staging: comedi: dt9812: remove some #if 0'ed out code

The struct dt9812_usb_cmd definition contains some #if 0'ed
out code that would cause build errors if enabled. Just remove
it.

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

staging: comedi: dt9812: use dev->class_dev for all dev_{level} messages

For consistency in the driver, use dev->class_dev for the device in
all the dev_{level} messages.

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

staging: comedi: dt9812: remove pr_fmt()

There are no pr_{level} messages in this driver. Remove the pr_fmt().

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

staging: comedi: dt9812: cleanup analog out subdevice (*insn_write)

For aesthetic reasons, rename the function to help with grepping and
rename some of the local vars.

The dt9812_analog_out() function can fail. Make sure to check for any
failure and return the errno.

The comedi core expects the (*insn_write) functions to return either
an errno or the number of samples written. Change the final return to
insn->n to make this clearer.

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

staging: comedi: dt9812: cleanup analog out subdevice (*insn_read)

For aesthetic reasons, rename the function to help with grepping and
rename some of the local vars.

This function returns the current analog output value for the channel
that is cached in the private data. Absorb dt9812_analog_out_shadow()
into this function to make it more concise.

The comedi core expects the (*insn_read) functions to return either
an errno or the number of samples read. Change the final return to
insn->n to make this clearer.

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

staging: comedi: dt9812: cleanup analog in subdevice (*insn_read)

For aesthetic reasons, rename the function to help with grepping and
rename some of the local vars.

dt9812_analog_in() can fail. Make sure to check for any failure and
return the errno.

The comedi core expects the (*insn_read) functions to return either
an errno or the number of samples read. Change the final return to
insn->n to make this clearer.

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

staging: comedi: dt9812: convert digital out subdevice to (*insn_bits)

Currently the (*insn_write) function for the digital output subdevice
only sets the state for a single channel. It's more efficent to use
the (*insn_bits) function and allow setting the state for all the
channels.

The comedi core can use the (*insn_bits) to emulate the (*insn_write)
if needed.

Also, use the subdevice 'state' variable to hold the current state
of the channels instead of 'do_shadow' in the private data.

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

staging: comedi: dt9812: convert digital in subdevice to (*insn_bits)

Currently the (*insn_read) function for the digital input subdevice
returns the state for a single channel. It's more efficent to use
the (*insn_bits) function and return the state for all the channels.

The comedi core can use the (*insn_bits) to emulate the (*insn_read)
if needed.

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

staging: comedi: dt9812: remove unused variables from private data

The vendor, product, and serial numbers read from the usb device
are only used for a dev_info() message about the device after it
is reset. Reading these values might not be required for the usb
device to function.

For now just remove the variables from the private data and just
use local variables.

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

staging: comedi: dt9812: convert to use comedi (*auto_attach)

Converting this driver to use the comedi (*auto_attach) mechanism
allows pushing the usb (*probe) into the comedi (*auto_attach) and
the usb (disconnect) into the comedi (*detach). This removes the
disconnect between the usb driver and the comedi driver. Now when
the comedi driver is attached it will always have a usb device
associated with it.

This removes the 16 usb device limitation and allows bringing all
the private data into a single struct that can be kzalloc'ed when
the comedi driver is (*auto_attached). It also allows removing the
the sanity checks that make sure a usb device is connected to the
comedi device in the helper functions.

For aesthetic reasons, add some whitespace to the subdevice init.

Also, fix the analog out subdevice. There are 2 analog output
channels available on the usb device.

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

staging: comedi: dt9812: factor the device reset out of dt9812_probe()

When this driver is converted to the comedi (*auto_attach) mechanism
the device reset will be done during the (*auto_attach). To make the
conversion cleaner, factor the device reset out of the (*probe).

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

staging: comedi: dt9812: factor the endpoint setup out of dt9812_probe()

When this driver is converted to the comedi (*auto_attach) mechanism the
endpoint setup will be done during the (*auto_attach). To make the
conversion cleaner, factor the endpoint setup out of the (*probe).

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

staging: comedi: dt9812: remove 'interface' from usb private data

The usb_interface * is only used in the usb (*probe) and (*disconnect).
Remove the pointer to the 'interface' in the usb private data.

