One chip family employs immediate Tx completion, where knowledge of
completed packets is given as part of the FW status. Another is only
notified of Tx completion via the FW status, and has to read the
completion status of the packets from a different location.

Implement the wl12xx tx completion as a delayed Tx completion.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Rates reported by HW can be different between chip families. Make the
rate-to-idx translation tables private per family and use them in a
common translation function. Add a global element to help determine
which rates are HW HT-rates.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

The number of spare Tx blocks must be changed when the GEM cipher is
engaged. Track set_key() operations to see if this is the case and
change the Tx HW spare block count accordingly. Set the number of spare
blocks for each operating mode from the low level driver.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Each chip family can have a different amount of Tx descriptors. These
are set on init.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

This storage is allocated in wlcore_alloc_hw and freed in free_hw. The
size of the storage is determined by the low-level driver.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Different chip families have the factory MAC address written in
different places.  Add a new hardware operation to read the MAC
address, if available.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

The top registers initialization is very specific to the actual
hardware used, even the way in which we read from and write to the top
registers varies from chip to chip.  This patch moves all top
registers initialization to wl12xx.  Also add a boot op for the wlcore
module to call at the right time and a few callbacks with the common
called to be called from the lower drivers boot operations.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

The PG version depends on the actual hardware.  This commit moves the
code used to read the PG version to the lower driver, by adding the
get_pg_ver hardware operation.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Move the code that identifies the chip ID and selects the appropriate
firmware to an operation implemented by the lower driver.  Also move
the quirks definitions into wlcore.h and rename to WLCORE_QUIRK_*.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Add register tables support in wlcore, add some new IO functions to
read and write to chip-specific register and data addresses.  Move
some common register values from wl12xx to wlcore and add the
registers table to wl12xx.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Add partition tables to wlcore, move and reorganize partition setting
functions.  Move wl12xx partition table to use the wlcore partition
table instead.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

In order to add chip-specific operations and prepare for future
elements that need to be set by the lower driver, move the wl1271
structure to the wlcore.h file and add an empty placeholder for the
operations structure.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

We need to set some parameters (eg. partition and register tables)
during probe of the lower driver, so split the probe function, leaving
most of it in wlcore, but moving the hw struct allocation to the lower
driver.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Create a new small wl12xx module that only contains the probe
functions and depends entirely on wlcore otherwise.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Rename the wl12xx driver directory to wlcore as an initial step
towards the split of the driver into wlcore and wl12xx.  We just
rename the directory first to keep git blame happy.
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Move wl12xx and wl1251 modules into a new drivers/net/wireless/ti
directory.  Add a TI WLAN Kconfig option and Makefile to support this
change.
Signed-off-by: NLuciano Coelho <coelho@ti.com>
Cc: John W. Linville <linville@tuxdriver.com>

Set the dwell times for sched scan according to the number
of probe requests which are going to be transmitted.
This should fix the too short dwell time problem which
prevented some of the probe requests from being transmitted
in cases of high number of SSIDs (10+) to be actively sched scanned.
However, in the common case of having up to 1-2 SSIDs that
require active scan, the dwell time would be kept to a minimum
which should conserve power. This is important as sched scan
also runs periodically while the host is suspended and there's
great importance to keep power consumption as low as possible.
Signed-off-by: NEyal Shapira <eyal@wizery.com>
[fixed a couple of new strict checkpatch warnings]
Signed-off-by: NLuciano Coelho <coelho@ti.com>

wlvif->probereq is zeroed when adding an interface but
the skb pointed to isn't freed when the interface is removed.
This would lead to a mem leak on every recovery.
Fix it by freeing the skb when removing the interface.
Signed-off-by: NEyal Shapira <eyal@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

The ap keys should be freed only when removing
ap role (otherwise, some arbitrary data might
get freed).
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

On the PandaBoard (omap_hsmmc + wl12xx_sdio) with DMA_API_DEBUG:

 WARNING: at lib/dma-debug.c:930 check_for_stack.part.8+0x7c/0xe0()
 omap_hsmmc omap_hsmmc.4: DMA-API: device driver maps memory fromstack
Signed-off-by: NMircea Gherzan <mgherzan@gmail.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

The iteration on the wlvif list in wl1271_op_resume/suspend was
perfomed before locking wl->mutex which would lead to a kernel
panic in case a recovery was queued at the same time
and would delete the wlvifs from the list.
Signed-off-by: NEyal Shapira <eyal@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

wl12xx sets the do_join flag (which later starts the
sta role) when the bssid was changed and the sta is
associated. However, this no longer happens after
the "mac80211: remove spurious BSSID change flag"
patch.

