OSDN Git Service

uclinux-h8/linux.git
5 years agobrcmfmac: Call brcmf_dmi_probe before brcmf_of_probe
Hans de Goede [Fri, 23 Nov 2018 09:11:48 +0000 (10:11 +0100)]
brcmfmac: Call brcmf_dmi_probe before brcmf_of_probe

ARM systems with UEFI may have both devicetree (of) and DMI data in this
case we end up setting brcmf_mp_device.board_type twice.

In this case we should prefer the devicetree data, because:
1) The devicerree data is more reliable
2) Some ARM systems (e.g. the Raspberry Pi 3 models) support both UEFI and
   classic uboot booting, the devicetree data is always there, so using it
   makes sure we ask for the same nvram file independent of how we booted.

This commit moves the brcmf_dmi_probe call to before the brcmf_of_probe
call, so that the latter can override the value of the first if both are
set.

Fixes: bd1e82bb420a ("brcmfmac: Set board_type from DMI on x86 based ...")
Cc: Peter Robinson <pbrobinson@gmail.com>
Tested-and-reported-by: Peter Robinson <pbrobinson@gmail.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: support STA info struct v7
Dan Haab [Fri, 9 Nov 2018 16:38:55 +0000 (09:38 -0700)]
brcmfmac: support STA info struct v7

The newest firmwares provide STA info using v7 of the struct. As v7
isn't backward compatible, a union is needed.

Even though brcmfmac does not use any of the new info it's important to
provide the proper struct buffer. Without this change new firmwares will
fallback to the very limited v3 instead of something in between such as
v4.

Signed-off-by: Dan Haab <dan.haab@luxul.com>
Reviewed-by: Rafał Miłecki <rafal@milecki.pl>
Reviewed-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agob43: Use cordic algorithm from kernel library
Priit Laes [Mon, 19 Nov 2018 18:01:25 +0000 (20:01 +0200)]
b43: Use cordic algorithm from kernel library

Kernel library has a common cordic algorithm which is identical
to internally implemented one, so use it and drop the duplicate
implementation.

Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Priit Laes <plaes@plaes.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agob43: Fix error in cordic routine
Larry Finger [Mon, 19 Nov 2018 18:01:24 +0000 (20:01 +0200)]
b43: Fix error in cordic routine

The cordic routine for calculating sines and cosines that was added in
commit 6f98e62a9f1b ("b43: update cordic code to match current specs")
contains an error whereby a quantity declared u32 can in fact go negative.

This problem was detected by Priit Laes who is switching b43 to use the
routine in the library functions of the kernel.

Fixes: 986504540306 ("b43: make cordic common (LP-PHY and N-PHY need it)")
Reported-by: Priit Laes <plaes@plaes.org>
Cc: Rafał Miłecki <zajec5@gmail.com>
Cc: Stable <stable@vger.kernel.org> # 2.6.34
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Priit Laes <plaes@plaes.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmsmac: Use cordic-related macros from common cordic library
Priit Laes [Mon, 19 Nov 2018 18:01:23 +0000 (20:01 +0200)]
brcmsmac: Use cordic-related macros from common cordic library

Current driver includes macro that is available from general cordic
library. Use that and drop unused duplicate and unneeded internal
definitions.

Signed-off-by: Priit Laes <plaes@plaes.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agolib: cordic: Move cordic macros and defines to header file
Priit Laes [Mon, 19 Nov 2018 18:01:22 +0000 (20:01 +0200)]
lib: cordic: Move cordic macros and defines to header file

Now that these macros are in header file, we can eventually
clean up the duplicate macros present in the drivers that
utilize the same cordic algorithm implementation.

Also add CORDIC_ prefix to nonprefixed macros.

Reviewed-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Priit Laes <plaes@plaes.org>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoMerge tag 'iwlwifi-next-for-kalle-2018-11-23' of git://git.kernel.org/pub/scm/linux...
Kalle Valo [Thu, 29 Nov 2018 09:00:00 +0000 (11:00 +0200)]
Merge tag 'iwlwifi-next-for-kalle-2018-11-23' of git://git./linux/kernel/git/iwlwifi/iwlwifi-next

second batch of iwlwifi patches intended for v4.21

* New FW debugging infrastructure;
* Some more work on 802.11ax;
* Improve support for multiple RF modules with 22000 devices;
* Remove an unused FW parameter;
* Other debugging improvements;

5 years agoiwlwifi: mvm: set HW capability VHT_EXT_NSS_BW
Shaul Triebitz [Wed, 18 Jul 2018 07:29:53 +0000 (10:29 +0300)]
iwlwifi: mvm: set HW capability VHT_EXT_NSS_BW

Enable the VHT extended NSS BW feature in iwlwifi/mvm.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: send the STA_HE_CTXT command in AP mode
Shaul Triebitz [Wed, 8 Aug 2018 10:24:50 +0000 (13:24 +0300)]
iwlwifi: mvm: send the STA_HE_CTXT command in AP mode

In AP mode, if AP supports HE (and the STA), send the
STA_HE_CTXT command.

This is needed mainly for PPE (packet extension) params.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: set MAC_FILTER_IN_11AX in AP mode
Shaul Triebitz [Wed, 8 Aug 2018 10:20:45 +0000 (13:20 +0300)]
iwlwifi: mvm: set MAC_FILTER_IN_11AX in AP mode

In AP mode, if AP supports 11ax, add the MAC_FILTER_IN_11AX
flag in MAC_CTXT command (needed for various 11ax stuff).

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: fw: support API change to debug TLVs
Sara Sharon [Mon, 13 Aug 2018 05:21:10 +0000 (08:21 +0300)]
iwlwifi: fw: support API change to debug TLVs

A new field was added. Since the code isn't operational (yet) no
need to worry about backward compatibility.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: add debugfs file to read fw debug data recording
Lior Cohen [Thu, 3 May 2018 07:17:05 +0000 (10:17 +0300)]
iwlwifi: add debugfs file to read fw debug data recording

FW debug data will oneshot read all data available in DRAM
and fill the supplied user buffer. In case the read request
is greater than the new data in DRAM, the driver will write
all data it has and return the buffer immediately.