For aesthetic reasons, rename the variable 'interface' to simply
'intf' in the (*probe) and (*disconnect).

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

staging: comedi: dt9812: move shadow values to private data

The analog and digital out shadow values are only used by the
comedi_device. Move the variables from the usb private data to
the comedi private data. Also, rename them to something a bit
shorter.

Move the initialization of the shadow values from the usb probe
to the comedi attach. Also rename the flag used to determine the
initial value from 'range_2_5' to 'is_unipolar'.

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

staging: comedi: dt9812: use module_comedi_usb_driver()

Now that the semaphore used by the comedi_driver is initialized
when the comedi_device private data is allocated, the module_init()
in this driver just clears the 'slot' data before calling
comedi_usb_driver_register(). This static data will already be
cleared when the driver is loaded.

Replace the module_{init,exit}() with the module_comedi_usb_driver()
macro to remove the boilerplate code.

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

staging: comedi: dt9812: move the sempaphore used by the comedi_driver

The semaphore in the struct slot_dt9812 is used by the comedi_driver
when reading and writing to the usb device. It has no real association
with the 'slot'. The 'slot' is protected with a static semaphore in
the driver.

Move the semaphore into the comedi_device private data and initialize
it after the private data is allocated.

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

staging: comedi: dt9812: pass the comedi_device * to dt9812_analog_out_shadow()

For aesthetic reasons, instead of passing the 'slot' pointer, pass the
comedi_device pointer to this function and rename the local var 'result'
to simply 'ret'.

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

staging: comedi: dt9812: tidy up the usb endpoint enumeration

Only the command_{write,read} usb endpoints are used by this driver.

Remove the message_pipe, and {read,write}_stream endpoint information
from the usb private data.

Rename the variables used for command endpoints to help keep the
lines < 80 chars.

Also, rename a couple of the local variables used to enumerate the
endpoints to also help keep the lines < 80 chars.

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

staging: comedi: dt9812: pass the comedi_device * to dt9812_analog_out()

For aesthetic reasons, instead of passing the struct usb_dt9812 pointer,
pass the comedi_device pointer to this function and rename the local
variable 'result' to simply 'ret'.

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

staging: comedi: dt9812: pass the comedi_device * to dt9812_configure_mux()

For aesthetic reasons, instead of passing the struct usb_dt9812 pointer,
pass the comedi_device pointer to this function.

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

staging: comedi: dt9812: pass the comedi_device * to dt9812_configure_gain()

For aesthetic reasons, instead of passing the struct usb_dt9812 pointer,
pass the comedi_device pointer to this function.

Tidy up the function and use the comedi_device class_dev for the
dev_err() message.

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

staging: comedi: dt9812: pass the comedi_device * to dt9812_analog_in()

For aesthetic reasons, instead of passing the 'slot' pointer, pass the
comedi_device pointer to this function and rename the local var 'result'
to simply 'ret'.

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

staging: comedi: dt9812: pass the comedi_device * to dt9812_digital_out_shadow()

For aesthetic reasons, instead of passing the 'slot' pointer, pass the
comedi_device pointer to this function and rename the local var 'result'
to simply 'ret'.

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

staging: comedi: dt9812: pass the comedi_device * to dt9812_digital_out()

For aesthetic reasons, instead of passing the 'slot' pointer, pass the
comedi_device pointer to this function and rename the local var 'result'
to simply 'ret'.

Also, initialize the 'reg', and 'value' when they are declared.

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

staging: comedi: dt9812: pass the comedi_device * to dt9812_digital_in()

For aesthetic reasons, instead of passing the 'slot' pointer, pass the
comedi_device pointer to this function and rename the local var 'result'
to simply 'ret'.

Remove the commented out printk().

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

staging: comedi: dt9812: tidy up usb_bulk_msg() calls

To clarify the code, add a local variable for the struct usb_device
pointer used in the usb_blk_msg() calls.

It's not necessary to initialize the 'count' when writing to the
usb device. The 'count' variable is used to get back the number
of bytes actually sent.

