In function zram_bvec_write(), previous data at the index is
already freed by function zram_free_page().
When failed to compress or zs_malloc, there is no way to restore old data.
Therefore, free previous data when it's about to update.

Also, no need to check whether table is not empty outside of
function zram_free_page(), because the function properly checks inside.
Signed-off-by: Sunghan Suh <sunghan.suh@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

No one has the hardware for it anymore, and there has not been any
development on it in a long time.

If someone shows up with the hardware, and wants to clean it up, this
can be easily reverted.
Reported-by: Lidza Louina <lidza.louina@gmail.com>
Cc: Jakub Schmidtke <sjakub@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

These functions are identical other than the which and how many urbs
are unlinked.

Consolidate the functions into one usbdux_unlink_urbs() function.
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>

For aesthetic reasons, move the defines used for the bulk transfer
commands out of the middle of the code and rename them so they are
a bit more readable.
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>

This define is not used, and it's actually incorrect. The usbdux device
only has 4 analog output channels.

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>

The 'dac_commands' buffer is used to pass the ao channel list from
usbdux_ao_cmd() to the urb callback, usbduxsub_ao_isoc_irq(). This
buffer does not need to be allocated.

Change it into a simple array of the correct size. Rename the variable
to 'ao_chanlist' to clarify what it actually is.
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>

This check is not necessary. The cmd->chanlist_len will always be
less than or equal to the subdevice len_chanlist that was setup
during the attach of the device.
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>

This check is not necessary. The cmd->chanlist_len will always be
less than or equal to the subdevice len_chanlist that was setup
during the attach of the device.
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>

Remove the ugly casting of datap to the ao urb->transfer_buffer. The format
of the data in the buffer is simply.

  [0] = # of channels to update (s->async->cmd.chanlist.len)
  [1] = lsb of value for 1st channel to update
  [2] = msb of value for 1st channel to update
  [3] = 1st channel to update (depvriv->dac_commands[0])

If more than 1 channel is to be updated they are simply added to the buffer:

  [4] = lsb of value for 2st channel to update
  [5] = msb of value for 2st channel to update
  [6] = 1st channel to update (depvriv->dac_commands[1])

etc.
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>

This buffer is used to cache the values that are written to the
analog output channels. Currently it only caches the single writes
to the channels using the (*insn_write) callback. The async command
writes are not cached. The buffer is also being kzalloc'ed during
the attach of the driver to a size much larger that required.

Rename the CamelCase buffer and change it to an array in the private
data of the correct size to cache the analog output channel values.

Modify the analog output urb callback so it updates the cached values
with those used for the asynchronous command to allow readback after
the command completes.

The sanity check of the index to dac_commands[] (i.e. the 'chan' being
written) is not needed. The chanlist_len will always be less than the
number of channels.

Also, fix the dev_err() message so it uses the proper device pointer.
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>

The usbdux and usbduxsigma drivers are _very_ similar. For aesthetic
reasons, rename the private data variables in this driver to match
the names in the usbduxsigma driver so we can start sharing the
common 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>

Use the comedi_range_is_bipolar() helper instead of checking the
'range' index against a magic number.

Also, use the s->maxdata to calculate the value needed to munge the
value for bipolar data instead of the magic 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>

For aesthetics, move this function closer to the (*auto_attach).

Also, rename some of the defined constants that are used by the
firmware upload.
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>

These defines are not used by the driver. Remove them.
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>

The endpoint defines are each only used in one place and don't help
clarify the code. Remove the defines and just 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>

Use the stop helpers instead of duplicating the code in the detach.
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>

For aesthetic reasons, pass the comedi_device pointer to the unlink
helpers instead of the private data pointer.

All the unlink helpers simply call usb_kill_urb() to cancel any pending
transfer requests. The usb passed to usb_kill_urb() can be NULL so the
extra sanity check is not required.

The unlink helpers will always return success so just make them void
functions.

Since the stop helpers will also always return success, make them void
functions as well.
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>

Add the missing down/up of the semaphore to prevent other commands
from being issued to the usb device while the pwn is being stopped.

Rename the local variable used for the private data pointer to the
comedi "norm".

Use memset() to initialize the urb transfer buffer.
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>

Add the missing down/up of the semaphore to prevent other commands
from being issued to the usb device while the pwn is being stopped.

Rename the local variable used for the private data pointer to the
comedi "norm".

Make sure to check that usbdux_pwm_stop() was successful before
sending command to the usb device to stop the pwm.
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>

For aesthetic reasons, pass the comedi_device pointer to this function
instead of the private data pointer. Rename the local variable used
for the private data pointer to the comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.
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>

Comedi (*insn_write) operations are supposed to write insn->n values.
Fix this function to work like the core expects.