Signed-off-by: Shahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: Lior Cohen <lior2.cohen@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: dbg: add apply point logic
Sara Sharon [Wed, 13 Jun 2018 12:19:47 +0000 (15:19 +0300)]
iwlwifi: dbg: add apply point logic

Add a function to be called when apply point occurs.
For each of the TLVs, the function will perform the
apply point logic:
- For HCMD - send the stored host command
- For buffer allocation - allocate the memory and send the
  buffer allocation command
- For trigger and region - update the stored configuration

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: dbg: disable triggers for ini
Sara Sharon [Wed, 13 Jun 2018 11:41:35 +0000 (14:41 +0300)]
iwlwifi: dbg: disable triggers for ini

When ini is loaded, disable all legacy trigger
configuration.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: trans: support loading ini TLVs from external file
Sara Sharon [Thu, 14 Jun 2018 07:49:03 +0000 (10:49 +0300)]
iwlwifi: trans: support loading ini TLVs from external file

Support loading and storing ini TLVs from external
file. Those TLVs are appended to the default TLVs,
so store them separately.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: trans: parse and store debug ini TLVs
Sara Sharon [Thu, 17 May 2018 07:14:30 +0000 (10:14 +0300)]
iwlwifi: trans: parse and store debug ini TLVs

The new debug ini TLVs can be either packed into firmware
binary or written in external file. Support loading them
from both. Store the data per apply point. Apply point is
a point during driver runtime, where the TLV becomes active.
For example, a trigger of hardware error may be configured
to collect a subset of data pre-alive, as a opposed to HW
error that occurs after alive.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: fw: add FW API of new TLV infrastructure
Sara Sharon [Wed, 13 Jun 2018 11:58:18 +0000 (14:58 +0300)]
iwlwifi: fw: add FW API of new TLV infrastructure

Add the FW API of the new debug infrastructure. Next patches
will introduce the utilization of this infra.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: add support for TWT capabilities
Emmanuel Grumbach [Wed, 8 Aug 2018 07:42:23 +0000 (10:42 +0300)]
iwlwifi: mvm: add support for TWT capabilities

We need to check the TWT support of the peer and to
propagte the capability to the firmware.
The current implementation will enable TWT only if the TWT
support is advertised in the HE CAP IE and in the Extended
Capability IE.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: add fw dump upon RT ucode start failure
Lior Cohen [Thu, 2 Aug 2018 08:48:18 +0000 (11:48 +0300)]
iwlwifi: add fw dump upon RT ucode start failure

FW dump was missing in case the RT FW ucode
section failed to load. This failure happens when
the RT section of the FW file is corrupted.

Signed-off-by: Lior Cohen <lior2.cohen@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: fix cfg structs for 22000 with different RF modules
Luca Coelho [Sat, 4 Aug 2018 11:45:42 +0000 (14:45 +0300)]
iwlwifi: fix cfg structs for 22000 with different RF modules

We have to choose different configuration and different firmwares
depending on the external RF module that is installed.  Since the
external module is not represented in the PCI IDs, we need to change
the configuration at runtime, after checking the RF ID of the module
installed.  We have a bit of a mess in the code that does this,
because it applies cfg's according to the RF ID only, ignoring the
integrated module that is in use.

Fix that for some devices by adding correct configurations for them
and not ignoring the integrated module's type when making the
decision.

Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: remove assignment of the reciprocal
Emmanuel Grumbach [Tue, 31 Jul 2018 12:34:20 +0000 (15:34 +0300)]
iwlwifi: mvm: remove assignment of the reciprocal

The firmware stopped looking at this field long ago.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: dbg: split fifos dump
Sara Sharon [Tue, 31 Jul 2018 06:54:26 +0000 (09:54 +0300)]
iwlwifi: dbg: split fifos dump

Split the dump of RXF and TXF. This is in order to
enable code reuse for INI, which may decide to dump
only RXF and not TXF, and vice versa.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: dbg: don't limit dump decisions to all or monitor
Sara Sharon [Mon, 30 Jul 2018 08:59:16 +0000 (11:59 +0300)]
iwlwifi: dbg: don't limit dump decisions to all or monitor

Currently opmode is limited to asking transport to either
dump all the dumps configured at startup, or monitor only.
Instead, pass to transport a bitmask, to allow flexibility.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoMerge tag 'iwlwifi-next-for-kalle-2018-11-11' of git://git.kernel.org/pub/scm/linux...
Kalle Valo [Thu, 15 Nov 2018 14:59:01 +0000 (16:59 +0200)]
Merge tag 'iwlwifi-next-for-kalle-2018-11-11' of git://git./linux/kernel/git/iwlwifi/iwlwifi-next

First set of iwlwifi patches for 4.21

* PCI IDs for some new 9000-series cards;
* Improve antenna usage on connection problems;
* Some improvements in the debugging code;
* Other clean-ups and small fixes;

5 years agoiwlwifi: fw: use helper to determine whether to dump paging
Sara Sharon [Mon, 30 Jul 2018 08:06:09 +0000 (11:06 +0300)]
iwlwifi: fw: use helper to determine whether to dump paging

Logic is there twice, and we'll need a third place
soon for ini dumping. In addition move the dumping
to a function, also to enable reuse.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: add new cards for 9560, 9462, 9461 and killer series
Ihab Zhaika [Tue, 31 Jul 2018 06:53:09 +0000 (09:53 +0300)]
iwlwifi: add new cards for 9560, 9462, 9461 and killer series

add few PCI ID'S for 9560, 9462, 9461 and killer series.