Just return the usb_blk_msg() result when it is the last operation
in a function.

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

staging: comedi: dt9812: remove attaching by serial number

Currently this driver supports attaching 16 usb devices to 16
comedi devices. The user can pass the 32-bit serial number of
the usb device when attaching to force a comedi device to attach
to a specific usb device.

It's also possible to attach a comedi device without having an
attached usb device. When the device is first opened the
comedi_device (*open) function in this driver then checks to
see if a usb device has been attached. If so the subdevice
information is updated based on the usb device and the comedi
device is then operational. If a usb device is not attached
the (*open) returns -ENODEV.

To simplify converting this driver to the comedi (*auto_attach)
mechanism, remove the attaching by serial number option.

Modify the usb (*probe) so that the first available slot is used.
If all the slots are used return -ENODEV.

Modify the comedi (*attach) so that the first unused slot that
has an attached usb device is used. If all the slots are used
return -ENODEV.

Since this ensures that the comedi device has an attached usb
device, remove the (*open) function and fully initialize the
subdevices during the (*attach).

Fix the comedi (*detach) so that the slot is made available
after detaching.

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

staging: comedi: dt9812: rename 'comedi' variable in struct slot_dt9812

Because of the usb (*probe) and comedi (*attach) disconnect, the struct
slot_dt9812 is used to pass the device data between the usb_driver and
the comedi_driver. The variable 'comedi' in this struct is used during
the comedi (*attach) to indicate if a slot is currently being used.

For aesthetic reasons, rename the variable to 'devpriv' since that is
what is actually saved in the pointer.

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

staging: comedi: dt9812: rename the private data struct

The comedi private data (dev->private) for this driver is stored
in a 'struct comedi_dt9812' that is allocated during the comedi
(*attach).

For aesthetic reasons, rename this struct to help make its use
clearer.

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

staging: comedi: dt9812: move the usb framework functions to EOF

In preparation of converting this manually attached comedi driver
into an auto attached comedi usb driver, move the usb framework
functions to the end of the file.

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

staging: comedi: usbduxfast: tidy up usbduxfast_ai_insn_read()

Remove the sanity check of devpriv, this function can only be called
by the comedi core if the private data is allocated.

Get the 'chan', 'range' and 'rngmask' from the function parameters
when the local variables are declared.

For aesthetic reasons, rename the local variable 'err' to 'ret'
since that name is more standard.

(*insn_read) functions are supposed to return the number of values
read or an error code. Change the final return to 'insn->n' to make
it clearer.

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

staging: comedi: usbduxfast: introduce usbduxfast_cmd_data()

Introduce a helper function to set the buffer used to transfer
commands to the usb device.

Each command consists of four uint8_t values that are stored at
specific indexes in the buffer. The helper function consolidates
the code that sets the buffer to reduce coding errors and make
the driver a bit easier to understand.

Note, the '0xff & rngmask' can be reduced to simply 'rngmask'
since the rngmask is always an 8-bit value.

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

staging: comedi: usbduxfast: rename usbduxfastsub_submit_InURBs()

Rename this CamelCase function to fix the checkpatch.pl warning.

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

staging: comedi: usbduxfast: rename CamelCase 'minSamplPer'

Rename the local variable to fix the checkpatch.pl warning.

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

staging: comedi: usbduxfast: consolidate the firmware upload

Absorb the usbduxfastsub_stop(), usbduxfastsub_upload(), and
usbduxfastsub_start() functions into usbduxfast_upload_firmware().

Each of them just do a usb_control_msg() to the device and output
an error message if it fails. A similar message is also output by
usbduxfast_upload_firmware() so the extra messages are redundant.

We can also share the malloc'ed local buffer needed for the
usb_control_msg().

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

staging: comedi: usbduxfast: move usbduxfast_upload_firmware()

For aesthetic reasons, move usbduxfast_upload_firmware() near its
only caller, usbduxfast_request_firmware().

Also, move the #define for the maximum firmware size so it's with
the other firmware defines.

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

staging: comedi: usbduxfast: tidy up firmwareUpload()

Rename the CamelCase function and just pass the const struct
firmware * directly instead of parsing out the data and size.

