In order to reconfigure the correct reg domain on
recovery, we have to save the current configuration
before clearing it (wl->reg_ch_conf_pending is
considered before configuring a new regdomain).
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

wlcore configures different dwell times according to number
of active interfaces (in order to prevent hurting VO during
scan).

However, determining active vif only according to
bss_conf->idle is not explicit enough, and might result
in non-started vifs being counted as started as well
(e.g. unassociated sta during sta).

Fix it by checking for explicit conditions according
to the vif type.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Make sure the FW is awake when entering recovery. This is useful for
reading the FW panic log and also some FW registers giving us crash
report stats.

We must do this before interrupts are disabled since we rely on an
interrupt to complete the wakeup.

If the wakeup fails, continue recovery normally. All read/writes will be
blocked and won't do any harm.
Signed-off-by: NBarak Bercovitz <barak@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

The FW panic log is read during recovery work.
It has to be stopped properly before reading. otherwise,
inconsistent data might be read which cause the driver
to freeze.

__wlcore_cmd_send has to work for the special case of
CMD_STOP_FWLOGGER, while in recovery, in order to stop
the fw log before it is read.
Signed-off-by: NBarak Bercovitz <barak@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Reading FW panic log is only supported in host mode (SDIO)
Signed-off-by: NBarak Bercovitz <barak@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

FW memory block size and FW log end marker parameters
are added to wl structure and are initialized per
chip architecture.

convert_hwaddr hw operation is added to convert chip
dependent FW internal address.

Copy from FW log is also simplified to copy the entire
memory block as FW logger utility is repsponsible
for parsing of FW log content.
Signed-off-by: NIgal Chernobelsky <igalc@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

number of fwlog mem_blocks can be configured using module param.
this is a fw debug feature: in case a large fw log data is busrted during
a short period of time, the memory get filled and data is lost.
this allows us to dynamicly set the fw log mem_block usage, although
configuring more mem_block for logger comes at the expense of TP.
Signed-off-by: NYair Shapira <yair.shapira@ti.com>
Signed-off-by: NIdo Reis <idor@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Change interrogate command prototype to have command size
and returned buffer length.
This fixes the issue when command parameters are needed to
be passed to FW in addition to acx header as in the case with
get RSSI command, where role_id has to be passed.
Signed-off-by: NIgal Chernobelsky <igalc@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Add new ap_event_mask field, to indicate events that
should be unmasked only when there's an ap interface.

This is done in order to avoid spurious wakeups
when we don't care about the incoming event anyway.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Print current active channel/s and a role type for
that channel in the driver_state debugfs.
Signed-off-by: NVictor Goldenshtein <victorg@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

This keeps the kernel happy when using edge-irqs and requesting a
threaded irq.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Channels 52-64 were mapped incorrectly.

Refactor and document wlcore_get_reg_conf_ch_idx() in
order to make it clear what's going on there.

While on it, fix the return value check to consider
0 as a valid return value as well (indicates channel 1).
Reported-by: NYaniv Machani <yanivma@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NIdo Reis <idor@ti.com>
Signed-off-by: NVictor Goldenshtein <victorg@ti.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

The fw doesn't support channels 7,9,11 in 5ghz band,
so don't advertise supporting them.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

wlvif could be passed as NULL from the wlcore_tx_work_locked()
to the wl1271_prepare_tx_frame() and to wl1271_skb_queue_head()
functions. This may lead to a Kernel panic, fix this by
validating that wlvif != NULL.
Signed-off-by: NVictor Goldenshtein <victorg@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Remove scan debug dumps which are rarely used.
Make scan debug prints more clear and short.
Signed-off-by: NVictor Goldenshtein <victorg@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

We need some stuff done on idle change, most notably we have to stop
sched-scanning. Take care of this by reintroducing idle handling.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

We now disable elp sleep during plt mode to allow normal operation of
plt tools such as calibrator.

Having elp_sleep enabled during plt mode is actually not required and
in fact it disrupt plt operations such as rx statistics etc.
Signed-off-by: NYair Shapira <yair.shapira@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Under this mode the chip is powered on including sdio
but no FW is downloaded and run, interrupts are not enabled, etc...

This mode is intended to allow RTTT to bridge sdio as a transport
to the chip.

Driver only provides sdio access using the dev_mem debugfs file.