Cc: stable@vger.kernel.org
Signed-off-by: Ihab Zhaika <ihab.zhaika@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: remove unused conversion table
Emmanuel Grumbach [Mon, 30 Jul 2018 12:57:08 +0000 (15:57 +0300)]
iwlwifi: mvm: remove unused conversion table

chanwidths isn't used now in debugfs-vif.c.  Remove it.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: dbg: use helper to check if debug type is set
Sara Sharon [Mon, 30 Jul 2018 06:43:24 +0000 (09:43 +0300)]
iwlwifi: dbg: use helper to check if debug type is set

This enables to incorporate more logic in one place.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: Send non offchannel traffic via AP sta
Andrei Otcheretianski [Tue, 24 Jul 2018 18:57:50 +0000 (21:57 +0300)]
iwlwifi: mvm: Send non offchannel traffic via AP sta

TDLS discovery response frame is a unicast direct frame to the peer.
Since we don't have a STA for this peer, this frame goes through
iwl_tx_skb_non_sta(). As the result aux_sta and some completely
arbitrary queue would be selected for this frame, resulting in a queue
hang.  Fix that by sending such frames through AP sta instead.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: add description to a few generic assert numbers
Naftali Goldstein [Wed, 25 Jul 2018 12:52:20 +0000 (15:52 +0300)]
iwlwifi: mvm: add description to a few generic assert numbers

FW asserts 0x70, 0x71, and 0x73 all just mean that the real error
happened in another MAC, and to look there for the problem. Add their
descriptions to the assert number lookup table so users get a nicer
error message in the logs.
Also, since the 4 most-significant bits of the assert number are
dynamic, and depend on which MAC the assert occurred on, ignore those
bits when looking up the assert name.

Signed-off-by: Naftali Goldstein <naftali.goldstein@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: dbg: avoid passing trigger around
Sara Sharon [Mon, 23 Jul 2018 10:50:02 +0000 (13:50 +0300)]
iwlwifi: dbg: avoid passing trigger around

The trigger structure is being passed around, when
all we care about is whether to dump only monitor
or not. Pass a bool instead.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: pcie: don't reset TXQ write pointer
Sara Sharon [Thu, 19 Jul 2018 06:15:11 +0000 (09:15 +0300)]
iwlwifi: pcie: don't reset TXQ write pointer

Currently code sets the write pointer when getting the TX queue
allocate response. This causes a redundant interrupt with any actual
change in the pointer. Remove this write altogether.

Fixes: 310181ec34e2 ("iwlwifi: move to TVQM mode")
Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: toggle tx antenna if tx fails during connection establishment
Avraham Stern [Wed, 11 Jul 2018 07:30:20 +0000 (10:30 +0300)]
iwlwifi: mvm: toggle tx antenna if tx fails during connection establishment

If tx fails during connection establishment, try another antenna for
the next tx. This will increase the chance to establish connection if
one of the antennas is blocked.  Note that the antenna is toggled even
when failing to tx data frames since connection establishment may use
EAPOLs for 802.1X authentication/ 4 way handshake.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: fw: do not set sgi bits for HE connection
Naftali Goldstein [Sun, 15 Jul 2018 12:26:27 +0000 (15:26 +0300)]
iwlwifi: fw: do not set sgi bits for HE connection

If the association supports HE, HT/VHT rates will never be used for Tx
and therefore there's no need to set the sgi-per-channel-width-support
bits, so don't set them in this case.

Fixes: 110b32f065f3 ("iwlwifi: mvm: rs: add basic implementation of the new RS API handlers")
Signed-off-by: Naftali Goldstein <naftali.goldstein@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: switch management tx antenna only on tx failure
Avraham Stern [Thu, 5 Jul 2018 14:34:03 +0000 (17:34 +0300)]
iwlwifi: mvm: switch management tx antenna only on tx failure

Switch the antenna used for management tx only if previous tx failed.
If previous tx succeeded, there is no reason to switch antennas.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: enable low latency for soft ap
Tova Mussai [Mon, 2 Jul 2018 06:08:29 +0000 (09:08 +0300)]
iwlwifi: mvm: enable low latency for soft ap

Enable low latency for softAP in all modes (standalone, SCM
and DCM).
This is in order to minimize the time the softAP leaves the channel for
other operations

Signed-off-by: Tova Mussai <tova.mussai@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: trans: Clear persistence bit when starting the FW
Shahar S Matityahu [Wed, 4 Jul 2018 12:31:36 +0000 (15:31 +0300)]
iwlwifi: trans: Clear persistence bit when starting the FW

In D3 suspend flow in 9260 gen2 HW, the NIC receives two PERST signals.
The first PERST is expected and indicates the device on coming resume flow.
The second PERST causes FW restart FW restart.
In order to avoid this issue, the FW set the persistence bit on.
Once this bit is set, the FW ignores reset attempts.
The problem is when the FW gets assert during D3 and then the persistence
bit is set and causes the FW to ignore reset.
To handle this issue, the FW opens the preg bit which allows access
to the persistence bit, so that the driver clear the persistence bit
and reset the NIC.

The flow is as follows:
the driver checks if the persistence bit is set.
If the bit is set, the driver checks if he can clear the bit.
If the driver can not clear the bit then there is no point to continue
configuring the NIC since it will fail.

The fix was added is in start HW flow instead of the resume flow since in
general, if the persistence bit is set, the driver can not start the FW.
So it is good to check it when we start configuring the NIC.

The driver does not need to close the preg bit since the FW close it
during the start flow.

Signed-off-by: Shahar S Matityahu <shahar.s.matityahu@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: remove queue_info_lock
Johannes Berg [Wed, 4 Jul 2018 21:19:13 +0000 (23:19 +0200)]
iwlwifi: mvm: remove queue_info_lock

All the queue management code runs under mvm->mutex, so there are
only very few cases of accessing the data structures without it:
 * TX path, which doesn't take any locks anyway
 * iwl_mvm_wake_sw_queue() and iwl_mvm_stop_sw_queue() where we
   just (atomically) read a bitmap, so the lock isn't needed.

