Commit Graph

375710 Commits

Author SHA1 Message Date
Ilan Peer 8eb3871076 iwlwifi: mvm: Update the supported interface combinations
iwlmvm does not support concurrent operation of AP with P2P Client/GO.
Update the interface limits to reflect that iwlmvm supports only
concurrent operation of station with AP and P2P Client/GO.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-11 20:17:06 +02:00
Emmanuel Grumbach 837fb69f10 iwlwifi: mvm: don't set the MCAST queue in STA's queue list
The MCAST queue should be enabled after DTIM only.
According to fw API, the MCAST must not be attached to any
station, but should appear in the mcast_qid of the AP's
mac context only.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-11 20:16:56 +02:00
Emmanuel Grumbach 5af01772ee iwlwifi: mvm: properly tell the fw that a STA is awake
The firmware API wasn't being used correctly, fix that.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-11 20:16:53 +02:00
Johannes Berg 8e00cc1261 Merge remote-tracking branch 'wireless-next/master' into iwlwifi-next 2013-06-11 20:16:44 +02:00
Johannes Berg 565f5a942b iwlwifi: include export.h instead of module.h
We only need EXPORT_SYMBOL_GPL, none of the other things
from module.h, so only include export.h.

Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-11 20:16:26 +02:00
Emmanuel Grumbach 476a9e09cb iwlwifi: ignore 0-length PHY DB sections
This can happen during development but can cause problems,
WARN (once) and go on.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Reviewed-by: Guy Cohen <guy.cohen@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-11 20:12:02 +02:00
Emmanuel Grumbach 0742a75a72 iwlwifi: mvm: fix irrelevant comment
This code moved and is now far away from the code that takes
the mutex.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-11 20:11:35 +02:00
Emmanuel Grumbach 629d0fa0da iwlwifi: mvm: remove unused wait_for_ba field
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-11 20:11:22 +02:00
Solomon Peachy 8b3e7be437 cw1200: Fix an assorted pile of checkpatch warnings.
Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-11 12:48:10 -04:00
Solomon Peachy 19db577868 cw1200: Eliminate the ETF debug/engineering code.
This is only really useful for people who are bringing up new hardware
designs and have access to the proprietary vendor tools that interface
with this mode.

It'll live out of tree until it's rewritten to use a less kludgy interface.

Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-11 12:48:10 -04:00
Solomon Peachy fa8eeae102 cw1200: Remove "ITP" debug subsystem.
This can live on as an out-of-tree patch for those that care.

Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-11 12:48:09 -04:00
Gabor Juhos 10af87c3d9 rt2x00: rt2x00queue: initialize data_queue fields earlier
Support for rt2800 device is broken since my
'rt2x00: rt2x00dev: use rt2x00dev->tx->limit'
patch. The changelog of that commit says that the
TX data queue is initialized already when the
rt2x00lib_probe_hw() function is called.

However as Jakub noticed it, this statement is not
correct. The queue->limit field is initialized in
the rt2x00queue_alloc_entries routine and that is
not yet called when rt2x00lib_probe_hw() runs.
Because the value of tx->limit contains zero, the
driver tries to allocate a kernel fifo with zero
size and kfifo_alloc rejects that with -EINVAL.

  PCI: Enabling device 0000:01:00.0 (0000 -> 0002)
  ieee80211 phy1: rt2x00_set_rt: Info - RT chipset 3071, rev 021c detected
  ieee80211 phy1: rt2x00_set_rf: Info - RF chipset 0008 detected
  ieee80211 phy1: rt2x00lib_probe_dev: Error - Failed to initialize hw
  rt2800pci: probe of 0000:01:00.0 failed with error -22

Move the data_queue field initialization from
the rt2x00queue_alloc_entries routine into the
rt2x00queue_init function. The initialization
code is not strictly related to the allocation,
and the change ensures that the queue_data fields
can be used in the probe routines.

