staging: comedi: s626: remove 'set_int_src' callback from encoder private data

There are two functions used for the 'set_int_src' callback, One is used for
the channel 0-2 encoders and the other for the channel 3-5 encoders.

Refactor the two callbacks into a single s626_set_int_src() function and use
the encoder channel number to handle the differences.

Remove the then unnecessary 'set_int_src' member and just call s626_set_int_src()
drectly.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: remove 'get_load_trig' callback from encoder private data

There are two functions used for the 'get_load_trig' callback. One is used for
the channel 0-2 encoders and the other for the channel 3-5 encoders.

Refactor the two callbacks into a single s626_get_load_trig() function and use
the encoder channel number to handle the differences.

Remove the then unnecessary 'get_enable' member from the encoder private data.

The 'get_load_trig' callbacks were not being used by the driver. For now block
the s626_get_load_trig() function with '#ifdef unused' to prevent a compiler
warning.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: remove 'set_load_trig' callback from encoder private data

There are two functions used for the 'set_load_trig' callback, One is used for
the channel 0-2 encoders and the other for the channel 3-5 encoders.

Refactor the two callbacks into a single s626_set_load_trig() function and use
the encoder channel number to handle the differences.

Remove the then unnecessary 'set_load_trig' member and just call s626_set_load_trig()
drectly.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: remove 'get_enable' callback from encoder private data

There are two functions used for the 'get_enable' callback, s626_get_enable_a()
function is used for the channel 0-2 encoders and s626_get_enable_b() is used
for the channel 3-5 encoders.

Refactor the two callbacks into a single s626_get_enable() function and use the
encoder channel number to handle the differenced.

Remove the then unnecessary 'get_enable' member and just call s626_set_enable()
directly.

The 'get_enable' callbacks were not being used by the driver. For now block the
s626_get_enable() function with '#ifdef unused' to prevent a compiler warning.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: remove 'set_enable' callback from encoder private data

There are two functions used for the 'set_enable' callback, s626_set_enable_a()
function is used for the channel 0-2 encoders and s626_set_enable_b() is used
for the channel 3-5 encoders.

Refactor the two callbacks into a single s626_set_enable() function and use the
encoder channel number to handle the differenced.

Remove the then unnecessary 'set_enable' member and just call s626_set_enable()
directly.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: remove my_latch_lsw from encoder private data

This member of the encoder private data can be calculated based on
the comedi channel stored in the 'chan' member of the private data.
Do that instead and remove the unnecessary member.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: define S626_LP_CNTR* registers based on channel number

Redefining the Counter Preload/Latch registers as a macro that calculates
the register offset based on the comedi channel number.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: remove my_cr[ab] from encoder private data

These members of the encoder private data can be calculated based on
the comedi channel stored in the 'chan' member of the private data.
Do that instead and remove the unnecessary members.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: add comedi "chan" number to struct s626_enc_info

Add the comedi "chan" number to the encoder private data so we can then
remove some of the members of the private data.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: s626: define CRA/CRB registers based on channel number

Start cleaning up the 'struct s626_enc_info' by redefining the my_cra
and my_crb members based on the comedi channel number.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_1564: move apci1564_interrupt() into addi_apci_1564.c

On moving the function into the driver proper, also check the device is
asserting the shared interrupt line.

This patch also fixes the interrupt handling for the digital input
change-of-state interrupts.

Signed-off-by: Chase Southwood <chase.southwood@gmail.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Cc: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_1564: add Change-of-State interrupt subdevice and required functions

This board supports an interrupt that can be generated by an AND/OR
combination of 16 of the input channels.

Create a separate subdevice to handle this interrupt.

The apci1564_di_config() function is used to configure which
inputs are used to generate the interrupt. Currently this function
is broken since it does not follow the comedi API for insn_config
functions. Fix this function by implementing the config instruction
INSN_CONFIG_DIGITAL_TRIG.

Add the remaining subdevice operations necessary for the interrupt
subdevice to support async commands.