Therefore, we can remove the spinlock. This enables some cleanup
in the ugly locking in iwl_mvm_inactivity_check().

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: synchronize TID queue removal
Johannes Berg [Wed, 4 Jul 2018 21:12:33 +0000 (23:12 +0200)]
iwlwifi: mvm: synchronize TID queue removal

When we mark a TID as no longer having a queue, there's no
guarantee the TX path isn't using this txq_id right now,
having accessed it just before we reset the value. To fix
this, add synchronize_net() when we change the TIDs from
having a queue to not having one, so that we can then be
sure that the TX path is no longer accessing that queue.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoiwlwifi: mvm: remove set but not used variable 'he_phy_data'
YueHaibing [Fri, 5 Oct 2018 10:47:01 +0000 (10:47 +0000)]
iwlwifi: mvm: remove set but not used variable 'he_phy_data'

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/net/wireless/intel/iwlwifi/mvm/rxmq.c: In function 'iwl_mvm_rx_mpdu_mq':
drivers/net/wireless/intel/iwlwifi/mvm/rxmq.c:1386:7: warning:
 variable 'he_phy_data' set but not used [-Wunused-but-set-variable]
   u64 he_phy_data;

'he_phy_data' never used since be introduce in
commit 18ead597daa1 ("iwlwifi: support new rx_mpdu_desc api")

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
5 years agoMerge ath-next from git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git
Kalle Valo [Fri, 9 Nov 2018 15:15:25 +0000 (17:15 +0200)]
Merge ath-next from git://git./linux/kernel/git/kvalo/ath.git

ath.git patches for 4.21. Major changes:

ath10k

* add support for WCN3990 firmware crash recovery

* add firmware memory dump support for QCA4019

wil6210

* add firmware error recovery while in AP mode

ath9k

* remove experimental notice from dynack feature

5 years agortlwifi: rtl8192de: Fix misleading REG_MCUFWDL information
Shaokun Zhang [Mon, 5 Nov 2018 11:25:30 +0000 (19:25 +0800)]
rtlwifi: rtl8192de: Fix misleading REG_MCUFWDL information

RT_TRACE shows REG_MCUFWDL value as a decimal value with a '0x'
prefix, which is somewhat misleading.

Fix it to print hexadecimal, as was intended.

Cc: Ping-Ke Shih <pkshih@realtek.com>
Cc: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Shaokun Zhang <zhangshaokun@hisilicon.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoiwlegacy: fix small typo
Yangtao Li [Thu, 1 Nov 2018 15:40:23 +0000 (11:40 -0400)]
iwlegacy: fix small typo

Signed-off-by: Yangtao Li <tiny.windzz@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agocw1200: fix small typo
Yangtao Li [Thu, 1 Nov 2018 15:33:19 +0000 (11:33 -0400)]
cw1200: fix small typo

Signed-off-by: Yangtao Li <tiny.windzz@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmutil: print invalid chanspec when WARN-ing
Rafał Miłecki [Fri, 26 Oct 2018 11:22:32 +0000 (13:22 +0200)]
brcmutil: print invalid chanspec when WARN-ing

On one of my devices I got WARNINGs when brcmfmac tried to decode
chanspec. I couldn't tell if it was some unsupported format or just a
malformed value passed by a firmware.

Print chanspec value so it's possible to debug a possible problem.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agowireless: airo: potential buffer overflow in sprintf()
Dan Carpenter [Wed, 24 Oct 2018 08:33:34 +0000 (11:33 +0300)]
wireless: airo: potential buffer overflow in sprintf()

It looks like we wanted to print a maximum of BSSList_rid.ssidLen bytes
of the ssid, but we accidentally use "%*s" (width) instead of "%.*s"
(precision) so if the ssid doesn't have a NUL terminator this could lead
to an overflow.

Static analysis.  Not tested.

Fixes: e174961ca1a0 ("net: convert print_mac to %pM")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agortlwifi: Remove same duplicated includes
zhong jiang [Wed, 24 Oct 2018 06:07:36 +0000 (14:07 +0800)]
rtlwifi: Remove same duplicated includes

Just remove same duplicated includes.

Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agortlwifi: rtl8723ae: Remove set but not used variables and #defines
zhong jiang [Wed, 24 Oct 2018 05:52:09 +0000 (13:52 +0800)]
rtlwifi: rtl8723ae: Remove set but not used variables and #defines

radiob_array_table' and 'radiob_arraylen' are not used after setting its value.
It is safe to remove the unused variable. Meanwhile, radio B array should be
removed as well. because it will no longer be referenced.

Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmsmac: never log "tid x is not agg'able" by default
Ali MJ Al-Nasrawy [Tue, 23 Oct 2018 16:12:35 +0000 (19:12 +0300)]
brcmsmac: never log "tid x is not agg'able" by default

This message greatly spams the log under heavy Tx of frames with BK access
class which is especially true when operating as AP. It is also not informative
as the "agg'ablity" of TIDs are set once and never change.
Fix this by logging only in debug mode.

Signed-off-by: Ali MJ Al-Nasrawy <alimjalnasrawy@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agozd1201: mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:48:26 +0000 (22:48 +0200)]
zd1201: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agortlwifi: rtl8821ae: phy: Mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:47:59 +0000 (22:47 +0200)]
rtlwifi: rtl8821ae: phy: Mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoray_cs: mark expected switch fall-throughs
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:47:20 +0000 (22:47 +0200)]
ray_cs: mark expected switch fall-throughs

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Addresses-Coverity-ID: 114948 ("Missing break in switch")
Addresses-Coverity-ID: 114949 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agort2x00: rt61pci: mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:46:47 +0000 (22:46 +0200)]
rt2x00: rt61pci: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agort2x00: rt2800lib: mark expected switch fall-throughs
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:46:03 +0000 (22:46 +0200)]
rt2x00: rt2800lib: mark expected switch fall-throughs

