb() was following the next pointer
to free all linked packets. However, it is only called with unlinked
packets so this can be removed.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Alwin Beukers <alwin@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: util: move brcmu_pkttotlen() function to brcmfmac

The functions brcmu_pkttotlen() is only used in brcmfmac driver
so it has been moved there. It also does not use the sk_buff
next pointer anymore but walks a skb queue to determine the total
length.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Alwin Beukers <alwin@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: smac: combine promiscuous mode functionality

Combined mac configuration for promiscious mode and monitor mode, and
removed unused monitor mode flag in pub structure.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Alwin Beukers <alwin@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: util: use sk_buff_head in precedence queue functions

Instead of dealing with sk_buff prev pointers the queue functions
now make use of the sk_buff_head functions provided by the kernel.

Reported-by: Johannes Berg <johannes@sipsolutions.net>
Reviewed-by: Alwin Beukers <alwin@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: smac: remove usage of brcmu_pkttotlen

The function brcmu_pkttotlen calculates the total length of a sk_buff
chain following the next pointer. In brcmsmac this is not needed as
in each place where it was used the provided sk_buff had a NULL pointer
as next field value.

Reviewed-by: Alwin Beukers <alwin@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: fmac: remove id retrieve code

sdio_chip.c is dedicated to sdio bus chip. No need to retrieve id
for bus core

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: fmac: add resetcore function for bcm4330 chip

This patch is part of the series of adding new backplane support

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: fmac: add coredisable function for bcm4330 chip

This patch is part of the series of adding new backplane support

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: fmac: add corerev function for bcm4330 chip

This patch is part of the series adding new backplane support

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: fmac: add iscoreup function for bcm4330 chip

New type of backplane interconnect support is needed
for bcm4330

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: smac: don't modify sta parameters when adding sta

Reported-by: Johannes Berg <johannes@sipsolutions.net>
Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Alwin Beukers <alwin@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: smac: remove duplicate definition of D11_PHY_HDR_LEN

The macro definition D11_PHY_HDR_LEN was defined in d11.h and pub.h.
It has been removed from pub.h.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Alwin Beukers <alwin@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: smac: replace own access category definitions with mac80211 enum

The brcmsmac had own definitions for the access categories. The mac80211
header provides these as well as they are used in the conf_tx callback.
As the definitions did not match the driver configured the tx parameters
to the wrong queue.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Alwin Beukers <alwin@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: smac: remove code under unused macro definitions

There were a couple of code segments left that were placed under
a macro definition that was not applicable or not used in the
brcmsmac driver. These pieces of code have been removed.

Reviewed-by: Pieter-Paul Giesberts <pieterpg@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Reviewed-by: Alwin Beukers <alwin@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: smac: fix endianess issue for OTP memory access

This fixes issue when using OTP memory. It was introduced by
following commit:

commit 028f78d43d80dcb8b1142ea38606067151dd3d51
Author: Arend van Spriel <arend@broadcom.com>
Date:   Fri Oct 21 16:16:35 2011 +0200

    brcm80211: smac: change buffer endianess convert function interface

Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: move nvm_device_type from iwl_priv to iwl_trans

The nvm_device_type is eeprom related and does not need to be part
of the iwl_priv structure.  Move it and eliminate access to the iwl_priv
structure.

Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: Remove dependence of iwl_priv from eeprom routines.

Make the eeprom routines less dependent on the iwl_priv structure.
Don't use the priv when bus structure is sufficient.

Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: fix rate-scaling algorithm for BT combo devices

iwlwifi tries to avoid using antenna B in BT combo devices when BT
is active. A bug in the rate-scaling algorithm was causing the combo
device to never attempt MIMO rates. Fix the algorithm to
opportunistically try MIMO rates when BT traffic is low.

Signed-off-by: Meenakshi Venkataraman <meenakshi.venkataraman@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: add debug information on queue stop / wake

Users complain that the traffic gets stalled sometimes. This will
allow easier debugging.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: remove the use of the QOS debug flag

This bit was used only once. Use IWL_DL_INFO instead.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: don't always split remain-on-channel

Due to the P2P context always having the associated
bit set when we got to this code, we were splitting
up the remain-on-channel durations unconditionally.
This isn't needed -- if the P2P context is in P2P
device type mode it doesn't impose restrictions on
timing to skip it in that case.

