and rename it to ath9k_cmn_init_channels_rates.
sbands are move to ath_common as well.
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Survey stats such as channel busy time, rx busy time
and noise are collected when sw_scan starts for every
switched new channel. This happens till sw_scan stops.
All the collected stats are shared up when get_survey()
is called.
This implements support for ACS feature from Hostapd.
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Nishant Sarmukadam <nishants@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Implement sw_scan start and stop callbacks.
Reset BBP registers for channel, RX busy and
average noise stats which clear on read.
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Nishant Sarmukadam <nishants@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Adding AP firmware command to access BBP
regs. This will be used for retrieving
different counters CCA/RX needed for ACS
Signed-off-by: Yogesh Ashok Powar <yogeshp@marvell.com>
Signed-off-by: Nishant Sarmukadam <nishants@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Function ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan} have been
replaced with ieee80211_{channel_to_frequency, frequency_to_channel}.
There should be no users of the two functions now. So remove them.
Cc: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Zhao, Gang <gamerh2o@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Replace ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan} with more
generic ieee80211_{channel_to_frequency, frequency_to_channel}.
File <net/cfg80211.h> has already been included. It's safe to use
IEEE80211_BAND_2GHZ here.
This change is a preparation for the removal of function
ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan}.
Signed-off-by: Zhao, Gang <gamerh2o@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Replace ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan} with more
generic ieee80211_{channel_to_frequency, frequency_to_channel}.
Function ieee80211_frequency_to_channel() will return 0 if frequency
is not correctly provided, so no need to set channel to 0 explicitly.
Include <net/cfg80211.h> for the defination of IEEE80211_BAND_2GHZ.
Because <net/cfg80211.h> includes <linux/ieee80211.h>, so we can
replace <linux/ieee80211.h> with <net/cfg80211.h>.
This change is a preparation for the removal of function
ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan}.
Signed-off-by: Zhao, Gang <gamerh2o@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Replace ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan} with more
generic ieee80211_{channel_to_frequency, frequency_to_channel}.
File <net/cfg80211.h> has already been included. It's safe to use
IEEE80211_BAND_2GHZ here.
This change is a preparation for the removal of function
ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan}.
Cc: Jussi Kivilinna <jussi.kivilinna@iki.fi>
Signed-off-by: Zhao, Gang <gamerh2o@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Replace ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan} with more
generic ieee80211_{channel_to_frequency, frequency_to_channel}.
Include <net/cfg80211.h> for the defination of IEEE80211_BAND_2GHZ.
Because <net/cfg80211.h> includes <linux/ieee80211.h>, so we can
replace <linux/ieee80211.h> with <net/cfg80211.h>.
This change is a preparation for the removal of function
ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan}.
Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: Zhao, Gang <gamerh2o@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Replace ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan} with more
generic ieee80211_{channel_to_frequency, frequency_to_channel}.
Include <net/cfg80211.h> for the defination of IEEE80211_BAND_2GHZ.
Because <net/cfg80211.h> includes <linux/ieee80211.h>, so we can
replace <linux/ieee80211.h> with <net/cfg80211.h>.
This change is a preparation for the removal of function
ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan}.
Cc: Simon Kelley <simon@thekelleys.org.uk>
Signed-off-by: Zhao, Gang <gamerh2o@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Replace ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan} with more
generic ieee80211_{channel_to_frequency, frequency_to_channel}.
Include <net/cfg80211.h> for the defination of IEEE80211_BAND_2GHZ.
Because <net/cfg80211.h> includes <linux/ieee80211.h>, so we can
replace <linux/ieee80211.h> with <net/cfg80211.h>.
This change is a preparation for the removal of function
ieee80211_{dsss_chan_to_freq, freq_to_dsss_chan}.
Signed-off-by: Zhao, Gang <gamerh2o@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Certain HW options (TX packet retry count, CW configuration and
TX power configuration) can be specified in both the TX packet
descriptor and also into HW "global" registers.
The HW is thus configured to honour the global register or the
TX descriptor field depending by the case.
This patch adds few comments that hopefully clarify in which cases
the driver uses one method and in which cases it uses the other.
Signed-off-by: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
During driver initialization, some skbs are preallocated for RX.
Currenly if the allocation fails, the driver's allocation routine
exits immediatly but it will return zero (success) anyway.
In this way the driver will continue initialization with buggy
pointers around.
This patch makes the driver's allocation routine to return
an error value and to print a complaint message when skb allocation
fails.
In this way its caller will not go further, avoinding the driver to
successfully load, and preventing dereferencing buggy pointers.
An hint is thus printed about why the driver failed.
Signed-off-by: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
During initialization a number of RX skbs are allocated and mapped
for DMA.
Currently if pci_map_single() fails, it will result in passing to the
HW a wrong DMA address (to write to!).
This patch adds check for this condition and eventually causes the
driver not to initialize, avoiding at least dangerous DMAs.
Signed-off-by: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Currently the driver uses pci_iomap() but iounmap() is called in
the error path
Change to use pci_iounmap() instead.