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Addresses-Coverity-ID: 145198 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agort2x00: rt2500pci: mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:45:19 +0000 (22:45 +0200)]
rt2x00: rt2500pci: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agort2x00: rt2400pci: mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:44:34 +0000 (22:44 +0200)]
rt2x00: rt2400pci: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agomwifiex: Mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:43:55 +0000 (22:43 +0200)]
mwifiex: Mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoprism54: islpci_dev: mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:43:05 +0000 (22:43 +0200)]
prism54: islpci_dev: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Addresses-Coverity-ID: 114947 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoprism54: isl_ioctl: mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:42:47 +0000 (22:42 +0200)]
prism54: isl_ioctl: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoprism54: isl_38xx: Mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:42:29 +0000 (22:42 +0200)]
prism54: isl_38xx: Mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Addresses-Coverity-ID: 114944 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoorinoco_usb: mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:42:04 +0000 (22:42 +0200)]
orinoco_usb: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoiwlegacy: common: mark expected switch fall-throughs
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:41:23 +0000 (22:41 +0200)]
iwlegacy: common: mark expected switch fall-throughs

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Addresses-Coverity-ID: 201384 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoiwlegacy: 4965-mac: mark expected switch fall-through
Gustavo A. R. Silva [Mon, 22 Oct 2018 20:40:53 +0000 (22:40 +0200)]
iwlegacy: 4965-mac: mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agortl8xxxu: Fix missing break in switch
Gustavo A. R. Silva [Mon, 22 Oct 2018 11:51:03 +0000 (13:51 +0200)]
rtl8xxxu: Fix missing break in switch

Add missing break statement in order to prevent the code from falling
through to the default case.

Fixes: 26f1fad29ad9 ("New driver: rtl8xxxu (mac80211)")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agortl8xxxu: Mark expected switch fall-throughs
Gustavo A. R. Silva [Mon, 22 Oct 2018 11:50:08 +0000 (13:50 +0200)]
rtl8xxxu: Mark expected switch fall-throughs

In preparation to enabling -Wimplicit-fallthrough, mark switch cases
where we are expecting to fall through.

Addresses-Coverity-ID: 1357355 ("Missing break in switch")
Addresses-Coverity-ID: 1357378 ("Missing break in switch")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agortlwifi: btcoex: remove set but not used variable 'ppsc'
YueHaibing [Mon, 22 Oct 2018 07:51:15 +0000 (07:51 +0000)]
rtlwifi: btcoex: remove set but not used variable 'ppsc'

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtcoutsrc.c: In function 'halbtc_leave_lps':
drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtcoutsrc.c:295:21: warning:
 variable 'ppsc' set but not used [-Wunused-but-set-variable]

drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtcoutsrc.c: In function 'halbtc_enter_lps':
drivers/net/wireless/realtek/rtlwifi/btcoexist/halbtcoutsrc.c:318:21: warning:
 variable 'ppsc' set but not used [-Wunused-but-set-variable]

It never used since introduction in
commit aa45a673b291 ("rtlwifi: btcoexist: Add new mini driver")

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoqtnfmac: fix error handling in control path
Sergey Matyukevich [Wed, 17 Oct 2018 14:02:47 +0000 (14:02 +0000)]
qtnfmac: fix error handling in control path

This patch fixes the following warnings:

- smatch
drivers/net/wireless/quantenna/qtnfmac/commands.c:132 qtnf_cmd_send_with_reply() warn: variable dereferenced before check 'resp' (see line 117)
drivers/net/wireless/quantenna/qtnfmac/commands.c:716  qtnf_cmd_get_sta_info() error: uninitialized symbol 'var_resp_len'.
drivers/net/wireless/quantenna/qtnfmac/commands.c:1668 qtnf_cmd_get_mac_info() error: uninitialized symbol 'var_data_len'.
drivers/net/wireless/quantenna/qtnfmac/commands.c:1697 qtnf_cmd_get_hw_info() error: uninitialized symbol 'info_len'.
drivers/net/wireless/quantenna/qtnfmac/commands.c:1753 qtnf_cmd_band_info_get() error: uninitialized symbol 'info_len'.
drivers/net/wireless/quantenna/qtnfmac/commands.c:1782 qtnf_cmd_send_get_phy_params() error: uninitialized symbol 'response_size'.
drivers/net/wireless/quantenna/qtnfmac/commands.c:2438 qtnf_cmd_get_chan_stats() error: uninitialized symbol 'var_data_len'.

- gcc-8.2.1
drivers/net/wireless/quantenna/qtnfmac/commands.c: In function 'qtnf_cmd_send_with_reply':
drivers/net/wireless/quantenna/qtnfmac/commands.c:133:54: error: 'resp' may be used uninitialized in this function [-Werror=maybe-uninitialized]

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sergey Matyukevich <sergey.matyukevich.os@quantenna.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: fix spelling mistake "Retreiving" -> "Retrieving"
Colin Ian King [Tue, 16 Oct 2018 17:43:42 +0000 (18:43 +0100)]
brcmfmac: fix spelling mistake "Retreiving" -> "Retrieving"

Trivial fix to spelling mistake in brcmf_err error message.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoqtnfmac: add support for Topaz chipsets
Sergey Matyukevich [Tue, 16 Oct 2018 10:23:58 +0000 (10:23 +0000)]
qtnfmac: add support for Topaz chipsets

This patch adds support for QSR1000/QSR2000 family of chipsets
to qtnfmac_pcie platform driver.

QSR1000/QSR2000 (aka Topaz) is a family of 80MHz, 11ac Wave2,
4x4/2x4/2x2 chips, including single and dual band devices.
Depending on specific chip model and firmware in use, either
STA or both STA and AP modes are supported.

Patch adds Topaz support to qtnfmac_pcie driver. Proper platform
bus will be selected on probing based on chip ID.