Signed-off-by: Chase Southwood <chase.southwood@gmail.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Cc: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_1564: introduce apci1564_private struct

The addi_private struct defined in addi-data/addi_common.h is very bloated
and contains many fields which addi_apci_1564 does not require.  In the
interest of eventually removing this driver's dependency on
addi_common.h, we can create a private data struct specifically for
addi_apci_1564 containing only the fields it will actually use.

Signed-off-by: Chase Southwood <chase.southwood@gmail.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Cc: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_1564: remove unused static variables

The global variables ui_InterruptStatus_1564 and ui_InterruptData are both
set but never used.  Just remove them from the driver.

Signed-off-by: Chase Southwood <chase.southwood@gmail.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Cc: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: addi_apci_1564: remove use of devpriv->b_OutputMemoryStatus

This member of the private data struct is only set at one location in the
entire driver, and then never even used for anything.  Let's just remove
its use.

Signed-off-by: Chase Southwood <chase.southwood@gmail.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Cc: H Hartley Sweeeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: automatically handle D/A data format

The DT2821 series board have jumpers that set the output range for
the two Analog Output channels. The range_table for the Analog Output
subdevice provides all possible ranges to the user. When a unipolar
range is selected the board expects the data to be in a straight
binary format. When a bipolar range is select the data should be in
two's complement format.

Currently, the user passes some configuration options when attaching
to the driver to select the data format for each channel. If the
user does not pass the config options, the data format is assumed to
be straight binary.

The Analog Output subdevice now has a range_table that provides the
user will all possible ranges. Use the range information to determine
if the data values need to be munged into two's complememnt values
and deprecate the config options.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: fix bug where DIO direction is reset to inputs

The direction of the 16 DIO channels is controlled with two bits in the
DACSR register. The HBOE bit controls the direction of channels 15-8 and
the LBOE bit channels 7-0.

The DACSR register is also used to control the Analog Output async command
support. Currently, starting or canceling an Analog Output command would
also reset the DIO directions so that all channels were inputs.

Fix the driver so that the DIO direction is not changed when starting or
canceling an Analog Output async command.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: tidy up the register map and bit defines

For aesthetics, rename the register map defines so that the bit defines
are associated with the registers. Also, convert all the bit defines into
bit-shifts to make them clear.

Remove all the verbose comments. All the registers and bits are named
according to the User Manual for the DT2821 Series, Document UM-05073-G.
That manual can be checked if necessary for the verbose information.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: avoid calculating the timer divisor multiple times

A common timer is used for analog input and output async commands. The
(*do_cmdtest) for both subdevices calculates the divisor as part of
Step 4 when validating the trigger arguments. The divisor is calculated
again in the (*do_cmd) for both subdevices in order to set the timer.

The comedi core only calls the (*do_cmd) is called after a successful
(*do_cmdtest). Save the divisor from the (*do_cmdtest) in the private
data and use that value in the (*do_cmd).

The extra check of the cmd->convert_arg in dt282x_ai_cmd() is not necessary.
The convert_arg was already checked in the (*do_cmdtest), and it's not used
in the (*do_cmd).

Tidy up dt282x_ns_to_timer(), the parameters are all unsigned int's and the
mask of the 'flags' can be moved here to simplify the callers.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: update the MODULE_DESCRIPTION

Change the MODULE_DESCRIPTION to something more useful than the
generic "Comedi low-level driver".

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: tidy up multi-line comments

Tidy up the multi-line comments at the beginning of the file to follow
the CodingStyle.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: cleanup analog output range table

The DT2821 series boards do not have programmable analog output ranges.
The range is set, per-channel, with jumpers on the board.

Currently the user has to pass some configuration options when attaching
to the driver in order to set the range for each channel. If they do not
pass the range options they will default to 0. If they pass an invalid
range option they will also default to 0. The 0 range option is bipolar
10V. If the user passes a range option that does not match the hardware
setting it doesn't cause any problems in the driver it just causes the
user space comedilib library to incorrectly convert the hardware values
to physical values.

Make the attach a bit simpler by deprecating these options and providing
a range table with all the valid ranges for the analog outputs. The user
can then select the correct range when they use the comedilib functions.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: fix Analog Input se/diff channel configuration

Most of the boards supported by this driver can do either single-ended or
differential Analog Input but there are a couple that can only do one or
the other. On the boards that can do both the type of input is set with
jumpers on the board.

