This patch fixes multiple issues with the current RTT
implementation in ath9k.

* The data that is obtained from the RTT interface registers
  are stored in 31:5 - mask out the extra bits when reading them.

* A history buffer is maintained which is not needed at all.
  Remove this array and just store the baseband data for each
  chain (or bank).

* A 'num_readings' variable was being used to handle the
  last entry. But it was being used in an improper manner, with
  the result that the RTT values were never being written
  to the RTT Interface registers. Fix this by using a simple
  flag.

* Stop baseband operations before programming the calibration values
  to the HW.

* Do not restore RX gain settings as part of RTT.
Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Use the new bool function ether_addr_equal to add
some clarity and reduce the likelihood for misuse
of compare_ether_addr for sorting.

Done via cocci script:

$ cat compare_ether_addr.cocci
@@
expression a,b;
@@
-	!compare_ether_addr(a, b)
+	ether_addr_equal(a, b)

@@
expression a,b;
@@
-	compare_ether_addr(a, b)
+	!ether_addr_equal(a, b)

@@
expression a,b;
@@
-	!ether_addr_equal(a, b) == 0
+	ether_addr_equal(a, b)

@@
expression a,b;
@@
-	!ether_addr_equal(a, b) != 0
+	!ether_addr_equal(a, b)

@@
expression a,b;
@@
-	ether_addr_equal(a, b) == 0
+	!ether_addr_equal(a, b)

@@
expression a,b;
@@
-	ether_addr_equal(a, b) != 0
+	ether_addr_equal(a, b)

@@
expression a,b;
@@
-	!!ether_addr_equal(a, b)
+	ether_addr_equal(a, b)
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Send the EAPOL and management frames in the same AC_VO queue. The issue
happens when the AP supports QOS, the management frames are sent to AC_VO
queue and EAP frame goes to AC_BE queue. Even though the EAP frame is
queued before the DEAUTH management frame, as they are queued on different
h/w queues, order of delivery between these frames cannot be controlled.

This fixes the connection failure seen in P2P case.
Signed-off-by: NVivek Natarajan <nataraja@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

This reverts commit a844adfd.
The commit a844adfd is degrading rx sensitivity of lower rate in
HT40 mode and it is confirmed that reverting the change is
improving rx sensitivity.

spur_freq_sd (for self-corr in AGC) is defined with respect to the
center of each 20MHz channel while spur_phase_delta (for self-corr
in Rx and spur data filter) is defined with respect to the center
of current RF channel.

So in short, we need to subtract spur_freq_sd (for self-corr in AGC)
by the offset between the center of primary20 and the center of RF
channel in SW. This offset could be +/10 MHz for dynamic 40.

Cc: Madhan Jaganathan <madhanj@qca.qualcomm.com>
Signed-off-by: NKai Shi <kaishi@qca.qualcomm.com>
Signed-off-by: NRajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Return status of ath6kl_htc_tx_issue() is ignored in
ath6kl_htc_tx_from_queue(), but failed tx packet is
is not cleaned up. To fix memory leak in this case, call
completion with error. Also, throw an error debug message
when tx fails in ath6kl_sdio_write_async() due to shortage
in bus request buffer.

kvalo: change the error message to WARN_ON_ONCE()
Signed-off-by: NVasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

There are few cases where the tx skb is dropped but netstats is
not updated, fix this.
Signed-off-by: NVasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

When our driver device is removed on the AHB bus, our IO memory is never unmapped.
Signed-off-by: NJonathan Bither <jonbither@gmail.com>
Reviewed-by: NPavel Roskin <proski@gnu.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

On the htc mbox layer, it will assign each endpoint (AC) with a
different Tx-packet-drop threshold, so lower priority AC is more
likely to drop packets and the cookies become more available to higher
priority AC.

On the htc pipe layer, assign the tx packet drop threshold as well, it
will let AC to drop packets when cookies below the tx packet drop
threshold.
Signed-off-by: NKevin Fang <kevin.fang@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

When a STA sends huge BK QoS data frame first and then sends BE/VI/VO
QoS data frame, the corresponding throughput becomes much lower than
that without sends BK QoS data frame before.