Signed-off-by: Igor Mitsyanko <igor.mitsyanko.os@quantenna.com>
Signed-off-by: Sergey Matyukevich <sergey.matyukevich.os@quantenna.com>
Signed-off-by: Andrey Shevchenko <ashevchenko@quantenna.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoqtnfmac_pcie: use single PCIe driver for all platforms
Igor Mitsyanko [Tue, 16 Oct 2018 10:23:56 +0000 (10:23 +0000)]
qtnfmac_pcie: use single PCIe driver for all platforms

Single PCIe driver can identify hardware type by reading CHIP ID at
probe time and invoking a correct initialization sequence.

Signed-off-by: Igor Mitsyanko <igor.mitsyanko.os@quantenna.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agowlcore: Fix the return value in case of error in 'wlcore_vendor_cmd_smart_config_start()'
Christophe JAILLET [Tue, 16 Oct 2018 07:39:40 +0000 (09:39 +0200)]
wlcore: Fix the return value in case of error in 'wlcore_vendor_cmd_smart_config_start()'

We return 0 unconditionally at the end of
'wlcore_vendor_cmd_smart_config_start()'.
However, 'ret' is set to some error codes in several error handling paths
and we already return some error codes at the beginning of the function.

Return 'ret' instead to propagate the error code.

Fixes: 80ff8063e87c ("wlcore: handle smart config vendor commands")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agolibertas: remove set but not used variable 'int_type'
YueHaibing [Mon, 15 Oct 2018 02:16:35 +0000 (02:16 +0000)]
libertas: remove set but not used variable 'int_type'

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/net/wireless/marvell/libertas/if_spi.c: In function 'if_spi_h2c':
drivers/net/wireless/marvell/libertas/if_spi.c:799:6: warning:
 variable 'int_type' set but not used [-Wunused-but-set-variable]

It never used since introduction in
commit d2b21f191753 ("libertas: if_spi, driver for libertas GSPI devices")

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Tested-by: Lubomir Rintel <lkundrak@v3.sk>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agorsi: fix spelling mistake "Initialzing" -> "Initializing"
Colin Ian King [Sat, 13 Oct 2018 15:37:30 +0000 (16:37 +0100)]
rsi: fix spelling mistake "Initialzing" -> "Initializing"

Trivial fix to spelling mistake in rsi_dbg debug message

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: Fix ccode from EFI nvram when necessary
Hans de Goede [Thu, 11 Oct 2018 09:51:07 +0000 (11:51 +0200)]
brcmfmac: Fix ccode from EFI nvram when necessary

In some cases the EFI-var stored nvram contains "ccode=ALL" or "ccode=XV"
to specify "worldwide" compatible settings, but these 2 ccode-s do not work
properly.

I've tested the different known "worldwide" ccode-s used in various nvram
sources with the latest firmwares from linux-firmware for various brcmfmac
models, here is a simplified (*) table with what each setting results in:

ALL: 12-14 disab, U-NII-1, U-NII-2 no-IR/radar, U-NII-3
XV:  12-14 no-IR, disables all 5G channels
XY:  12-13 enab, 14 disab, U-NII-1 enab, U-NII-2 no-IR/radar, U-NII-3 disab
X2:  12-13 no-IR, 14 dis, U-NII-1 no-IR, U-NII-2 no-IR/radar, U-NII-3 no-IR

Where 12,13,14 are 2.4G channels 12-14 and U-NII-1/2/3 are the 3 different
5G channel groups. no-IR is no-Initiate-Radiation, we will never send on
these channels without first having received valid wifi traffic there.

This immediately shows that both ALL and XV are not as worldwide as we want
them to be. ALL causes channels 12 and 13 to not be available and XV causes
all 5GHz channels to not be available. Also ALL unconditionally enables the
U-NII-1 and U-NII-3 5G groups, while we really should be using no-IR for
these.

This commit replace XV and ALL with X2, which allows usage of chan 12-13
and 5G channels, but only after receiving valid wifi traffic there first.

Note that this configure the firmware's channel limits, the kernels own
regulatory restrictions based on e.g. regulatory info received from the
access-point, will be applied on top of this.

This fixes channels 12+13 not working on the Asus T200TA and the Lenovo
Mixx 2 8 and 5G channels not working on the Asus T100HA.

This has been tested on the following models: Acer Iconia Tab8 w1-810,
Acer One 10, Asus T100CHI, Asus T100HA, Asus T100TA, Asus T200TA and a
Lenovo Mixx 2 8.

*) There are some exceptions to this table:
1) On really old firmware e.g. linux-firmware's 2011 brcmfmac4330-sdio.bin
   ALL really means all, unconditionally enabling everything
2) The exact meaning might be influenced by setting the regrev nvram var.
   Specifically using ccode=XV + regrev=1 on brcmfmac43241b4 leads to:
   12-14 no-ir, U-NII-1 no-ir, U-NII-2 no-ir/radar, U-NII-3 no-ir
   But only on the brcmfmac43241b4 and not on e.g. the brcmfmac43340

Tested-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: Add support for getting nvram contents from EFI variables
Hans de Goede [Thu, 11 Oct 2018 09:51:06 +0000 (11:51 +0200)]
brcmfmac: Add support for getting nvram contents from EFI variables

Various X86 laptops with a SDIO attached brcmfmac wifi chip, store the
nvram contents in a special EFI variable. This commit adds support for
getting nvram directly from this EFI variable, without the user needing
to manually copy it.

This makes Wifi / Bluetooth work out of the box on these devices instead of
requiring manual setup.

This has been tested on the following models: Acer Iconia Tab8 w1-810,
Acer One 10, Asus T100CHI, Asus T100HA, Asus T100TA, Asus T200TA and a
Lenovo Mixx 2 8.

Tested-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: Cleanup brcmf_fw_request_done()
Hans de Goede [Wed, 10 Oct 2018 11:01:03 +0000 (13:01 +0200)]
brcmfmac: Cleanup brcmf_fw_request_done()

The "cur" variable is now only used for a debug print and we already
print the same info from brcmf_fw_complete_request(), so the debug print
does not provide any extra info and we can remove it.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: Set board_type from DMI on x86 based machines
Hans de Goede [Wed, 10 Oct 2018 11:01:02 +0000 (13:01 +0200)]
brcmfmac: Set board_type from DMI on x86 based machines

For x86 based machines, set the board_type used for nvram file selection
based on the DMI sys-vendor and product-name strings.

