* Handle all usec parameters in one function. It's much cleaner
 this way.
Signed-off-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

 * Prepare for half/quarter/turbo support, introduce a new
 ah_bwmode parameter and get rid of ah_turbo. Bwmode stands
 for "bandwidth mode" and can have 4 values, default (20MHz),
 turbo (40MHz), half rate (10MHz), and quarter rate (5MHz).
Signed-off-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

 * No functional changes

 * Clean up reset:
 Introduce init functions for each unit and call them instead
 of having everything inside ath5k_hw_reset (it's just c/p for
 now so nothing changes except calling order -I tested it with
 various cards and it's ok-)

 * Further cleanups:
 ofdm_timings belongs to phy.c
 rate_duration belongs to pcu.c
 clock functions are general and belong to reset.c (more to follow)

 * Reorder functions for better organization:
 We start with helpers and other functions follow in categories,
 init functions are last
Signed-off-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This patch moves code out from wireless drivers where two different
functions are defined in three code locations for the same purpose and
provides a common function to sign extend a 32-bit value.
Signed-off-by: NAndreas Herrmann <andreas.herrmann3@amd.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Signed-off-by: NFelix Fietkau <nbd@openwrt.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

As reported by Ryan Niemi, some bitmasks in the register definition for the PCU
Diagnostic register (DIAG_SW) were missing a zero at the end. While at it fix
some typos and add more comments.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Don't generate calibration errors messages when not needed.
Signed-off-by: NFabio Rossi <rossi.f@inwind.it>
Acked-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

If the symbol offset is 46, it will be counted in both
the third and fourth bytes of the mask, and in this
case the shift will be negative which can pollute
high order bits in the mask.  This may negatively impact
OFDM symbol detection.
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Fix some comments:
    s/transmition/transmission/
    s/puting/putting/
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This fixes a few misspellings, word repetitions, and some grammar
nits in ath5k comments.  No code changes.
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

In commit 39d5b2c8 "ath5k: update
AR5K_PHY_RESTART_DIV_GC values to match masks" i introduced a regression on PHY
chips older than AR5K_SREV_PHY_5413, which caused signal values to be about
10dB less that before. This patch reverts the AR5K_PHY_RESTART_DIV_GC values to
the same values which were effectively used before (without the bitmask
mistake). This brings signal levels back to normal on these PHY chips.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Collect all pieces concering the antenna switch table into one function.
Previously it was split up between ath5k_hw_reset() and
ath5k_hw_commit_eeprom_settings().

Also we need to set the antenna switch table when ath5k_hw_set_antenna_mode()
is called manually (by "iw phy0 antenna set", for example).

I'm not sure if we need to set the switchtable at the same place in
ath5k_hw_reset() as it was before - it is set later thru
ath5k_hw_set_antenna_mode() anyways - but i leave it there to avoid
problems(?).

Plus print switchtable registers in the debugfs file.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

#define AR5K_PHY_RESTART_DIV_GC               0x001c0000
is 3 bit wide.

The previous values of 0xc and 0x8 are 4bit wide and bigger than the mask.

Writing 0 and 1 to AR5K_PHY_RESTART_DIV_GC is consistent with the comments and
initvals we have in the HAL.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

If the channel is not set yet and we configure the antennas just store the
setting. It will be activated during the next reset, when the channel is set.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Now that we have ftrace, it is not needed any more.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Seperate noise floor calibration from other PHY calibration and move it to the
tasklet. This is the first step to more separation of different calibrations.

Also move out ath5k_hw_request_rfgain_probe(ah) so we have one clean function
for I/Q calibration on 5111x parts.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This patch removes from drivers/net/ all the unnecessary
return; statements that precede the last closing brace of
void functions.

It does not remove the returns that are immediately
preceded by a label as gcc doesn't like that.

It also does not remove null void functions with return.

Done via:
$ grep -rP --include=*.[ch] -l "return;\n}" net/ | \
  xargs perl -i -e 'local $/ ; while (<>) { s/\n[ \t\n]+return;\n}/\n}/g; print; }'

with some cleanups by hand.

Compile tested x86 allmodconfig only.
Signed-off-by: NJoe Perches <joe@perches.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Resolution of a merge conflict upstream accidentally removed a hunk of
"ath5k: IQ calibration for AR5211 is slightly different", so restore it.
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We check the bounds on pdadc once when correcting for
negative curves but not when we later copy values from
from the pdadc_tmp array, leading to a potential overrun.

Although we shouldn't hit this case in practice, let's
be consistent.
Reported-by: NDan Carpenter <error27@gmail.com>
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Acked-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

These channels aren't selectable anyway, but our calculations
for 2.5 mhz frequencies are incorrect.  The value is supposed to
be:

  (frequency - reference) * (10/25)

i.e., divide by 2.5, but we were instead doing:

  (10 * frequency - reference) / 25.