Reported-by: Huqiu Liu <liuhq11@mails.tsinghua.edu.cn>
Signed-off-by: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Currently the contention window enable/disable HW flag is not
touched by the driver.
This patch explicitly set it to the correct value to make sure
contention window is enabled (AFAIK contention window must be
enabled in most (if not all) cases.
Signed-off-by: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Add device tree binding documentation for Texas Instrument's wl1251
wireless lan chip. For now only the SPI binding is documented.
Signed-off-by: Sebastian Reichel <sre@debian.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Add device tree support for the spi variant of wl1251.
Signed-off-by: Sebastian Reichel <sre@debian.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This patch adds support for requesting the regulator powering
the vio pin.
Signed-off-by: Sebastian Reichel <sre@debian.org>
Reviewed-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Move the power GPIO handling from the board code into
the driver. This is a dependency for device tree support.
Signed-off-by: Sebastian Reichel <sre@debian.org>
Reviewed-by: Pavel Machek <pavel@ucw.cz>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Move the wl1251 part of the wl12xx platform data structure into a new
structure specifically for wl1251. Change the platform data built-in
block and board files accordingly.
Signed-off-by: Luciano Coelho <coelho@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Reviewed-by: Felipe Balbi <balbi@ti.com>
Reviewed-by: Sebastian Reichel <sre@debian.org>
Reviewed-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
As many Surface Pro I & II users have found out, the mwifiex_usb
doesn't support usb autosuspend, and it has caused some system
stability issues.
Bug 69661 - mwifiex_usb on MS Surface Pro 1 is unstable
Bug 60815 - Interface hangs in mwifiex_usb
Bug 64111 - mwifiex_usb USB8797 crash failed to get signal
information
USB autosuspend get triggered when Surface Pro's AC power is
removed or powertop enables power saving on USB8797 device.
Driver's suspend handler is called here, but resume handler
won't be called until the AC power is put back on or powertop
disables power saving for USB8797.
We need to refactor the suspend/resume handlers to support
usb autosuspend properly. For now let's just remove it.
Cc: <stable@vger.kernel.org> # 3.5+
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The stop_scan_complete function was used as an intermediate step before
doing the actual connection creation. Since we're using hci_request
there's no reason to have this extra function around, i.e. we can simply
put both HCI commands into the same request.
The single task that the intermediate function had, i.e. indicating
discovery as stopped is now taken care of by a new
HCI_LE_SCAN_INTERRUPTED flag which allows us to do the discovery state
update when the stop scan command completes.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
The discovery process has a timer for disabling scanning, however
scanning might be disabled through other means too like the auto-connect
process. We should therefore ensure that the timer is never active
after sending a HCI command to disable scanning.
There was some existing code in stop_scan_complete trying to avoid the
timer when a connect request interrupts a discovery procedure, but the
other way around was not covered. This patch covers both scenarios by
canceling the timer as soon as we get a successful command complete for
the disabling HCI command.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Some devices may refuse to re-encrypt with the LTK if they haven't
received all our keys yet. This patch adds a 250ms delay before
attempting re-encryption with the LTK.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
It's not strictly speaking required to re-encrypt a link once we receive
an LTK since the connection is already encrypted with the STK. However,
re-encrypting with the LTK allows us to verify that we've received an
LTK that actually works.
This patch updates the SMP code to request encrypting with the LTK in
case we're in master role and waits until the key refresh complete event
before notifying user space of the distributed keys.
A new flag is also added for the SMP context to ensure that we
re-encryption only once in case of multiple calls to smp_distribute_keys.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
LE connection attempts do not have a controller side timeout in the same
way as BR/EDR has (in form of the page timeout). Since we always do
scanning before initiating connections the attempts are always expected
to succeed in some reasonable time.
This patch adds a timer which forces a cancellation of the connection
attempt within 20 seconds if it has not been successful by then. This
way we e.g. ensure that mgmt_pair_device times out eventually and gives
an error response.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
This patch adds defines for the initiator filter policy parameter values
of the HCI_LE_Create_Connection command. They will be used in a
subsequent patch to check whether we should have a timeout for the
connection attempt or not.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Now that we have nicely tracked values of the initiator and responder
address information we can pass that directly to the smp_c1 function
without worrying e.g. about who initiated the connection. This patch
updates the two places in smp.c to use the new variables.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
For SMP we need the local and remote addresses (and their types) that
were used to establish the connection. These may be different from the
Identity Addresses or even the current RPA. To guarantee that we have
this information available and it is correct track these values
separately from the very beginning of the connection.
For outgoing connections we set the values as soon as we get a
successful command status for HCI_LE_Create_Connection (for which the
patch adds a command status handler function) and for incoming
connections as soon as we get a LE Connection Complete HCI event.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
We shouldn't update the hci_conn state to BT_CONNECT until the moment
that we're ready to send the initiating HCI command for it. If the
connection has the BT_CONNECT state too early the code responsible for
updating the local random address may incorrectly think there's a
pending connection in progress and refuse to update the address.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Different controllers behave differently when HCI_Set_Random_Address is
called while they are advertising or have a HCI_LE_Create_Connection in
progress. Some take the newly written address into use for the pending
operation while others use the random address that we had at the time
that the operation started.
