Wakeup interrupt has been updating write pointers (indexes, actually) only
for tx queues 0-5.  This is adequate just for 3945, but inadequate for other
devices, all of which have more tx queues.  Now updating all tx/command queues,
so device can be aware of all new tx and host commands enqueued while
device was asleep.

This can potentially improve data traffic bandwidth and/or latency.
Signed-off-by: NBen Cahill <ben.m.cahill@intel.com>
Signed-off-by: NReinette Chatre <reinette.chatre@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Add/clarify comments and debug messages for interrupt used only for uCode load
Signed-off-by: NBen Cahill <ben.m.cahill@intel.com>
Signed-off-by: NReinette Chatre <reinette.chatre@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Invoke workaround to avoid instability in L0->L0S->L1 transition on PCIe bus.
Workaround disables L0S state so device moves directly from L0->L1.
Workaround needed on all devices since and including 4965; add to 6000/1000.

Describe bug and workaround better in comments.
Signed-off-by: NBen Cahill <ben.m.cahill@intel.com>
Signed-off-by: NReinette Chatre <reinette.chatre@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

To save power, don't run apm_ops.init() until needed at "up" time.

EEPROM (5000 and earlier devices) may be read without running apm_ops.init(),
but OTP reads (6000 and newer devices) require a powered-up chip.
Therefore, remove apm_ops.init() from the general path in XXXX_pci_probe(),
and call it only if device uses OTP.  Once done with OTP read, call
apm_ops.stop() to reset chip and save power until "up" time comes around.

NOTE:  This patch depends on removal of priv->lock from iwl_apm_stop();
lock does not get initialized until later in flow.  See patch
"remove unneeded locks from apm_stop()".
Signed-off-by: NBen Cahill <ben.m.cahill@intel.com>
Signed-off-by: NReinette Chatre <reinette.chatre@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Since priv->reg_lock was added to keep multi-access register manipulation
atomic, priv->lock protection is no longer needed.  Remove this from
iwl_apm_stop_master() and iwl_apm_stop().

Add warning of timeout when polling for busmaster disablement confirmation,
and some comments.

NOTE:  This is needed to enable use of apm_ops.stop() within iwl_eeprom_init();
priv->lock does not get initialized until after this flow.  See patch
"remove power-wasting calls to apm_ops.init()"
Signed-off-by: NBen Cahill <ben.m.cahill@intel.com>
Signed-off-by: NReinette Chatre <reinette.chatre@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

By setting the firmware version in wiphy it is possible to obtain this
information via ethtool. Some examples,

	    # ethtool -i wlan1
	    driver: iwlagn
	    version: 2.6.32-rc5-wl-56840-g26d8540
	    firmware-version: 228.57.2.23
	    bus-info: 0000:03:00.0
	    # ethtool -i wlan0
	    driver: iwl3945
	    version: 2.6.32-rc5-wl-56840-g26d8540
	    firmware-version: 15.28.2.8
	    bus-info: 0000:02:00.0
Signed-off-by: NReinette Chatre <reinette.chatre@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The multi-line code in this macro wasn't wrapped
in do {} while (0) so we cannot use it in an if()
branch safely in the future -- fix that.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This patch removes the internal command and data buffers that the GSPI driver
maintained and instead relies on the Libertas core to synchronize access
to the command and data ports as with the other interface drivers.  This
cleanup reduces the GSPI driver's memory footprint and should improve
performance by removing the need to copy to these internal buffers.
This also simplifies the bottom half of the interrupt handler.

This is an incremental cleanup: after removing the redundant buffers, we
can further improve the driver to use a threaded IRQ handler instead of
maintaining its own thread.  However I would like a few folks to test
the buffer removal first and make sure that I'm not introducing
regressions.