The root cause is that when station send BK stream, the tx credits of
BK stream don't return back to higher priority QoS stream such as BE,
VI, and VO stream. This patch will handle BK stream properly, when
there is higher priority QoS stream, it will seek tx credits from BK
stream properly.
Signed-off-by: NKevin Fang <kevin.fang@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

bss filter is configured to allow the frames from all the bss other
than the currenly connected one, this is done when a scan is requested
in connected state. There is no reason to filter out the currently
connected bss, configure the filter to allow all the bss. This would
fix the reporting of stale rssi of the current bss while scanning.
Reported-by: NNaveen singh <navesing@qca.qualcomm.com>
Signed-off-by: NVasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

With firmwares which do not support rsn capability override
(ATH6KL_FW_CAPABILITY_RSN_CAP_OVERRIDE) from host would
cause 4-way handshake failure when HT cap is advertised.
To fix this, do not advertise HT cap with cfg80211 for
those fw.

kvalo: fix a checkpatch warning
Signed-off-by: NVasanthakumar Thiagarajan <vthiagar@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

This patch enables all multicast packets in non suspend mode
and enable multicast filtering in wow suspend mode. This also
fixes a bug in multicast where the driver assumed disable
multicast-all command disabled/filtered all multicast
packets, which was wrong assumption, because firmware will
apply the programmed filter.

Multicast requirements
 - Enable forward all multicast packets(no filtering) in
      non suspend mode.
 - Enable multicast filtering in wow suspend mode for both
      AP and CLIENT.

kvalo: fix a checkpatch warning and drop unrelated newline removal
Signed-off-by: NNaveen Gangadharan <ngangadh@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

The patch "ath6kl: support fw reporting phy capabilities" gave the
firmware the ability to disable certain wiphy supported bands. Check if
this is the case in ath6kl_wmi_beginscan_cmd to avoid dereferencing a
NULL pointer.

kvalo: change the patch so that there's no code between declarations
Signed-off-by: NThomas Pedersen <c_tpeder@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Do not assume anyone before us includes it for us.
Signed-off-by: NLuis R. Rodriguez <mcgrof@do-not-panic.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

* PS_WAIT_FOR_TX_ACK is used in network-sleep mode and checking
  it for handling IDLE transitions is incorrect. Fix this.

* RX PCU/DMA engines have to be stopped before setting the chip into
  full-sleep mode - otherwise the chip becomes mute.

* Make things a bit clear by checking explicitly for network-sleep
  mode in the tx() routine and add a couple of debug statements
  to aid PS debugging.
Signed-off-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Extend debugfs entry for dfs_stats with DFS detection events
and shared pool statistics.
Signed-off-by: NZefir Kurtisi <zefir.kurtisi@neratec.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Fix the following bug: in tkip mode, qos-null ps on/off packets
are dropped due to incorrect packet length so that ath9k softap
can't handle powersave state transition of peer STA correctly.
Signed-off-by: NMichael Liang <mliang@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

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

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

It does not work properly and reduces throughput.
Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The delay calculation is the same for all chips, however some parts of the
code missed the extra delay factor for half/quarter.
Clean up the code and move the delay calculation to a common place.
Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

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

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

For some reason the MAC timing is a bit off when waiting for ACKs, so add
some extra delay to the ACK timeout values. Significantly reduces the
number of retransmissions in my tests.
Also disable the 2.4 GHz ACK timeout workaround in half/quarter mode, it
is not required there.
Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

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

AR9462 uses modified version of 3-Wire hw scheme for btcoex.
MCI itself is not a separate hw scheme but it aids to manage
multiple bt profiles. In ar9462, bt priority traffic is identified
by the number of bt profile types instead of gpio. So that this
patch removes MCI hw scheme.
Signed-off-by: NRajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

As of now beacon configuration is being called multiple times
in bss info change notification. This patch avoids multiple
configuration and make it simpler.
Signed-off-by: NRajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This adds locking of the detector's shared pulse and PRI sequence
pools to enable multi-wiphy operation on SMP systems.
Signed-off-by: NZefir Kurtisi <zefir.kurtisi@neratec.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The current code unconditionally reads the target
power values for all modes from the EEPROM. However
In 'ar9003_hw_set_power_per_rate_table' the regulatory
caps are applied only on a mode specific subset of the
power values.