Rename the CamelCasae local val 'fwBuf'.

Use goto to provide a common exit path that frees the local
buffer for the firmware when an error is encountered.

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

staging: comedi: usbduxfast: fix a > 80 char line issue

Fix a line over 80 characters issue reported by checkpatch.pl.

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

staging: comedi: usbduxfast: tidy up usbduxfast_ai_cancel()

Remove the unnecessary comments and dev_err() noise.

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

staging: comedi: usbduxfast: tidy up usbduxfast_ai_stop()

Remove the devpriv sanity check, all the callers have previously
verified that its valid.

Absorb usbduxfast_unlink_urbs() into this function since it's the
only caller and the only thing it does is call usb_kill_urb().

This function will always succeed so just return 0 and remove the
local variable 'ret'.

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

staging: comedi: usbduxfast: tidy up usbduxfastsub_unlink_InURBs()

Rename this function so it has namespace associated with the driver.

Remove the local variables 'j' and 'ret'. They are not used and the
function always returns '0'.

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

staging: comedi: usbduxfast: tidy up send_dux_commands()

Rename this function so it has namespace associated with the driver.

For aesthetic reasons, rename the local variable 'tmp' to 'ret' as
this is more common for checking errno values.

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

staging: comedi: usbduxfast: remove extra dev_err() messages

A dev_err() message is output if send_dux_commands() fails. Remove
the extra dev_err() messages output by the callers.

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

staging: comedi: usbduxfast: rename 'dux_commands' in private data

For aesthetic reasons, rename this variable to 'duxbuf' and also
rename the define used for its size.

This helps fix some of the > 80 char lines reported by checkpatch.pl.

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

staging: comedi: usbduxfast: rename 'transfer_buffer' in private data

For aesthetic reasons, rename this variable to 'inbuf' to match the
define used for its size (SIZEINBUF).

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

staging: comedi: usbduxfast: rename 'insnBuffer' from private data

This buffer space is kmalloc'ed and free'ed but never used. Just
remove it.

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

staging: comedi: usbduxfast: rename 'urbIn' in private data

Rename the CamelCase variable 'urbIn' in the private data to
simply 'urb'.

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

staging: comedi: usbduxfast: remove 'usb' from the private data

The usb_device pointer can be fetched when needed from the comedi_device
by using comedi_to_usb_interface() then interface_to_usbdev().

Do that instead and remove the pointer from the private data.

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

staging: comedi: usbduxfast: remove 'intf' from the private data

The usb_interface pointer is already saved in the comedi_device
by the comedi core.

Remove the pointer from the private data and use comedi_to_usb_interface()
to get the usb_interface when needed.

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

staging: comedi: usbduxfast: use dev->class_dev for all dev_{level} messages

For consistency in the driver, change all the dev_{level} messages
so they use dev->class_dev for the device.

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

staging: comedi: usbduxfast: remove 'comedidev' from the private data

Now that the comedi_device is passed to the internal functions, the
back pointer to it is not needed in the private data. Remove it.

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

staging: comedi: usbduxfast: don't pass the private data directly

Instead of passing the private data to the internal functions, pass
the comedi_device pointer and get the private data from it.

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

staging: comedi: usbduxfast: remove 'ifnum' from private data

The interface number is only needed during the attach of the
device. Remove it from the private data.

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

staging: comedi: usbduxfast: remove 'attached' from private data

The subdevice function pointers are only setup if the device is
attached. The 'attached' flag in the private data is unnecessary.
Remove it.

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

staging: comedi: usbduxfast: remove 'probed' from private data

The 'probed' flag in the private data was used to handle the
disconnect between the usb (*probe) and the comedi_driver
(*auto_attach) so that if the comedi_driver was attached without
an associated usb_device the subdevice functions would return
-ENODEV.

Now that the usb_driver (*probe) is part of the comedi_driver
(*auto_attach), the comedi_driver can only attach if the usb
device is present.

Remove the unnecessary variable from the private data and its
uses in the driver.

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

staging: comedi: usbduxfast: absorb tidy_up() into the (*detach)

Now that the usb (*probe) is handled in the comedi_driver (*auto_attach)
the comedi_driver (*detach) will be called automatically if the
attach fails.