Since on some models these strings are too generic, this commit also adds
a quirk table overriding the strings for models listed in that table.

The board_type setting is used to load the board-specific nvram file with
a board-specific name so that we can ship files for each supported board
in linux-firmware.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: Set board_type used for nvram file selection to machine-compatible
Hans de Goede [Wed, 10 Oct 2018 11:01:01 +0000 (13:01 +0200)]
brcmfmac: Set board_type used for nvram file selection to machine-compatible

For of/devicetree using machines, set the board_type used for nvram file
selection to the first string listed in the top-level's node compatible
string, aka the machine-compatible as used by of_machine_is_compatible().

The board_type setting is used to load the board-specific nvram file with
a board-specific name so that we can ship files for each supported board
in linux-firmware.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: Add support for first trying to get a board specific nvram file
Hans de Goede [Wed, 10 Oct 2018 11:01:00 +0000 (13:01 +0200)]
brcmfmac: Add support for first trying to get a board specific nvram file

The nvram files which some brcmfmac chips need are board-specific. To be
able to distribute these as part of linux-firmware, so that devices with
such a wifi chip will work OOTB, multiple (one per board) versions must
co-exist under /lib/firmware.

This commit adds support for callers of the brcmfmac/firmware.c code to
pass in a board_type parameter through the request structure.

If that parameter is set then the code will first try to load
chipmodel.board_type.txt before falling back to the old chipmodel.txt name.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: Remove recursion from firmware load error handling
Hans de Goede [Wed, 10 Oct 2018 11:00:59 +0000 (13:00 +0200)]
brcmfmac: Remove recursion from firmware load error handling

Before this commit brcmf_fw_request_done would call
brcmf_fw_request_next_item to load the next item, which on an error would
call brcmf_fw_request_done, which if the error is recoverable (*) will
then continue calling brcmf_fw_request_next_item for the next item again
which on an error will call brcmf_fw_request_done again...

This does not blow up because we only have a limited number of items so
we never recurse too deep. But the recursion is still quite ugly and
frankly is giving me a headache, so lets fix this.

This commit fixes this by removing brcmf_fw_request_next_item and by
making brcmf_fw_get_firmwares and brcmf_fw_request_done directly call
firmware_request_nowait resp. firmware_request themselves.

*) brcmf_fw_request_nvram_done fallback path succeeds or
   BRCMF_FW_REQF_OPTIONAL is set

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agobrcmfmac: Remove firmware-loading code duplication
Hans de Goede [Wed, 10 Oct 2018 11:00:58 +0000 (13:00 +0200)]
brcmfmac: Remove firmware-loading code duplication

brcmf_fw_request_next_item and brcmf_fw_request_done both have identical
code to complete the fw-request depending on the item-type.

This commit adds a new brcmf_fw_complete_request helper removing this code
duplication.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath9k: dynack: remove 'experimental' tag
Lorenzo Bianconi [Fri, 2 Nov 2018 20:49:59 +0000 (21:49 +0100)]
ath9k: dynack: remove 'experimental' tag

Remove experimental tag from dynack Kconfig entry since it has
been tested on outdoor 25Km links

Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath9k: dynack: make ewma estimation faster
Lorenzo Bianconi [Fri, 2 Nov 2018 20:49:58 +0000 (21:49 +0100)]
ath9k: dynack: make ewma estimation faster

In order to make propagation time estimation faster,
use current sample as ewma output value during 'late ack'
tracking

Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath9k: dynack: check da->enabled first in sampling routines
Lorenzo Bianconi [Fri, 2 Nov 2018 20:49:57 +0000 (21:49 +0100)]
ath9k: dynack: check da->enabled first in sampling routines

Check da->enabled flag first in ath_dynack_sample_tx_ts and
ath_dynack_sample_ack_ts routines in order to avoid useless
processing

Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath9k: dynack: move debug log after buffer increments
Lorenzo Bianconi [Fri, 2 Nov 2018 20:49:56 +0000 (21:49 +0100)]
ath9k: dynack: move debug log after buffer increments

Move debug log in ath_dynack_sample_tx_ts and ath_dynack_sample_ack_ts
after timestamp buffer head/tail increments in order to make debugging
more user friendly

Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath9k: dynack: use authentication messages for 'late' ack
Lorenzo Bianconi [Fri, 2 Nov 2018 20:49:55 +0000 (21:49 +0100)]
ath9k: dynack: use authentication messages for 'late' ack

In order to properly support dynack in ad-hoc mode running
wpa_supplicant, take into account authentication frames for
'late ack' detection. This patch has been tested on devices
mounted on offshore high-voltage stations connected through
~24Km link

Reported-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath10k: fix struct init whitespace in SNOC driver
Brian Norris [Fri, 2 Nov 2018 17:17:49 +0000 (10:17 -0700)]
ath10k: fix struct init whitespace in SNOC driver

Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath10k: don't assume this is a PCI dev in generic code
Brian Norris [Fri, 2 Nov 2018 17:17:48 +0000 (10:17 -0700)]
ath10k: don't assume this is a PCI dev in generic code

Not all devices using this driver are backed by a PCI device, so it's
not fair to assume 'to_pci_dev()' is valid. Fortunately, we only were
using this to convert straight back to a bare 'device', which means we
were only doing no-op pointer arithmetic, and not actually accessing
potentially out-of-bounds memory. But this is still bad practice.

Just use 'ar->dev' directly, since that's really all we're looking for.