Additionally, the check for (frequency % 5 == 2) had an extra
subtraction that wasn't in madwifi HAL.
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Acked-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

It's not a phy related funtion; It has more to do with the interrupt handler
and tasklet scheduling, so it belongs to base.c.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Optimize ath5k_hw_calibration_poll() since it is called on every singe
interrupt.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

We don't need to generate a software interrupt (SWI) just to schedule a tasklet
- we can just schedule the tasklet directly.

Rename constants, names, etc to reflect the fact that we don't use SWI any more.

Also move the flag handling into the tasklet and prepare it to behave correctly
when there are multiple flags present.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Remove static variable ath5k_calinterval which was used as a constant. Use a
#define instead. Also we don't need ah_cal_intval.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h

percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files.  percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.

percpu.h -> slab.h dependency is about to be removed.  Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability.  As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.

  http://userweb.kernel.org/~tj/misc/slabh-sweep.py

The script does the followings.

* Scan files for gfp and slab usages and update includes such that
  only the necessary includes are there.  ie. if only gfp is used,
  gfp.h, if slab is used, slab.h.

* When the script inserts a new include, it looks at the include
  blocks and try to put the new include such that its order conforms
  to its surrounding.  It's put in the include block which contains
  core kernel includes, in the same order that the rest are ordered -
  alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
  doesn't seem to be any matching order.

* If the script can't find a place to put a new include (mostly
  because the file doesn't have fitting include block), it prints out
  an error message indicating which .h file needs to be added to the
  file.

The conversion was done in the following steps.

1. The initial automatic conversion of all .c files updated slightly
   over 4000 files, deleting around 700 includes and adding ~480 gfp.h
   and ~3000 slab.h inclusions.  The script emitted errors for ~400
   files.

2. Each error was manually checked.  Some didn't need the inclusion,
   some needed manual addition while adding it to implementation .h or
   embedding .c file was more appropriate for others.  This step added
   inclusions to around 150 files.

3. The script was run again and the output was compared to the edits
   from #2 to make sure no file was left behind.

4. Several build tests were done and a couple of problems were fixed.
   e.g. lib/decompress_*.c used malloc/free() wrappers around slab
   APIs requiring slab.h to be added manually.

5. The script was run on all .h files but without automatically
   editing them as sprinkling gfp.h and slab.h inclusions around .h
   files could easily lead to inclusion dependency hell.  Most gfp.h
   inclusion directives were ignored as stuff from gfp.h was usually
   wildly available and often used in preprocessor macros.  Each
   slab.h inclusion directive was examined and added manually as
   necessary.

6. percpu.h was updated not to include slab.h.

7. Build test were done on the following configurations and failures
   were fixed.  CONFIG_GCOV_KERNEL was turned off for all tests (as my
   distributed build env didn't work with gcov compiles) and a few
   more options had to be turned off depending on archs to make things
   build (like ipr on powerpc/64 which failed due to missing writeq).

   * x86 and x86_64 UP and SMP allmodconfig and a custom test config.
   * powerpc and powerpc64 SMP allmodconfig
   * sparc and sparc64 SMP allmodconfig
   * ia64 SMP allmodconfig
   * s390 SMP allmodconfig
   * alpha SMP allmodconfig
   * um on x86_64 SMP allmodconfig

8. percpu.h modifications were reverted so that it could be applied as
   a separate patch and serve as bisection point.

Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: NTejun Heo <tj@kernel.org>
Guess-its-ok-by: NChristoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>

This code was commented-out when it was added about a year ago and
remains unchanged -- seems as if we don't need it...
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

"ath5k: remove stale function declarations, make some functions static"
commented-out some unused functions.  This removes them.
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>
Acked-by: NBob Copeland <me@bobcopeland.com>

according to the HAL sources the calculation of the Q value is slightly
different for AR5211 chips.

i couldn't test this since IQ calibration never finishes on older parts. this
is a different problem...
Signed-off-by: NBruno Randolf <br1@einfach.org>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

save antenna settings and preserve across resets.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

I/Q calibration was completely broken, resulting in a high number of CRC errors
on received packets. before i could see around 10% to 20% CRC errors, with this
patch they are between 0% and 3%.

1.) the removal of the mask in commit "ath5k: Fix I/Q calibration
(f1cf2dbd)" resulted in no mask beeing used
when writing the I/Q values into the register. additional errors in the
calculation of the values (see 2.) resulted too high numbers, exceeding the
masks, so wrong values like 0xfffffffe were written. to be safe we should
always use the bitmask when writing parts of a register.