Tested on Blackfin BF527 with DMA disabled due to an issue with the SPI
host controller driver in the current bleeding-edge Blackfin kernel.  I
would appreciate it if someone with working DMA could test this patch
and provide feedback.
Signed-off-by: NAndrey Yurovsky <andrey@cozybit.com>
Tested-by: NGeorge Shore <george.shore@imgtec.com>
Acked-by: NDan Williams <dcbw@redhat.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This patch is adapted from the submission by Michael Buesch <mb@bu3sch.de>
for a bounce-buffer copying problem with b43.
Signed-off-by: NLarry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

It's not right to do something here when returning an
error, and hostapd should never have relied on it as
it only fixes up a small part of the problem anyway.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This variable is set once, and tested once.
However, the code path that can set it is
mutually exclusive with the code path that
tests it, so the test is always true. Thus
we also don't need to set it either and can
just remove the variable.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We drop nullfunc frames, but not qos-nullfunc frames,
even though those could be used for PS state control
as well.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

The check after the loop which checks whether the initialization timed-out
was wrong.  If the initialization would succeed exactly in the 20000th time
(the value set for INIT_LOOP), the driver would bail out and claim that
initialization failed.
Reported-by: NJuha Leppanen <juha_motorsportcom@luukku.com>
Signed-off-by: NLuciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: NJanne Ylalehto <janne.ylalehto@nokia.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

When mac80211 is asked to buffer multicast frames
in AP mode, it will not set the flag indicating
that the frames should be sent after the DTIM
beacon for those frames buffered in software. Fix
this little inconsistency by always setting that
flag in the buffering code path.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This value is unused by mac80211, because it was only
be used by wireless extensions, and turned out to not
be useful there because the quality value needs to be
comparable between scan results and the current value
which is impossible when the qual value is calculated
taking into account noise, for example.

Since it is unused anyway, this patch deprecates it
in the hope that drivers will remove their sometimes
quite expensive calculations of the value.

I'm open to actual uses of the value, but the best
way of using it seems to be what the Intel drivers do
which should probably be generalised if we have noise
values from the hardware.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Compared to ieee80211_beacon_get(), the new function
ieee80211_beacon_get_tim() returns information on the
location and length of the TIM IE, which some drivers
need in order to generate the TIM on the device. The
old function, ieee80211_beacon_get(), becomes a small
static inline wrapper around the new one to not break
all drivers.
Signed-off-by: NJohannes Berg <johannes@sipsolutions.net>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

airo_cs uses spy interfaces so it needs to select WEXT_SPY.

ERROR: "iw_handler_set_thrspy" [drivers/net/wireless/airo.ko] undefined!
ERROR: "wireless_spy_update" [drivers/net/wireless/airo.ko] undefined!
ERROR: "iw_handler_get_spy" [drivers/net/wireless/airo.ko] undefined!
ERROR: "iw_handler_get_thrspy" [drivers/net/wireless/airo.ko] undefined!
ERROR: "iw_handler_set_spy" [drivers/net/wireless/airo.ko] undefined!
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This fixed a BUG_ON in __skb_trim() when paged rx is used in
iwlwifi driver. Yes, the whole mac80211 stack doesn't support
paged SKB yet. But let's start the work slowly from small
code snippets.
Reported-and-tested-by: NAbhijeet Kolekar <abhijeet.kolekar@intel.com>
Signed-off-by: NZhu Yi <yi.zhu@intel.com>
Signed-off-by: NReinette Chatre <reinette.chatre@intel.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

To do this we reorder callers in order in which they are called.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Force bias is a fix for usage of AR5416 radios on the 2.4 GHz band
for orientation sensitivity. This was only partially implemented
with the ath9k_hw_decrease_chain_power() but first -- this was being
called for all chipsets which is not correct and second -- it was
missing the actual orientation code.