The 'it->options[4]' passed by the user during the board (*attach) is used
to determine how the Analog Inputs are configured. This is used to set
the subdevice flags and number of channels correctly.

Fix the check of this option so that it only applies when the board can
actually do differential input. Also, default to differential inputs if
that is the only mode the board supports.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: tidy up Analog Input subdevice init

For aesthetics, add some whitespace to the Analog Input subdevice init.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: tidy up Analog Output subdevice init

For aesthetics, add some whitespace to the Analog Output subdevice init.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: remove the 'opt_*' enum

This enum is used to index the 'options' values that are passed
by the user attaches to the driver. These options are identified
in the comedi comment block.

For aesthetics, remove the enum and just open code the index values.

Refactor the Analog Input subdevice init so that the it->options[4],
indicating differential inputs, does not have to be checked multiple
times.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: factor board initialization out of (*attach)

For aesthetics, factor the code that initializes the then checks to see
if it is actually a DT2821 series board out of the (*attach).

For clarity, remove the MASK and VAL defines used to check the board and
jsut open code the values.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: fix dt282x_ao_insn_read()

The comedi core expects the (*insn_read) functions to return 'insn->n'
samples. Fix this function to work like the core expects.

For aesthetics, rename the private data 'ao' member to clarify its
use. Also, remove the unnecessary comment.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: tidy up dt282x_ai_dma_interrupt()

The main interrupt handler already has the comedi_subdevice pointer
needed in this function. Pass that subdevice pointer so that it does
not need to be fetched as a local variable.

Tidy up the function a bit.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: tidy up dt282x_ao_dma_interrupt()

The main interrupt handler already has the comedi_subdevice pointer
needed in this function. Pass that subdevice pointer so that it does
not need to be fetched as a local variable.

Tidy up the function a bit.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: remove sanity checks of 's->async->prealloc_buf'

The comedi core sanity checks that the subdevice has a buffer allocated
before allowing an async command to start. The helper functions in comedi_buf
will also sanity check the buffer as a side effect of the 'alloc' helpers
with reading or writing data for the buffer.

Remove the unnecessary sanity checks in the ai/ao dma interrupt handlers.
This will allow the handlers to disable the dma and then fail when they
try to read/write the buffer. Currently dma is left enabled due to the
sanity checks.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: tidy up Digital I/O subdevice init

For aesthetics, add some whitespace to the Digital I/O subdevice init.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: rename the prep_{ai,ao}_dma() functions

Rename these functions so they have namespace associated with the driver.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: remove unused chan_to_DAC() macro

This macro is not used in the driver. Remove it.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: convert private data flags to bit-fields

Convert the '2scomp' flags in the private data into bit-fields to save
a bit of space.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: remove DT2821_SIZE define

This define is only used in the comedi_request_region() call to specify
the size of the I/O region. Remove it and just open code the value.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: IRQ and DMA are needed for async command support

This driver requires an IRQ and 2 DMA channels in order to execute async
commands.

Modify the (*attach) so that command support is only hooked up if
DMA is available. This allows removing the extra checks in the
command support functions.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: remove 'volatile' on private data variables

As indicated by checkpatch.pl, "WARNING: Use of volatile is usually
wrong: ...". The variables in the private data that are marked
volatile don't need to be. Remove the volatile.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: fix dt282x_ao_insn_write()

The (*insn_write) functions are expected to write 'insn->n' samples to the
hardware. Fix this function so it works like the comedi core expects.

The comedi core sanity checks that the samples are within the 'maxdata' range
of the subdevice before calling the (*insn_write) so this function does not
need to mask each sample by 's->maxdata'.

Also, the wrong '*2scomp' flag in the private data was being checked to see
if the sample needs to be munged before being written. Fix this.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: convert 'dabits' boardinfo to 'ao_maxdata'

This member of the boardinfo is the number of bits in the analog output data.
Convert this member into the subdevice 'maxdata', (1 << board->dabits) - 1,
to avoid needing to do the calculation.