Fix it by setting the do_join flag on BSS_CHANGED_ASSOC
(for IBSS, do_join is already set on BSS_CHANGED_IBSS)
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Since some of the HT code pre-dates 802.11n-2009
some names are wrong. The one that bothers me most
is that "HT operation" is called "HT information"
in our code and that causes confusion.

Rename "HT information" to "HT operation" and also
the control_chan field to primary_chan to match
the name used in the spec.
Signed-off-by: NJohannes Berg <johannes.berg@intel.com>
Acked-by: NBing Zhao <bzhao@marvell.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Track freed FW blocks during Tx. If no blocks were freed during a
predefined timeout, initiate a HW recovery. This helps in situations
when the FW watchdog fails.

Don't trigger recovery during activities that can temporarily stop
Tx. This includes:
- scanning
- buffering packets for sleeping stations (AP role)
- ROC on any role
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Add the packet length to the tx debug print.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Add a flag to indicate we initiated a recovery work
on purpose, in order to avoid triggering BUG() (when
the bug_on_recovery module param was set).
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

We didn't update the beacon interval on association
(or on a change notification when working as sta),
so the default interval (100ms) was always used.

Update the beacon interval according to the bss_conf
before starting the sta role (on association).
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

return EBUSY on scan when there is any role
in ROC (not necessarily the one we are going
to use)
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Currently, CHANGE_CHANNEL indication while
associated is considered as roaming attempt.

However, with the new auth/assoc redesign,
we no longer have to handle this case,
so remove it.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

In some cases, the wrong vif type was set in the
change_interface callback (P2P_CLIENT/P2P_GO instead
of STA/AP)
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

wl12xx currently looks for AP authorization by registering
a netdev notifier and waiting for the IF_OPER_UP notification,
which is quite cumbersome.

Use the newly introduced sta_state callback (waiting
for assoc -> auth notification) instead, in order to
simplify it.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Implement sta_state callback instead of the
sta_add/remove callbacks.

Update the fw regarding peer state and ht caps
only after the station was authorized. Otherwise,
the fw might try establishing BA session before
the sta is authorized.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

On disassociation, check only whether the current vif
is idle, instead of checking whether the device is idle.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

device role is used for scanning and sending packets
before connection. however, since we don't need to
send packets before ibss creation, there is no need
to start the device on idle-off.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

This ensures the user won't encounter lag associated with getting in and
out of PSM when the card is in use.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Flush our Tx queues before suspending or changing the channel due to a
channel_switch element in the AP beacon.
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Before, the link was first freed (invalidating it in the map), and later
on vif removal, all valid wlvif-related links were reset. Since these
links were already invalid, we failed to reset them.
The bug was made worse by op_stop, which set the tx_queue_count to 0
arbitrarily. This resulted in a negative tx_queue_count in some scenarios.

Fix this by resetting the Tx-queues of a link when freeing it. Add a
WARN_ON and reset all link Tx-queues in op_stop, to avoid a negative
tx_queue_count.

[changed WARN_ON to WARN_ON_ONCE -- Luca]
Signed-off-by: NArik Nemtsov <arik@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

"ret" has already been checked at this point, and we don't need to check
it again.  This was left around from a previous patch 49d750ca
"wl12xx: 1281/1283 support - New radio structs and functions".
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Delete the global wl->vif (and the checks on it),
so multiple vifs could be added.
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>

Enter forced psm when changing fw, in order to make the
sta a bit more disconnection-persistent.
(DPM doesn't know about the incoming recovery, so it
won't enter psm by itself)
Signed-off-by: NEliad Peller <eliad@wizery.com>
Signed-off-by: NLuciano Coelho <coelho@ti.com>