Some fixes done to the code that ensures that PLT mode and normal
driver power mode (ifconfig/add_interface) are mutually excluded.
Signed-off-by: NYair Shapira <yair.shapira@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Start a ROC on the AP channel beforing sending the authentication reply
to a connecting STA. This ROC is held up to 1 second via a timer. If the
station is authorized and added by mac80211, the ROC is extended until
the station is fully authorized.
We make sure not to ROC twice when several stations are connecting in
parallel and to only release the ROC when both the pending-reply timer
and the STA-state callbacks do not require it.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@intel.com>

Use the wrapper function for retrieving the platform data instead of
accessing dev->platform_data directly. This is a cosmetic change
to make the code simpler and enhance the readability.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

After the last architecture switched to generic hard irqs the config
options HAVE_GENERIC_HARDIRQS & GENERIC_HARDIRQS and the related code
for !CONFIG_GENERIC_HARDIRQS can be removed.
Signed-off-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>

mac80211 currently sets WIPHY_FLAG_SUPPORTS_SCHED_SCAN based on whether
the start_sched_scan operation is supported or not, but that will not
be correct for all drivers, we're adding scheduled scan to the iwlmvm
driver but it depends on firmware support.

Therefore, move setting WIPHY_FLAG_SUPPORTS_SCHED_SCAN into the drivers
so that they can control it regardless of implementing the operation.

This currently only affects the TI drivers since they're the only ones
implementing scheduled scan (in a mac80211 driver.)
Acked-by: NLuciano Coelho <luca@coelho.fi>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Pass the wdev from cfg80211 on to the driver as the vif
if given and it's valid for the driver.
Signed-off-by: NDavid Spinadel <david.spinadel@intel.com>
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Currently packet patterns and it's enum/structures are used only
for WoWLAN feature. As we intend to reuse them for new feature
packet coalesce, they are renamed in this patch.

Older names are kept for backward compatibility purpose.
Signed-off-by: NAmitkumar Karwar <akarwar@marvell.com>
Signed-off-by: NBing Zhao <bzhao@marvell.com>
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Removing some boilerplate by using module_spi_driver instead of calling
register and unregister in the otherwise empty init/exit functions
Signed-off-by: NPeter Huewe <peterhuewe@gmx.de>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

The fw_status wasn't zeroed during allocation, resulting
in uninitialized var usage, and finally causing AP
traffic stop after recovery.

The wrong value in fw_status_2->counters.tx_lnk_free_pkts
led to a bad lnk->allocated_pkts calculation in
wlcore_fw_status(), causing wl18xx_lnk_low_prio() to return
FALSE (lnk->allocated_pkts > thold).
This eventually blocked the link in wlcore_tx_work_locked(),
as wl1271_skb_dequeue() continuously returned NULL.

Fix it by zeroing wl->fw_status_1/2 during allocation.
Signed-off-by: NVictor Goldenshtein <victorg@ti.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

u32 can be incorrect (too small) for some architectures.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Otherwise, if the work is pending, we might get
a bad dereference after the interface is removed.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

When associating to an AP with WEP set the
default key upon association by implementing
the set_deafult_key_idx op.

Fixes auto-arp sent with wrong key_idx bug.
Signed-off-by: NYoni Divinsky <yoni.divinsky@ti.com>
Signed-off-by: NEyal Shapira <eyal@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

The current code configures the peer caps only on BSS_CHANGED_HT
notification. However, we have to configure the peer caps
(and rates) even when HT is not enabled. Otherwise, the fw
continues working with low rates.

Configure the peer caps when sta_exists is true (i.e. when
we extracted the sta rates, e.g. on association).
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Instead of doing all the sysfs file handling in the main file, move it
to a new sysfs source file to reduce the amount of code in a single
file.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Remove unnecessary includes; remove duplicate and useless defines; fix
copyright notice and remove some unnecessary line breaks.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

There's no need to take up the space for devices that don't
support WoWLAN, and most drivers can even make the support
data static const (except where it's modified at runtime.)
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Drivers that don't use chanctxes cannot perform VHT association because
they still use a "backward compatibility" pair of {ieee80211_channel,
nl80211_channel_type} in ieee80211_conf and ieee80211_local.
Signed-off-by: NKarl Beldan <karl.beldan@rivierawaves.com>
[fix kernel-doc]
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>

Spin locks and completions are expensive in hard IRQ context and cause
problems with RT kernels.  In RT kernels, both spin locks and
completions can schedule(), so we can't use them in hard irq context.

Move handling code into the irq thread function to avoid that.
Reported-by: NGregoire Gentil <gregoire@alwaysinnovating.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

We were printing out all the hex dumps regardless of whether dynamic
debugging was enabled or not.  Now that print_hex_dump_debug() has
been implemented, we can use that instead.
Reported-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Don't ignore dummy packets when our queues are empty. This causes dummy
packets never to be sent when traffic is not suspended by FW thresholds,
which happens only in high Tx throughput situations. This may hurt Rx
performance.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

The wlcore state was checked without the mutex being taken. This leads
to WARN_ONs sometimes if a notification arrives when the driver
is on, but the mutex is only taken after it is off. This usually
happens if stopping the driver while connected to a network.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Save the sequence number of the broadcast AP link in the wlvif. For each
connected station, save the sequence number in the drv_priv part of
ieee80211_sta. Use the saved numbers on recovery/resume, with the
obligatory increment on recovery.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>