2.) using a (s32) cast for q_coff is a wrong conversion to signed, since we
convert to a signed value later by substracting 128. this resulted in too low
numbers for Q many times, which were limited to -16 by the boundary check later
on.

3.) checked everything against the HAL sources and took over comments and minor
optimizations from there.

4.) we can't use ENABLE_BITS when we want to write a number (the number can
contain zeros). also always write the correction values first and set ENABLE
bit last, like the HAL does.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Cc: stable@kernel.org
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

when using a fixed antenna we should use the antenna number in all tx
descriptors, otherwise the hardware will sometimes send the frame out on the
other antenna. it seems like the hardware does not always respect the default
antenna and diversity settings (esp.  AR5K_STA_ID1_DEFAULT_ANTENNA).

also i would like to note that antenna diversity does not always work correctly
on 5414 (at least) when only one antenna is connected: for example all frames
might be received on antenna A but still the HW tries to send on antenna B some
times, causing packet loss.

this is both verified with the antenna statistics output of the previous patch
and a spectrum analyzer.
Signed-off-by: NBruno Randolf <br1@einfach.org>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

ath5k_hw_register_timeout() was duplicated between phy.c and reset.c.
Since it is too big and too much used to be an inline function, move it
away from the ath5k.h header into reset.c.  Remove _ATH5K_RESET and
_ATH5K_PHY defines.
Signed-off-by: NPavel Roskin <proski@gnu.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

Remove all unnecessary function declarations from ath5k.h.  Comment out
unused functions.  Remove ath5k_hw_get_tsf32(), which is too trivial to
be commented out.  Make functions static if suggested by sparse.  Make
ath5k_pm_ops static.
Signed-off-by: NPavel Roskin <proski@gnu.org>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

That is "success", "unknown", "through", "performance", "[re|un]mapping"
, "access", "default", "reasonable", "[con]currently", "temperature"
, "channel", "[un]used", "application", "example","hierarchy", "therefore"
, "[over|under]flow", "contiguous", "threshold", "enough" and others.
Signed-off-by: NAndré Goddard Rosa <andre.goddard@gmail.com>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>

The sign of correction coefficients was lost in the calculations, which
caused high packetloss in 802.11a mode after the results were applied.
Fixed by removing unneccesary and broken AND with a bit mask.
Signed-off-by: NLukas Turek <8an@praha12.net>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

As a holdover from earlier code when we used to set
the power limit to '0' after a reset to configure the
default transmit power, ath5k interprets txpower=0 as
12.5 dBm.  Fix that by just passing 0 through.

This fixes http://bugzilla.kernel.org/show_bug.cgi?id=14567

Cc: stable@kernel.org
Reported-by: NDaniel Folkers <daniel.folkers@task24.nl>
Tested-by: NDaniel Folkers <daniel.folkers@task24.nl>
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This updates ath5k to calibrate the noise floor similar to the
way it is done in the madwifi hal and ath9k.  Of note:

- we start NF measurement at the same time as AGC calibration,
  but do not actually read the value until the periodic (long)
  calibration
- we keep a history of the last few values read and write the
  median back to the hardware for CCA
- we do not complain if NF calibration isn't complete, instead
  we keep the last read value.
Signed-off-by: NBob Copeland <me@bobcopeland.com>
Acked-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

This moves the shared regulatory structure into the
common structure. We will use this ongoing for common
data.
Signed-off-by: NLuis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

* Get rid of calibration timer, instead use a software interrupt
  to schedule the calibration tasklet.

 a) We don't need a timer for this, there is no need for accuracy
   even with round_jiffies i think this is a waste of resources.
   Also we don't need to run calibration if we are idle (no
   interrupts).

 b) When we add ANI support we 'll just extend the poll function
   and calibration tasklet and handle all periodic phy calibration
   on one place (much cleaner).

 c) Having calibration on a tasklet is better since during calibration
   we can't transmit or receive (antennas are detached to measure
   noise floor), previously calibration could run in parallel with
   tx/rx and interfere (packet loss).

 v2: kill tasklet on stop_hw, stop/wake queues
 v3: use time_is_before_eq_jiffies to compare timestamp with current
     time
Signed-off-by: NNick Kossifidis <mickflemm@gmail.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>

* Set correct xpd curve indices for high/low gain curves during
   rfbuffer setup on RF5112B with both calibration curves available.

 * Don't return zero min power when we have the same pcdac value
   twice because it breaks interpolation. Instead return the right
   x barrier as we do when we have equal power levels for 2 different
   pcdac values.
Signed-off-by: NNick Kossifidis <mickflemm@gmail.com>
Acked-by: NBob Copeland <me@bobcopeland.com>
Signed-off-by: NJohn W. Linville <linville@tuxdriver.com>