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.
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>

Comedi (*insn_read) operations are supposed to read and return insn->n
values. Fix this function to work like the core expects.

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.

Tidy up the exit path using goto to ensure that the semaphore is
released.
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>

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.

Tidy up the exit path using goto to ensure that the semaphore is
released.
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>

Tidy up this 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>

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.

Make sure an ao command is not already running and return -EBUSY.

Tidy up the exit path using goto to ensure that the semaphore is
released.
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>

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.

Tidy up the exit path using goto to ensure that the semaphore is
released.

Return -EBUSY instead if an ao command is already running.
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>

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.

Tidy up the exit path using goto to ensure that the semaphore is
released.

Return -EBUSY instead of 0 if the (*insn_write) cannot be done because
a command is running.

Tidy up the for() loop that writes the data. The dux_commands[1] and [4]
can be set outside the loop since they are constant. Use a local pointer
for dux_commands[2] to load the value to write. Only the last value needs
to be cached in the private data for read back.
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>

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.
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>

Use the comedi_range_is_bipolar() helper instead of checking the
'range' index against a magic number.

Also, use the s->maxdata to calculate the value needed to munge the
value for bipolar data instead of the magic 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>

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.

Tidy up the exit path using goto to ensure that the semaphore is
released.

Return -EBUSY instead of 0 if the (*insn_read) cannot be done because
a command is running.
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>

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.

Tidy up the exit path using goto to ensure that the semaphore is
released.
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>

Rename the local variable used for the private data pointer to the
comedi "norm".

Remove the unnecessary sanity check of the private data pointer. This
function can only be called is the private data was allocated during
the attach.

Tidy up the exit path using goto to ensure that the semaphore is
released.

Return -EBUSY if an ai command is already running.
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>

As pointed out by Ian Abbott, the *_cmd_running flags should be set
before submitting the urbs. There is a possible race condition where
an urb could complete and the flag is checked in the completion
routine before it's set.
Reported-by: Ian Abbott <abbotti@mev.co.uk>
Acked-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

A spin lock is taken here so we should use GFP_ATOMIC.
Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Acked-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The COMEDI_NI_ATMIO configuration option causes the "ni_atmio" module to
be built, which handles various National Instruments ISA-PNP data
acquisition cards.  The configuration option currently depends on
ISAPNP.

The "ni_atmio" module is a "legacy" comedi driver with no
auto-configuration support and it doesn't register as a PNP card driver.
The only difference in initialization from the other comedi ISA card
drivers is that it will use the ISA-PNP functions to find, attach and
activate the card if no I/O base address has been specified by the user
in the COMEDI_DEVCONFIG ioctl.

The module compiles fine without the ISAPNP configuration option as the
<linux/isapnp.h> and <linux/pnp.h> headers provide dummy inline
functions if the ISAPNP and PNP options, respectively, are not enabled.

Remove the dependancy on the ISAPNP option since the module builds
without it and can be used without it.

(There is a very slim chance that comedi will support proper ISA PNP
drivers one day, in which case the ni_atmio driver could be converted to
use this support and not support manual configuration.  However, not all
the PnP IDs of boards supported by this driver are currently known to
us.)
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The COMEDI_NI_ATMIO16D configuration option causes the "ni_atmio16d"
module to be built, which handles the National Instruments AT-MIO-16 and
AT-MIO-16D cards.  The configuration option currently depends on ISAPNP,
but the module doesn't use any ISAPNP functions and the cards are legacy
ISA cards with base I/O port, IRQ and DMA channel configured by jumpers.
(The driver doesn't use DMA though.)

Remove the ISADNP dependancy from COMEDI_NI_ATMIO16D and fix up the help
text - the card names are missing a dash and they're not ISA-PNP.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Resolves warnings from the "sparse" checker of the form "warning: cast to
restricted __le??".
Signed-off-by: Shaun Laing <shaun@xresource.ca>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Eli Billauer <eli.billauer@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

As reported, this driver causes problems when built into the kernel, so
force it to be a module until those issues are fixed.

Also delete the "default n" lines in the Kconfig, as they are not
needed.
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Eli Billauer <eli.billauer@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This is the driver for Xillybus, which is a general-purpose interface for
data communication with FPGAs (programmable logic). Please refer to the
README included in this patch for a detailed explanation.

It was previously submitted for misc-devices, but it appears like noone's
willing to review the code (which I can understand, given its magnitude).
Hence submitted as a staging driver.
Signed-off-by: Eli Billauer <eli.billauer@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>