Use the subdevice 'maxdata' in the driver instead of using the boardinfo and
use the comedi_offset_munge() helper to convert the two's complement data to
the offset binary form that the comedi core expects.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: convert 'adbits' boardinfo to 'ai_maxdata'

This member of the boardinfo is the number of bits in the analog input data.
Convert this member into the subdevice 'maxdata', (1 << board->adbits) - 1,
to avoid needing to do the calculation.

Use the subdevice 'maxdata' in the driver instead of using the boardinfo and
use the comedi_offset_munge() helper to convert the two's complement data to
the offset binary form that the comedi core expects.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: remove '= 0' boardinfo

All unlisted members in the boardinfo declaraction will default to '0'.
Remove the unnecessary code.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: move boardinfo declaration

For asethetics, move the boardinfo declaration closer to the struct
definition.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: dt282x: remove forward declarations

Move some of the functions to remove the need for the forward
declarations.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: introduce comedi_buf_n_bytes_ready()

Introduce an inline helper to return the number of bytes that are
ready to read from the comedi_async buffer. Use the helper in the
comedi drivers that currently do the calculation as part of the
(*poll) operation.

Also, use the helper in comedi_fops where the calculation is used as
part of the subdevice going nonbusy.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: ke_counter: expose clock source options to user space

As pointed out by Ian Abbott, the INSN_CONFIG_SET_CLOCK_SRC options should
be exposed in the user header comedi.h.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: rtl8192e: adjust error handling

This patch removes a test in error handling code by adding a return
path.

The Coccinelle semantic match that found the problem is:

// <smpl>
@@
expression E,E1,E2;
@@

E = alloc_etherdev(...)
... when != E = E1
if (...) { ... free_netdev(E); ... return ...; }
... when != E = E2
(
  if (...)
   {
   ... when != free_netdev(E);
   return dev; }
|
* if (...)
   {
   ... when != free_netdev(E);
   return ...; }
|
register_netdev(E)
)

// </smpl>

Signed-off-by: Himangi Saraogi <himangi774@gmail.com>
Acked-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8192u: r8192U_core.c: Cleaning up variable is set more than once

A struct member variable is set to the same value more than once

This was found using a static code analysis program called cppcheck.

Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: rtl8192u: r8192U_wx.c Fixed a code-style error about trailing whitespace

Fixed the following error reported by running checkpatch.pl

ERROR: trailing whitespace

Signed-off-by: Chaitanya Hazarey <c@24.io>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: rtl8192u: r8192U_wx.c Fixed some more code-style warnings about spaces

Fixed the following warnings issued by checkpatch.pl

WARNING: please, no space before tabs

Signed-off-by: Chaitanya Hazarey <c@24.io>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: rtl8192u: r8192U_wx.c Fixed code-style warnings

Fixed some of the following warning generated by checkpatch.pl:

WARNING: Missing a blank line after declarations

Signed-off-by: Chaitanya Hazarey <c@24.io>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: serqt_usb2: fixed frivolous else statement warnings

This is a patch to the serqt_usb2.c file that fixes three instances
of the following checkpatch.pl warning:
WARNING: else is not generally useful after a break or return

Signed-off-by: Vincent Heuken <me@vincentheuken.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fix sparse warnings for static symbols

This patch fixes the following sparse warnings:

drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:46:20: warning: symbol 'the_ksocklnd' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:49:19: warning: symbol 'ksocknal_ip2iface' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:67:15: warning: symbol 'ksocknal_create_route' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:103:1: warning: symbol 'ksocknal_create_peer' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:218:1: warning: symbol 'ksocknal_unlink_peer_locked' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:248:1: warning: symbol 'ksocknal_get_peer_info' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:323:1: warning: symbol 'ksocknal_associate_route_conn_locked' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:368:1: warning: symbol 'ksocknal_add_route_locked' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:409:1: warning: symbol 'ksocknal_del_route_locked' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:511:1: warning: symbol 'ksocknal_del_peer_locked' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:566:1: warning: symbol 'ksocknal_del_peer' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:624:14: warning: symbol 'ksocknal_get_conn_by_idx' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:661:15: warning: symbol 'ksocknal_choose_scheduler_locked' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:685:1: warning: symbol 'ksocknal_local_ipvec' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:713:1: warning: symbol 'ksocknal_match_peerip' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:745:1: warning: symbol 'ksocknal_select_ips' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:847:1: warning: symbol 'ksocknal_create_routes' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:997:1: warning: symbol 'ksocknal_connecting' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:1853:1: warning: symbol 'ksocknal_push_peer' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:1886:1: warning: symbol 'ksocknal_push' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:1935:1: warning: symbol 'ksocknal_add_interface' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:1997:1: warning: symbol 'ksocknal_peer_del_interface_locked' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2038:1: warning: symbol 'ksocknal_del_interface' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2213:1: warning: symbol 'ksocknal_free_buffers' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2256:1: warning: symbol 'ksocknal_base_shutdown' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2349:1: warning: symbol 'ksocknal_new_incarnation' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2364:1: warning: symbol 'ksocknal_base_startup' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2496:1: warning: symbol 'ksocknal_debug_peerhash' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2603:1: warning: symbol 'ksocknal_enumerate_interfaces' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2660:1: warning: symbol 'ksocknal_search_new_ipif' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2702:1: warning: symbol 'ksocknal_start_schedulers' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2748:1: warning: symbol 'ksocknal_net_start_threads' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2860:1: warning: symbol 'ksocknal_module_fini' was not declared. Should it be static?
drivers/staging/lustre/lnet/klnds/socklnd/socklnd.c:2866:1: warning: symbol 'ksocknal_module_init' was not declared. Should it be static?

Signed-off-by: James A Shackleford <shack@linux.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: OnAssocReq23a(): Use cfg80211_find_vendor_ie() to locate VMM

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Only validate vendor specific ies, when they it is vendor specific

Don't run rtw_validate_vendor_specific_ies() on anything but
WLAN_EID_VENDOR_SPECIFIC. In addition use the proper
WLAN_OUI_TYPE_MICROSOFT_* definitions.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_append_pmkid(): Adjust for IEs array not arrying fixed elements

Another case using hard coded offsets counting on the presence of
fixed beacon elements in the IEs array.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_restruct_wmm_ie23a(): Use cfg80211_find_vendor_ie()

Rewrite rtw_restruct_wmm_ie23a() using cfg80211_find_vendor_ie()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Eliminate set_wep_key() and call set_group_key() directly

This also eliminates another round of determine algorithm based on key
length, followed by determine key length based on algorithm.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Validate keys in cfg80211_rtw_add_key()

We validate the parameters in cfg80211_rtw_add_key() so no need to do
it all again in rtw_cfg80211_{ap_}set_encryption()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove obsolete struct ieee_parm

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_{ap_}set_encryption(): Stop passing ieee_parms argument

struct ieee_parms is obsolete, so stop passing it on.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_{ap_}set_encryption(): Add sta_addr argument

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_{ap_}set_encryption(): No more string parsing to determine cipher

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Stop setting ieee_parm arguments that aren't used

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_{ap_}set_encryption(): Add set_tx argument

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_ap_set_encryption(): obtain key from key_params

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_ap_set_encryption(): Pass key_index argument

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_set_encryption(): Pass key_index as argument

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_set_encryption(): Obtain key from key_params->key

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_set_encryption(): Use key_params->cipher

Use key_params->cipher to determine WEP/TKIP

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_set_encryption(): Obtain key_len from struct key_params

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_set_encryption(): Pass additional struct key_params

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_ap_set_encryption(): Use keyparms->cipher to set WEP size

We already know which type of WEP we are using when we get here, no
point trying to determine it from key_len.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_ap_set_encryption(): Stop using param->u.crypt.alg

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_ap_set_encryption(): Use key len from struct key_params

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_ap_set_encryption() is only called for WIFI_AP_STATE

No point checking it again in here, twice!