The reported TX power level is calculated from the
maximum of the power values. Because some of these
values are uncapped in certain cases, the reported
TX power will be wrong.

On the older chipset, we don't have such problems
because only the mode specific subset of the power
levels are retrieved from the EEPROM on those. Do
the same for the AR9003 chips to fix the issue.
Signed-off-by: NGabor Juhos <juhosg@openwrt.org>
Acked-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This fixes two compile warnings, and removes a useless
cast when assigning the 'sc' variable.
Reported-by: NGabor Juhos <juhosg@openwrt.org>
Signed-off-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

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

The 'ath9k_hw_update_regulatory_maxpower'
helper function has been introduced by
commit a55f8588
(ath9k_hw: Cleanup TX power calculation for AR9287).

Updating of the max_power_level value has been moved
into the helper function in that change, however the
removed code from 'ath9k_hw_ar9287_set_txpower' has
not been replaced with a call of the new helper
function.

Due to that missing call, retrieving tx power for 2x2
and 3x3 chainmask is not handled properly. During the
calculation of the tx power for 2x2 and 3x3 chainmasks
the values are reduced. Those reductions must be
compensated during retrieving.

Fix this by adding the missing call of the helper
function.
Signed-off-by: NGabor Juhos <juhosg@openwrt.org>
Acked-by: NSujith Manoharan <c_manoha@qca.qualcomm.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

In order to unifying regulatory limit handling
commit ca2c68cc
(ath9k_hw: clean up tx power handling) introduced
a new helper function 'ath9k_hw_apply_txpower',
and the direct calls of 'ah->eep_ops->set_txpower'
has been replaced by a call of the helper function.

This caused a change in the behaviour of the
'ath9k_hw_set_txpowerlimit' function. The purpose
of that function is to calculate and store the
rate txpower table and the regulatory limit without
touching the hardware registers. Before the commit,
the 'test' parameter of the function was passed to
the 'ah->eep_ops->set_txpower'. Now the calling of
the 'set_txpower' function happens indirectly through
'ath9k_hw_apply_txpower', so the 'test' argument of
the 'set_txpower' is always 'false'.

This patch restores the original behaviour of
'ath9k_hw_set_txpowerlimit' by adding a new
argument to 'ath9k_hw_apply_txpower.'
Signed-off-by: NGabor Juhos <juhosg@openwrt.org>
Acked-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Complete the waiting fwlog_block_read on debugfs cleanup, otherwise
userspace and module unload might softlock.
Signed-off-by: NThomas Pedersen <c_tpeder@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

This is supposed to be pushed in one indent level.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

The original conditions are always true.  I think && was intended here,
but I don't have the hardware to test.  Could you take a look?

kvalo: Chilam confirmed that the fix is valid
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

The ath6kl fw does not have enough memory to support the WoW filters of
more than one vif. Disallow WoW suspend if this is the case.
Signed-off-by: NThomas Pedersen <c_tpeder@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Now that we know the supported PHY capabilities, only restore supported
bands / HT capabilities in firmware when stopping AP.
Signed-off-by: NThomas Pedersen <c_tpeder@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

Currently the supported bands are just hard coded in the driver.
However, the ath6kl FW will include its 11n and band capabilites in a
WMI_READY event. Handle this and report capabilites to cfg80211
accordingly.
Signed-off-by: NThomas Pedersen <c_tpeder@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>

ath6kl: If wow deepsleep pno suspend failure driver will choose cutpower.
in this scenario MMC_PM_KEEP_POWER should be cleared so that mmc will power
off the ar6000

kvalo: remove extra parenthesis
Signed-off-by: NMing Jiang <mjiang@qca.qualcomm.com>
Signed-off-by: NKalle Valo <kvalo@qca.qualcomm.com>