The patch also introduces a helper function in
order to be able to get the correct data_queue_desc
structure for a given queue. This helper is only
needed temporarily and it will be removed later.

Reported-by: Jakub Kicinski <moorray@wp.pl>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-10 15:37:34 -04:00
Yunlian Jiang ec71997eff rtlwifi: initialize local array and set value.
GCC 4.8 is spitting out uninitialized-variable warnings against
"drivers/net/wireless/rtlwifi/rtl8192de/dm.c".

drivers/net/wireless/rtlwifi/rtl8192de/dm.c:941:31:
error: 'ofdm_index_old[1]' may be used uninitialized in this
function [-Werror=maybe-uninitialized]
     rtlpriv->dm.ofdm_index[i] = ofdm_index_old[i];

This patch adds initialization to the variable and properly sets its value.

Signed-off-by: Yunlian Jiang <yunlian@google.com>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-10 15:37:34 -04:00
John W. Linville 76015021aa Merge branch 'for-john' of git://git.kernel.org/pub/scm/linux/kernel/git/iwlwifi/iwlwifi-next 2013-06-10 14:51:17 -04:00
Dan Carpenter f7a01cac73 cw1200: handle allocation failure in wsm_event_indication()
Check for allocation failures and return -ENOMEM.  The caller
already expects it.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-10 14:41:26 -04:00
Dan Carpenter f28bc92c95 cw1200: read beyond end of array in debug code
This has only one caller and rates[] is an array with
IEEE80211_TX_MAX_RATES (4) elements.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-10 14:41:26 -04:00
Solomon Peachy aa63e18e3d cw1200: Sanity-check arguments in copy_from_user()
The optional debugfs interface to the vendor's engineering tools wasn't
bounds checking at all, which made it trivial to perform a buffer
overflow if this interface was compiled in and then explicitly enabled
at runtime.

This patch checks both the length supplied as part of the data to ensure
it is sane, and also the amount of data compared to the remaining buffer
space.  If either is too large, fail immediately.

(This bug was spotted by Dan Carpenter <dan.carpenter@oracle.com>)

Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-10 14:41:25 -04:00
Joe Perches cc2588eabb cw1200: hwio: Remove an unnecessary goto
goto after return is wrong.

The other code in this block needs to set an
error value then goto an error release block.

This one doesn't need to release anything and
was likely a copy/paste remainder.

Signed-off-by: Joe Perches <joe@perches.com>
Acked-By: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-10 14:41:25 -04:00
Johannes Berg ad19428835 iwlwifi: mvm: remove iwl_mvm_dbgfs_set_fw_dbg_log declaration
This function doesn't actually exist, remove its declaration.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-10 13:27:06 +02:00
Eytan Lifshitz fded313ec0 iwlwifi: mvm: don't request SMPS on non-STA iface
The Thermal Throttling code could do that, fix it.

Signed-off-by: Eytan Lifshitz <eytan.lifshitz@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-05 09:06:13 +02:00
Emmanuel Grumbach 9116a36839 iwlwifi: mvm: fix MCAST in AP mode
In multicast, there is no retries nor RTS since there is no
specific recipient that can ACK or send CTS. This means
that we must not use the rate scale table for multicast
frames.
This true for any frame that doesn't have a valid
ieee80211_sta pointer.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-05 09:06:13 +02:00
Alexander Bondar 5ec42ecf49 iwlwifi: mvm: enable PM always in unassociated mode
In unassociated BSS STA mode FW verifies both power save and power
management flags to decide on switching power off. The driver currently
sets power management flag according to mac80211 decision. As result, in
unassociated mode power management flag is down and power consumption is
high. Change power management enablement. When unassociated in BPS and
LP power save modes enable power management regardless of mac80211
decision. Rely on mac80211 decision if associated. Add power management
state update during associated/disassociated modes transitions.