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_cfg80211_ap_set_encryption(): Use key_params->cipher to determine encryption

Passing strings around for something we have defines for is just
silly.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused RSN_VERSION_BSD23A

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_get_wps_attr_content23a(): Remove unused argument

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused #define P80211CAPTURE_VERSION

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused #define EID_BSSIntolerantChlReport

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Get rid of obsolete _FIXED_IE_LENGTH_ + _BEACON_IE_OFFSET

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Stop carrying half the beacon frame header in the stored IE array

This gets rid of the odd carrying of half the beacon frame in the IE
array stored for the network. Instead we rely on the relevant fields
(timestamp, beacon_interval, and capability) stored in struct
wlan_bssid_ex.

Carrying only half the ieee80211_mgmt header led to a number of bugs
and simply obfuscated the code.

I have tried catching all instances relying on these three elements in
the IEs array, but missed cases may still need to be tracked down.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_generate_ie23a(): update network info as well as IE

We need to make sure to update the network info as well as the IE
array.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: issue_beacon23a(): Determine ie buffer location using 80211_mgmt

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: issue_beacon23a(): Fix another case issuing a corrupted beacon

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused dump_txrpt_ccx_8723a()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove no-op CheckFwRsvdPageContent23a()

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: ConstructProbeRsp(): Use struct ieee80211_mgmt

Use ieee80211_mgmt to construct the frame instead of relying on hard
coded offsets.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Use _FIXED_IE_LEN_ instead of hard coded value

This should help catching all the users once we stop carrying the
fixed beacon elements in the IE array.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_check_beacon_data23a(): Capability is already uptodate in cur_network

Beacon capability was already added to mlme_priv->cur_network before
calling this function, so no need to retrieve it from the IE array
again.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_add_beacon(): Construct a full beacon frame for validation

Clean up and correct the beacon frame validation using a full beacon
frame, and pass that to rtw_check_beacon_data23a(). Previous we went
through hoops to construct the frame, minus the ieee80211_3addr header
which just made it more complicated, and resulted in inconsistencies
and bugs.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: rtw_get_cur_max_rate(): Use _FIXED_IE_LENGTH_ instead of hardcoded value

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused IE offset #defines

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove usage of _{RE, }ASOCREQ_IE_OFFSET_

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: rtl8723au: Remove unused size defines for mgmt frame elements

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge 3.16-rc2 into staging-next

We want the staging fixes here as well.

Linux 3.16-rc2

Merge branch 'i2c/for-next' of git://git./linux/kernel/git/wsa/linux

Pull i2c new drivers from Wolfram Sang:
 "Here is a pull request from i2c hoping for the "new driver" rule.

  Originally, I wanted to send this request during the merge window, but
  code checkers with very recent additions complained, so a few fixups
  were needed.  So, some more time went by and I merged rc1 to get a
  stable base"

So the "new driver" rule is really about drivers that people absolutely
need for the kernel to work on new hardware, which is not so much the
case for i2c.  So I considered not pulling this, but eventually
relented.

Just for FYI: the whole (and only) point of "new drivers" is not that
new drivers cannot regress things (they can, and they have - by
triggering badly tested code on machines that never triggered that code
before), but because they can bring to life machines that otherwise
wouldn't be useful at all without the drivers.

So the new driver rule is for essential things that actual consumers
would care about, ie devices like networking or disk drivers that matter
to normal people (not server people - they run old kernels anyway, so
mainlining new drivers is irrelevant for them).

* 'i2c/for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: sun6-p2wi: fix call to snprintf
  i2c: rk3x: add NULL entry to the end of_device_id array
  i2c: sun6i-p2wi: use proper return value in probe
  i2c: sunxi: add P2WI (Push/Pull 2 Wire Interface) controller support
  i2c: sunxi: add P2WI DT bindings documentation
  i2c: rk3x: add driver for Rockchip RK3xxx SoC I2C adapter

Merge tag 'locks-v3.16-2' of git://git.samba.org/jlayton/linux

Pull file locking fixes from Jeff Layton:
 "File locking related bugfixes

  Nothing too earth-shattering here.  A fix for a potential regression
  due to a patch in pile #1, and the addition of a memory barrier to
  prevent a race condition between break_deleg and generic_add_lease"

* tag 'locks-v3.16-2' of git://git.samba.org/jlayton/linux:
  locks: set fl_owner for leases back to current->files
  locks: add missing memory barrier in break_deleg