The tidy_up() calls in the (*auto_attach) are not needed. Remove them
and absorb the code in tidy_up() into the (*detach).

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

staging: comedi: usbduxfast: remove some unused defines

Now that the private data is kzalloc'ed when the device is attached,
the NUMUSBDUXFAST define is not used. Remove it.

The SUBDEV_AD define is only used when setting up the subdevice.
Remove it and just open-code the value.

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

staging: comedi: usbduxfast: Move usb_driver (*probe) into comedi_driver

Currently, the usb_driver (*probe) locates a free place in the static
usbduxfast array to use for the private data in this driver. It then
does some initial setup of the usb device and allocates the buffers
needed. The firmware for the device is then requested and uploaded
before calling comedi_usb_auto_config() to allow the comedi core to
complete the (*auto_attach) of the device.

Move the bulk of the (*probe) into the comedi_driver (*auto_attach).
This allows the private data to be kzalloc'ed. We can then remove the
static array along with the semaphore that protected it.

This also removes the 16 attached devices limitation.

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

staging: comedi: usbduxfast: Use comedi_usb_auto_unconfig() for (*disconnect)

The usb_driver (*disconnect) calls comedi_usb_auto_unconfig() then
frees any allocations by calling tidy_up(). comedi_usb_auto_unconfig()
calls the comedi_driver (*detach) function, usbduxfast_detach().

Move the tidy_up() call into usbduxfast_detach() and use the
comedi_usb_auto_unconfig() directly for the (*disconnect).

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

staging: comedi: usbduxfast: tidy up usbduxfastsub_ai_Irq()

Rename the CamelCase function,

Rename some of the local variables to the normal names used in
comedi drivers. Add a local variable for the comedi_async *.

Remove the sanity checking. The urb that causes this function to
be called can only be submitted if the sanity checks already
passed.

Change the way the comedi_subdevice is fetched. The attach function
sets the dev->read_subev pointer to the correct subdevice. Use that
instead of accessing the dev->subdevices array directly.

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

staging: comedi: usbduxfast: remove all COMEDI_DEBUG messages

These messages are just function trace noise. Remove them.

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

staging: comedi: usbduxfast: tidy up usbduxfast_attach_common()

Remove the unnecessary comments and tidy up the whitespace in
this function.

Also, remove the dev_info() at the end. It's just added noise.

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

staging: comedi: usbduxfast: remove 'owner' from usb_driver

The struct usb_driver does not have an 'owner' field. Remove the
code in this driver that initializes it.

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

staging: comedi: usbduxfast: rename usbduxfastsub_s variables

For aesthetic reasons, globally rename the variables used for the
struct usbduxfastsub_s * in this driver to 'devpriv'. This variable
pointes to the comedi_device private data.

Also, rename the struct to 'usbduxfast_private' to make its use
a bit clearer.

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

staging: comedi: usbduxfast: rename usb_device variables

For aesthetic reasons, globally rename the variables used for the
struct usb_device * in this driver to simply 'usb'.

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

staging: comedi: usbduxfast: rename usb_interface variables

For aesthetic reasons, globally rename the variables used for the
struct usb_interface * in this driver to simply 'intf'.

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

staging: comedi: usbduxfast: use a local var for the usbduxfastsub_s

Add a local variable to hold the pointer to the place in the
usbduxfast array the is being configured for the device.

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

staging: comedi: usbduxfast: remove COMEDI_DEBUG noise in usb (*probe)

The COMEDI_DEBUG noise in the usb (*probe) is just added function
trace noise. Remove it along with the final dev_info().

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

staging: comedi: usbduxfast: request firmware synchronously

Change the request_firmware_nowait() to a request_firmware() so
that the usb_driver (*probe) and continue with the comedi_driver
(*auto_attach).