Signed-off-by: Alexander Bondar <alexander.bondar@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-05 09:06:07 +02:00
Alexander Bondar 4bf881f5d5 iwlwifi: mvm: Change location of vif_count verification for PM
Currently vif_count verification for power management enablement appear
in different places. Move these verifications to one place in
iwl_mvm_update_power_mode().

Signed-off-by: Alexander Bondar <alexander.bondar@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:16:11 +02:00
Emmanuel Grumbach 86a91ec757 iwlwifi: mvm: correctly configure MCAST in AP mode
The AP mode needs to use the MCAST fifo for the MCAST
frames sent after the DTIM. This fifo needs to be
configured with the same parameters as the VOICE FIFO.

A separate SCD queue is mapped to this fifo - the cab_queue
(cab stands for Content After Beacon). This queue isn't
connected to any station, but rather to the MAC context.
This queue should (and is already) be set as the MCAST
queue - this is part of the of MAC context command.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Reviewed-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:16:02 +02:00
Emmanuel Grumbach 497b49d2f6 iwlwifi: mvm: reorder Rx handler for performance purposes
Since SCAN related handlers are much less likely than
beacon related handlers, reorder between them.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:15:49 +02:00
Ilan Peer 8d683b7948 iwlwifi: mvm: Fix quota command settings
According to the FW implementation, the quota command should
have a valid entry for each active binding (where 'active' in
this context means that the binding is known to the FW). In case
the binding should not get any quota, the 'quota' should be set
to zero.

Not setting an 0 quota for an active binding when all the MACs
in the binding are idle, i.e., not associated in case of managed
interface, will result in preventing the FW scheduler from entering
IDLE state and the FW from transitioning to low PS.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:15:27 +02:00
Eran Harary 1094fa2646 iwlwifi: don't return -ERFKILL if SEND_IF_RFKILL is set
When CMD_SEND_IN_RFKILL is set, it is perfectly legitimate
to send a host command while RFKILL is asserted. In this
case, the host command sending functions should return 0
even if RFKILL is asserted.

Signed-off-by: Eran Harary <eran.harary@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:14:46 +02:00
Emmanuel Grumbach 36946ce69a iwlwifi: mvm: don't enable MIMO when BT is active
Another step in the rate control / BT Coex integration

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:13:29 +02:00
Emmanuel Grumbach 7c8e4159ba iwlwifi: mvm: don't start BA agreement when BT is active
Otherwise WiFi would kill BT.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:13:20 +02:00
Emmanuel Grumbach e96d551f4b iwlwifi: mvm: limit the length of the AMPDU when BT is running
This holds for existing BA agreements.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:13:11 +02:00
Emmanuel Grumbach 6767583175 iwlwifi: mvm: fix the LUT for BT Coex
There was a typo in the Loose LUT for BT Coex.
Fix that.

Reported-by: Roi Cohen <roi.cohen@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-04 13:12:55 +02:00
John Greene 99e9494069 brcmsmac: Reduce log spam in heavy tx, make err print in debug
Move message to debug mode to reduce log spam under heavy tx (iperf) load.
This message prints in ht debug mode only:
brcms_c_ampdu_dotxstatus_complete: Pkt tx suppressed, illegal channel
possibly 153

Signed-off-by: John Greene <jogreene@redhat.com>
Acked-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:48 -04:00
Sujith Manoharan c24bd3620c ath9k: Do not maintain ANI state per-channel
ANI state can be maintained globally instead of per-channel.
This reduces memory usage and since default values are used
during a scan run, per-channel state is not required.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:48 -04:00
Sujith Manoharan 6e4d291eec ath9k: Print ANI statistics in debugfs
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:47 -04:00
Sujith Manoharan 4f4395c692 ath9k: Set ofdmWeakSigDetect directly
The macros ATH9K_ANI_USE_OFDM_WEAK_SIG can be removed.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:47 -04:00
Sujith Manoharan e323300d97 ath9k: Simplify ANI initialization
The check "enable_ani" is not required since it is always
set to true and the logic for disabling/enabling ANI via
debugfs is done at a higher layer.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:47 -04:00
Sujith Manoharan 380013fc94 ath9k: Remove unused structure ath_dbg_bb_mac_samp
Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:46 -04:00
Sujith Manoharan 846e438f5f ath9k: Enable WoW only for AR9462
The only card with which WoW has been tested and verified is
AR9462. Do not enable it for all cards since WoW is really quirky
and needs to be tested properly with each chip.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:46 -04:00
Sujith Manoharan 74a97755f2 ath9k: Return early for invalid rates
Process and update the internal RSSI average, which
is used by ANI, after verifying that the received
frame has valid rate information.