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>

iwlagn: convert remain-on-channel duration to TU

The device expects TU but we get ms, so convert.

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>

iwlwifi: fix unused label in iwl_send_cmd_sync

Warning introduced by c847474b7dfdda304d0d8ffcc5a9db546b1cb3e9

iwlwifi: check status before send command

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: check status before send command

Check the status before sending host command, if any of the condition
match, cancel the host command before queue

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: set "echo" host command length

"echo" host command has no data, set the length to 0

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: Move the core suspend function to iwl-agn-lib

The core suspend function is part of agn, iwl_mac80211 should only
handle mac80211 I/F operations.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: move station functions to mac80211

The station related mac80211 callback functions should belong to _mac80211

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: move hw_scan into _mac80211 file

iwlagn_mac_hw_scan should belong to _mac80211 callback.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: move more mac80211 callback function

Move more mac80211 related functions to _mac80211 file

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: explicitly program P2P QoS parameters

In P2P device mode, the device needs to have valid
QoS parameters. We currently have those because we
program parameters from any virtual interface into
all contexts, but not only do we want to get rid of
this -- it is also unpredictable since on the BSS
context we might have any parameters, and there it
might even be programmed for HT.

Explicitly program default QoS parameters into the
PAN context for P2P (the defaults are 11g but with
QoS disabled) to make device behaviour predictable.

This also helps when in a follow-up patch we will
use TX QoS parameters from mac80211 only for the
context they were meant for -- without this first
that would completely break P2P device discovery.

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>

iwlagn: use per-vif AC parameters

Eliad added the ability to configure AC parameters
per virtual interface; make use of this in iwlwifi
and set the parameters in the right context. Since
storage and uploading to the device is already per
context, this is sufficient.

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>

iwlwifi: move all mac80211 related functions to one place

There are many mac80211 callback functions in the driver, as current
implementation, those functions are scatter everywhere which is very difficult
to find and maintain, move all the mac80211 callback functions into single
file. There should be no functional changes

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: move ucode files out of the iwl_priv structure

Relocate the ucode files and update relevant code.
More code refactoring.

Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: push knowledge of ucode image lower down

Move the knowledge of the ucode image to lower level routines.  Also do
not pass the iwl_priv structure lower than it needs to be.

Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: remove knowledge of ucode image location from upper layers

The upper layers of the driver do not need to know where the ucode is
stored.  It already passes in an enum of which ucode to load.  Let the
lower layer routines select the ucode to load.

Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: move ucode_write_complete from priv to trans structure

ucode_write_complete is used for ucode loading.  Move it as part of
restructuring work out of the priv structure.

Signed-off-by: Don Fry <donald.h.fry@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: remove un-supported SKUs

BG only SKUs are no longer supported by 2000 and 1x5 series. Remove it

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlagn: fix NULL ptr deref when reprogramming sta w/o LQ

Reinette reports a crash in iwl_reprogram_ap_sta(). The
debugging shows:

  b1 16    mov    $0x16,%cl
 *f3 a5    rep movsl %ds     <-- trapping instruction:(%rsi),%es:(%rdi)

which is a memcpy of 22 (0x16) words (movsl). this points
to "priv->stations[sta_id].lq" being NULL since that is
the memcpy() of that size here.

The only way I see for this to happen is if we try to
do some RXON reprogramming while connecting to an AP,
after tx_sync() but before full setup, but that seems
like something that might very well happen.

Fix this by checking if the LQ is present and only then
reprogramming it.

Reported-by: Reinette Chatre <reinette.chatre@intel.com>
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>

iwlagn: check for SMPS mode

Check and report WARN only when its invalid

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwlwifi: two more SKUs for 6x05 series

Add two more SKUs for 6x05 series of device.
First SKU has low 5GHz channels actives, the other SKU has high 5GHz channels actives.

Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>