We now ensure to only enable force bias only for AR5416 and BUG_ON()
on other chipsets. Although ath9k_hw_decrease_chain_power() was enabled
for newer chipsets I suspect that it never ran unless the EEPROM had
ATH9K_ANT_FIXED_A or ATH9K_ANT_FIXED_B for antenna diversity.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This only differs between single-chip solutions and non single-chip
solutions.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This reorders phy.c routines in the order in the order in which they are used
and also moves the spur mitigation helpers for each type of chip into phy.c
as they are RF related.

This patch has no functional changes.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This avoids a branch on every channel change.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This allows us to later define a callback for both.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This clarifies this is only required for external radios.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This is calling an allocation and checking for it, simplify
this process in a macro.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

ath9k_hw_rfattach() was just calling a helper and this helper was
doing nothing for single-chip devices, and for non single-chip devices
it is just allocating memory for banks to program the RF registers
at a later time. Simplify this by having the hw initialization call
the rf bank allocation directly for external radios.

Also, propagate an -ENOMEM properly now upon failure.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We a huge branch for old hardware and nothing for newer
hardware. Instead of doing this just bail out early for
newer hardware.

This patch has no functional changes.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Document what we can about the RF analog front ends (radios)
of Atheros 802.11n devices. What should be clearer now is the
what we do for old pre AR5416 and AR5418 MAC based devices in
comparison to the modern sigle-chip 802.11n solutions.

All devices after AR9280 are single chip and require less
programming -- the RF registers no longer need to be initialized
as they all have the RF analog front end embedded together with
the MAC/BB; this includes the AR9271. Older devices such as the
ones with the AR5416 MACs (PCI) or AR5418 MACs (PCI-E) have an
external 2.4 GHz AR2133 radio or a dual band 2.4 GHz / 5 GHz
AR5133 radio. These external radios require additional programming
of the RF registers.

Clarify which parts are for what devices and which code is
shared. This patch has no functional changes.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We adjust the core clock for ar9271 to 117 MHz; this also
requires us to adjust the baud divider based on the targetted
baud rate.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This update the register initialization/reset values (aka initvals)
for ar9271 based on the last recommended values on 2009-06-04 by our
systems engineering team.

The changes account for:

  * Supporting ar9271 1.0 and ar9271 1.1 together, the difference
    is bb_spectral_scan_ena, for 1.0 we'll set this to 0x1.

  * Ensuring we get the correct noise floor values -115 ~ -118
    when we enable bb_enable_ant_div_lnadiv=0 and
    mc_tx_def_ant_sel=1. Previous to this we would get noise
    floor values in the range -50 ~ -80. To fix settings for
    the registers:

     - bb_ch1_xatten1_db
     - bb_ch1_xatten2_db
     - bb_ch1_xatten1_margin
     - bb_ch1_xatten2_margin
     - bb_ch1_gain_force
     - bb_ch1_xatten2_hyst_margin
     - bb_ch1_xatten1_hyst_margin
     - bb_ch1_max_oc_gain

  * 0x8120[2] mc_mic_new_location_enable is changed to 0x1. The MAC team
    suggest to set this value.

  * 0x9910[0] bb_spectral_scan_ena is changed to 0x0.
    For ar9271 1.1 we don't need to enable this bit.

Cc: Stephen Chen <Stephen.Chen@atheros.com>
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

With the WLAN_PRE80211 drivers moved to drivers/staging, this
distinction becomes unnecessary.
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Move the netwave driver to drivers/staging.  This is another pre-802.11
driver that has seen virtually no non-API-fixup activity in years, and
for which no active hardware is likely to still exist.  This driver
represents unnecessary ongoing maintenance for no clear benefit.

This patch brought to you by the "hacking" session at the 2009 Kernel
Summit in Tokyo, Japan...
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Move the wavelan driver to drivers/staging.  This is another pre-802.11
driver that has seen virtually no non-API-fixup activity in years, and
for which no active hardware is likely to still exist.  This driver
represents unnecessary ongoing maintenance for no clear benefit.

This patch brought to you by the "hacking" session at the 2009 Kernel
Summit in Tokyo, Japan...
Acked-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>