Signed-off-by: Sujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:45 -04:00
Yijing Wang bda96e8c6a net, ipw2x00: remove redundant D0 power state set
Pci_enable_device() will set device power state to D0,
so it's no need to do it again in ipw2100_pci_init_one().

Signed-off-by: Yijing Wang <wangyijing@huawei.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:45 -04:00
Felix Fietkau ec26bcc055 ath9k: advertise support for active monitor interfaces
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:55:45 -04:00
John W. Linville 133eb82f09 Merge branch 'for-john' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next 2013-06-03 15:55:37 -04:00
Solomon Peachy 4da2a54a84 cw1200: rename the cw1200 platform definition header
My previous patch just moved the file, but it also needed to be renamed
to conform to proper conventions.

Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:54:56 -04:00
Solomon Peachy 7c0b6f49db cw1200: Rework SDIO platform support to prevent build problems.
Based on discussions with And Bergmann, this patch changes the SDIO
platform code to default to supporting the Sagrad devices, allowing for
it to be overridden in board setup code.  This renders the cw1200_sagrad
module suplerflous, so it is now removed.

It also moves the documentation that was in the cw1200_sagrad source to
the platform header.

Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:54:55 -04:00
Solomon Peachy 6dd64a304e cw1200: Replace use of 'struct resource' with 'int' for GPIO fields.
The only advantage of 'struct resource' is that it lets us assign names
as part of the platform data.  Unfortunately since we are using platform
data, we are already limited to a single instance of each driver,
rendering this moot.

So, replace the struct resources with ints, resulting in cleaner code.

This was based on a suggestion from Arnd Bergmann.

Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:54:55 -04:00
Solomon Peachy 7b19bc2ca9 cw1200: Reference correct 'powerup' GPIO signal.
Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:54:54 -04:00
Solomon Peachy c992219825 cw1200: move platform_data header to correct location.
(As suggested by Arnd Bergmann)

Signed-off-by: Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:54:54 -04:00
Solomon Peachy 911373cca1 cw1200: Rename 'sbus' to 'hwbus'
This avoids problems when building on SPARC targets due to the driver
calling the bus abstraction layer 'sbus'.  Not that any SBUS-sporting
SPARC targets are likely to have an SDIO controller, but this is the
correct thing to do.

See http://kisskb.ellerman.id.au/kisskb/buildresult/8846508/

Signed-off-by:  Solomon Peachy <pizza@shaftnet.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
2013-06-03 15:54:37 -04:00
Johannes Berg ed405be5cb mac80211: fix sdata locking around __ieee80211_request_smps
My cfg80211/mac80211 locking unification broke the sdata
locking in ieee80211_set_power_mgmt, it needs to acquire
the lock for __ieee80211_request_smps(). Add the locking.

Reported-by: Jakub Kicinski <kubakici@wp.pl>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-06-03 13:51:59 +02:00
Johannes Berg a8cd57b304 Revert "mac80211_hwsim: remove P2P_DEVICE support"
This reverts commit e3ee68b7b0.

This wasn't intended to be included here, my mistake. I
accidentally merged a mac80211 fixes tree here that had
this change, when it wasn't even intended to be there.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
2013-05-31 10:23:05 +02:00