Due to this undefined behavior and for the fact that we want to reliably
determine the initiator address of all connections for the sake of SMP
it's best to simply prevent the random address update if we have these
problematic operations in progress.
This patch adds a set_random_addr() helper function for the use of
hci_update_random_address which contains the necessary checks for
advertising and ongoing LE connections.
One extra thing we need to do is to clear the HCI_ADVERTISING flag in
the enable_advertising() function before sending any commands. Since
re-enabling advertising happens by calling first disable_advertising()
and then enable_advertising() all while having the HCI_ADVERTISING flag
set. Clearing the flag lets the set_random_addr() function know that
it's safe to write a new address at least as far as advertising is
concerned.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
If SMP fails we should not leave any keys (LTKs or IRKs) hanging around
the internal lists. This patch adds the necessary code to
smp_chan_destroy to remove any keys we may have in case of pairing
failure.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
The random numbers in Bluetooth Low Energy are 64-bit numbers and should
also be little endian since the HCI specification is little endian.
Change the whole Low Energy pairing to use __le64 instead of a byte
array.
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
It doesn't make much sense to calculate the ring
size fill count because it already is memoized in
a separate variable.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
One of the premises was to guarantee serialized
completion handling for upper layers
(HTC/WMI/HTT). Since quite some time now it is no
longer necessary.
The other premise was to batch up tx/rx
completions to take advantage of hot caches.
However frame tx/rx completion indications come in
on a single pipe already so they are already
batched up. More meaningful batching is done in
HTT itself.
This means PCI completion is no longer necessary
to keep around. It just wastes memory, cycles and
SLOC.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
It is inefficient to grab irqsave spinlocks for
skb lists for each queue/dequeue action.
Using rx_ring.lock and tx_lock allows to use less
heavy bh spinlock functions and moving locking
upwards allows to toggle spinlocks less often.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
HTT Rx endpoint processes both frame rx
indications and frame tx completion indications.
Those completions typically come in batches and
may be mixed so it makes sense to defer processing
hoping to get a bunch of them and take advantage
of hot caches.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Going through full htc tx path for htt tx is a
waste of resources. By skipping it it's possible
to easily submit scatter-gather to the pci hif for
reduced host cpu load and improved performance.
The new approach uses dma pool to store the
following metadata for each tx request:
* msdu fragment list
* htc header
* htt tx command
The htt tx command contains a msdu prefetch.
Instead of copying it original mapped msdu address
is used to submit a second scatter-gather item to
hif to make a complete htt tx command.
The htt tx command itself hands over dma mapped
pointers to msdus and completion of the command
itself doesn't mean the frame has been sent and
can be unmapped/freed. This is why htc tx
completion is skipped for htt tx as all tx related
resources are freed upon htt tx completion
indication event (which also implicitly means htt
tx command itself was completed).
Since now each htt tx request effectively consists
of 2 copy engine items CE_HTT_H2T_MSG_SRC_NENTRIES
is updated to allow maximum of
TARGET_10X_NUM_MSDU_DESC msdus being queued. This
keeps the tx path resource management simple.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
PCI is capable of handling scatter-gather lists.
This can be used to avoid copying memory.
Change the name of the callback while at to
reflect its purpose.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
The flag wasn't used anymore. No need to keep it.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
There's no real benefit from using them. DMA-API
already provides debugging. Some skbuffs are
already mapped directly with DMA-API since wrapper
arguments were insufficient and extending them
would be pointless.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
If some of the cleanup commands caused by mgmt_set_powered(off) never
complete we should still force the adapter to be powered down. This is
rather easy to do since hdev->power_off is already a delayed work
struct. This patch schedules this delayed work if at least one HCI
command was sent by the cleanup procedure.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
When powering off and disconnecting devices we should also consider
connections which have not yet reached the BT_CONNECTED state. They may
not have a valid handle yet and simply sending a HCI_Disconnect will not
work.
This patch updates the code to either disconnect, cancel connection
creation or reject incoming connection creation based on the current
conn->state value as well as the link type in question.
When the power off procedure results in canceling connection attempts
instead of disconnecting connections we get a connection failed event
instead of a disconnection event. Therefore, we also need to have extra
code in the mgmt_connect_failed function to check if we should proceed
with the power off or not.
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
When the LE white list gets changed via HCI commands make sure that
the internal storage of the white list entries gets updated.
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
The current LE white list entries require storing in the HCI controller
structure. So provide a storage and access functions for it. In addition
export the current list via debugfs.
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
When starting up a controller make sure that all LE white list entries
are cleared. Normally the HCI Reset takes care of this. This is just
in case no HCI Reset has been executed.
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