Fixes: 209b2a68de76 ("ath10k: add platform regulatory domain support")
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath10k: assign 'n_cipher_suites' for WCN3990
Brian Norris [Fri, 2 Nov 2018 17:17:47 +0000 (10:17 -0700)]
ath10k: assign 'n_cipher_suites' for WCN3990

Commit 2ea9f12cefe4 ("ath10k: add new cipher suite support") added a new
n_cipher_suites HW param with a fallback value and a warning log. Commit
03a72288c546 ("ath10k: wmi: add hw params entry for wcn3990") later
added WCN3990 HW entries, but it missed the n_cipher_suites.

Rather than seeing this warning every boot

  ath10k_snoc 18800000.wifi: invalid hw_params.n_cipher_suites 0

let's provide the appropriate value.

Cc: Rakesh Pillai <pillair@qti.qualcomm.com>
Cc: Govind Singh <govinds@qti.qualcomm.com>
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath10k: ath10k_snoc_{read,write}32() should be static
Brian Norris [Fri, 2 Nov 2018 18:10:00 +0000 (11:10 -0700)]
ath10k: ath10k_snoc_{read,write}32() should be static

They're provided as callbacks in ath10k_hif_ops and should be accessed
that way, if needed outside of snoc.c, and anyway, they're currently
unused outside snoc.c.

Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath10k: snoc: fix unbalanced clock error handling
Brian Norris [Mon, 5 Nov 2018 12:35:22 +0000 (14:35 +0200)]
ath10k: snoc: fix unbalanced clock error handling

Similar to regulator error handling, we should only start tearing down
the 'i - 1' clock when clock 'i' fails to enable. Otherwise, we might
end up with an unbalanced clock, where we never successfully enabled the
clock, but we try to disable it anyway.

Fixes: a6a793f98786 ("ath10k: vote for hardware resources for WCN3990")
Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath10k: snoc: relax voltage requirements
Brian Norris [Mon, 5 Nov 2018 12:35:09 +0000 (14:35 +0200)]
ath10k: snoc: relax voltage requirements

I rarely see drivers specify precise voltage requirements like this, but
if we really have to...let's at least give a little wiggle room. Board
designs (and accompanying device trees) may not provide exactly the
voltage listed here, and we shouldn't fail to probe just because of
this.

Round these ranges down to the nearest volt, and provide a 0.05V margin.
The regulator should provide its own supported ranges, which will
helpfully intersect with these ranges.

I would just as well remove these ranges entirely, but if I understand
correctly, there's some reason that QCOM SoC's like to set zero /
non-zero voltages.

Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath10k: snoc: fix unabalanced regulator error handling
Brian Norris [Mon, 5 Nov 2018 12:34:57 +0000 (14:34 +0200)]
ath10k: snoc: fix unabalanced regulator error handling

If a regulator fails to set its voltage, we end up with an unbalanced
call to regulator_disable(), because the error path starts with the
current regulator (which was never enabled).

Factor out the "on" function to perform (and unwind if failed) a single
regulator at a time, and then main loop (ath10k_snoc_vreg_on()) can just
worry about unwinding the regulators that were already enabled.

It also helps to factor out the "off" function, to avoid repeating some
code here.

Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agoath10k: snoc: remove 'wcn3990' from generic resource handling
Brian Norris [Mon, 5 Nov 2018 12:34:46 +0000 (14:34 +0200)]
ath10k: snoc: remove 'wcn3990' from generic resource handling

ath10k_wcn3990_clk_info and ath10k_wcn3990_vreg_info are not
WCN3990-specific structures. They hold generic data. So don't name them
with wcn3990 specifics.

Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agowil6210: fix locking in wmi_call
Lior David [Wed, 31 Oct 2018 08:52:24 +0000 (10:52 +0200)]
wil6210: fix locking in wmi_call

Switch from spin_lock to spin_lock_irqsave, because
wmi_ev_lock is used inside interrupt handler.

Signed-off-by: Lior David <liord@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agowil6210: fix freeing of rx buffers in EDMA mode
Ahmad Masri [Wed, 31 Oct 2018 08:52:23 +0000 (10:52 +0200)]
wil6210: fix freeing of rx buffers in EDMA mode

After being associated with some EDMA rx traffic, upon "down" driver
doesn't free all skbs in the rx ring.
Modify wil_move_all_rx_buff_to_free_list to loop on active list of rx
buffers, unmap the physical memory and free the skb.

Signed-off-by: Ahmad Masri <amasri@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agowil6210: remove unnecessary alignment code from rx flow
Ahmad Masri [Wed, 31 Oct 2018 08:52:22 +0000 (10:52 +0200)]
wil6210: remove unnecessary alignment code from rx flow

Rx buffers in EDMA mode are initialized to 4 bytes aligned size.
Remove the unnecessary alignment code applied on rx buffer size.

Signed-off-by: Ahmad Masri <amasri@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agowil6210: fix RGF_CAF_ICR address for Talyn-MB
Maya Erez [Wed, 31 Oct 2018 08:52:20 +0000 (10:52 +0200)]
wil6210: fix RGF_CAF_ICR address for Talyn-MB

RGF_CAF_ICR register location has changed in Talyn-MB.
Add RGF_CAF_ICR_TALYN_MB to support the new address.

Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agowil6210: fix L2 RX status handling
Maya Erez [Wed, 31 Oct 2018 08:52:19 +0000 (10:52 +0200)]
wil6210: fix L2 RX status handling

L2 RX status errors should not be treated as a bitmap and the actual
error values should be checked.
Print L2 errors as wil_err_ratelimited for easier debugging
when such errors occurs.

Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
5 years agowil6210: fix debugfs memory access alignment
Ahmad Masri [Wed, 31 Oct 2018 08:52:18 +0000 (10:52 +0200)]
wil6210: fix debugfs memory access alignment

All wil6210 device memory access should be 4 bytes aligned. In io
blob wil6210 did not force alignment for read function, this caused
alignment fault on some platforms.
Fixing that by accessing all 4 lower bytes and return to host the
requested data.

Signed-off-by: Ahmad Masri <amasri@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>