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

Staging: speakup: Fix return value in synth.c

The function return type is a pointer. Hence return NULL instead
of 0.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: speakup: Fix sparse warning in varhandlers.c

'cp' is a pointer. Fix the incorrect comparison with integer to
avoid the below warning:
drivers/staging/speakup/varhandlers.c:283:19: warning:
Using plain integer as NULL pointer

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: speakup: Return NULL instead of 0 in speakup_decpc.c

The function returns a pointer. Hence return NULL instead of 0.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Acked-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: remove custom locking macro definitions

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: synth: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: thread: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: soft: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: keypc: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: dtlk: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: dectlk: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: decpc: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: decext: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: apollo: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: acntpc: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: serialio: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: remove spk_trylock macro definition

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: main: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: i18n: remove custom locking macros

Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: devsynth: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: speakup: kobjects: remove custom locking macros

Signed-off-by: William Hubbs <w.d.hubbs@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: wlags49_h2: fix up "205179d staging: wlags49_h2: strncpy, need checking the memory length" for avoiding warnings

Type of 'probe_rsp->rawData[1]' is 'hcf_8' which is 'u8', it will
compare with 'int', so use min_t() to cast to 'u8' to avoid related
warning.

If use min(), it has '(void) (&_min1 == &_min2);', so if no type cast,
the compiler will report 'pointer types lacks a cast':

  drivers/staging/wlags49_h2/wl_main.c:3174:122: warning: comparison of distinct pointer types lacks a cast [enabled by default]

Signed-off-by: Chen Gang <gang.chen@asianux.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: nvec: Convert to use devm_input_allocate

nvec_kbd is converted to use devm_input_allocate.
This simplifies error handling and remove path.

Signed-off-by: Leon Romanovsky <leon@leon.nu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: ni_labpc: remove *_ai_gain_bits tables

The bits needed to set the analog input gain can be simply calculated
based on the 'range'.

The LabPC versions of the board do not have the '0x10' gain that the
LabPC+ board supports. By incrementing the range appropriately the
correct gain bits can still be calculated.

This allows removing the two gain tables, as well as the export, along
with the 'ai_range_code' data in the boardinfo.

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

Staging: ft1000: Change C99 comments in ft1000_dnld.c

This simple patch fixes the ERRORs found by checkpatch.pl in the file
ft1000_dnld.c by switching to the standard Kernel comment style.

Signed-off-by: Dylan Socolobsky <dsocolobsky@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: ft1000: fix brace coding style in ft1000_dnld.c

This simple patch fixes a brace warning raised by checkpatch.pl in ft1000_dnld.c

Signed-off-by: Dylan Socolobsky <dsocolobsky@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/xgifb: Fix always false condition in vb_setmode.c (smatch)

Smatch complains:
drivers/staging/xgifb/vb_setmode.c:3181
XGI_SetLockRegs() warn: bitwise AND condition is false here

Since SetNTSCTV is defined as 0 in drivers/video/sis/initdef.h this is
correct.
-> Change the condition to == to fix this.

Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/xgifb: Remove unnecessary binary calculation in GetXG27FPBits

If temp <= 2 no other bits than the lowest two (0x03) can be set anyway
-> this operation can be removed.

Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/xgifb: Consolidate function almost identical calls in XGINew_GetXG21Sense

Instead of calling xgifb_reg_and_or with almost identical parameters in
a simple if/else scenario, we assign the changing parameter to a temp
variable and call xgifb_reg_and_or only from one location.

-> Easier to read, easier to understand (especially wrt the line breaks)

(For the if condition we don't need the Temp variable, so we can use the
value directly).

Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/xgifb: Remove unnecessary cases in XGINew_SenseLCD

The switch statement sets the temp value to zero for certain cases and
leaves it untouched for other cases -> all these other cases can be
combined in the default case.

Since an empty case containing only a break, it can be removed.

The patch also removes the if statement, as it uses the same value as the switch
for comparison, and includes the code into the switch.

Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/xgifb: Remove uneccessary/simply if in XGINew_GetXG20DRAMType

Instead of storing the value in a temp variable, anding it with 1 and
assign data conditionally we can assign the value directly.

Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging/sb105x: remove asm/segment.h dependency

sb105x doesn't seem to actually need <asm/segment.h> (builds on x86
without it) and ppc/ppc64 doesn't provide it so it fails to build there.

This patch removes the dependency.

CC: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>