staging: lustre: replace MIN with min, cast (__kernel_size_t)

Switch from MIN to min and fix the new type warning.  The
warning is produced because a comparison between iov_len,
which is a __kernel_size_t, is made to kiov_len, which is an
unsigned int (include/linux/lnet/types.h).  Fix the warning
by casting kiov_len to __kernel_size_t.

Signed-off-by: Jeremiah Mahler <jmmahler@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: replace MIN/MAX with min_t/max_t

Switch from MIN/MAX to min_t/max_t with a size_t type.  The size_t type
was chosen because one operand is a size_t and all the others are
immediate integer values.

Signed-off-by: Jeremiah Mahler <jmmahler@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: replace MIN with min_t

Switch from MIN to the built in min_t with the int type.

Signed-off-by: Jeremiah Mahler <jmmahler@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: use min/max instead of MIN/MAX, simple cases

Custom MIN/MAX operations are being used which are not as robust
as the built in min/max operations which will warn about potentially
problematic type comparisons.

For the simple cases, where no type warning is produced, simply
replace MIN/MAX with min/max.

Signed-off-by: Jeremiah Mahler <jmmahler@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: cleanup align switch and case

Align switch and case to be at the same indent.

Signed-off-by: Samuel Kihahu <skihahu@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: lustre: error handling tweaks in obd_ioctl_getdata()

1)  The places which called copy_from_user() were returning the number
of bytes not copied instead of -EFAULT.

2) The user could trigger a memory leak if the condition
"(hdr.ioc_len != data->ioc_len)" was true.  Instead of adding a new call
to OBD_FREE_LARGE(), I created a free_buf label and changed everything
to use that label.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: lustre: obdclass: lprocfs_status.c: Tidied up some code in a case statement

Tidied up some code in a case statement after a discussion.

This was partly 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: lustre: lustre: obdclass: lprocfs_status.c: Removed an unnecessary NULL check

Removed an unnecessary NULL check.

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

lustre: ptlrpc: add "ptlrpc_internal.h" to sec_gc.c

ptlrpc_internal.h contains the prototypes for sptlrpc_gc_init() and
sptlrpc_gc_fini(), which are defined in sec_gc.c.

This removes the following sparse warnings:

drivers/staging/lustre/lustre/ptlrpc/sec_gc.c:217:5: warning: symbol 'sptlrpc_gc_init' was not declared. Should it be static?
drivers/staging/lustre/lustre/ptlrpc/sec_gc.c:241:6: warning: symbol 'sptlrpc_gc_fini' was not declared. Should it be static?

Signed-off-by: Spencer Baugh <sbaugh@andrew.cmu.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: lustre: mdc: mdc_lib.c: Removed an unnecessary NULL check

Removed an unnecessary NULL check.
I have checked the only place this function is called from.

Signed-off-by: Rickard Strandqvist <rickard_strandqvist@spectrumdigital.se>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers: staging: lustre: Track sign separately

The `mult' parameter is negated if the user data begins with a '-' so
that the final value has the appropriate sign.  But `mult' is only used
if the user data does not include a "units" suffix.  In this case,
`mult' is overridden with the numeric scale conveyed by the units suffix,
but retains the sign of the original value.

Having `mult' serving double-duty works but is confusing.  Use a new
local variable to store the sign of the user data instead.  This also
fixes a pitfall of passing 0 to `mult', expecting it to be ignored when
a units suffix is specified, but having the effect of taking the
absolute value of the user-provided data.

Signed-off-by: Chris Rorvick <chris@rorvick.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers: staging: lustre: Use mult if units not specified

Units can be passed to lprocfs_write_frac_u64_helper() via a suffix
(e.g., "...K", "...M", etc.) tacked onto the value.  A comment states
that "specified units override the multiplier," though the multiplier is
overridden regardless.  Update the conditional logic so that it only
applies when units are specified.

Signed-off-by: Chris Rorvick <chris@rorvick.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: libcfs: fix sparse warnings about static declaration

make sparse happy since these two fuchtion are only used in module.c.
tested by successful compilation.

Signed-off-by: Jia He <hejianet@gmail.com>
Cc: Oleg Drokin <oleg.drokin@intel.com>
Cc: Andreas Dilger <andreas.dilger@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: Cleanup cfs_str2mask in libcfs_string.c

- Replace body-less for-loop with while loop
- Use '\0' for null character instead of 0

Signed-off-by: Matthew Tyler <matt.tyler@flashics.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: Fix minor style error in libcfs_string.c

Signed-off-by: Matthew Tyler <matt.tyler@flashics.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

lustre: don't use iovec instead of kvec

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: lustre: obd: Fix a coding style issue

Fixed a brace coding style issue for functions.

Signed-off-by: Dean Michael Ancajas <dbancajas@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging:lustre:lustre:ptlrpc: fix sparse warnings in lproc_ptlrpc.c

Fixed the below warning in sparse:

drivers/staging/lustre/lustre/ptlrpc/lproc_ptlrpc.c:184:6:
warning: symbol 'ptlrpc_lprocfs_register' was not declared.
Should it be static?

Signed-off-by: Fred Chou <fred.chou.nd@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: replace static value with define

This patch replace the value '80' used in several files in the lustre source code
with a define LUSTRE_MDT_MAXNAMELEN.

This value is used in 4 different structures as the maximum len for a service name.
According to the comments, these names follow a convention which make it possible
to use the same define for LCS, LSS, LCF, and LSF.

Signed-off-by: Tristan Lelong <tristan@lelong.xyz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fix sparse warning on LPROC_SEQ_FOPS macros

This patch fix a sparse warning in lustre sources

warning: incorrect type in argument 1 (different address spaces)
    expected void [noderef] <asn:1>*to
    got char *<noident>

This is done by adding the missing __user attribute on userland pointers inside the LPROC_SEQ_FOPS like macros:
- LPROC_SEQ_FOPS
- LPROC_SEQ_FOPS_RW_TYPE
- LPROC_SEQ_FOPS_WR_ONLY
- LDLM_POOL_PROC_WRITER

The patch also updates all the functions that are used by this macro:
- lprocfs_wr_*
- *_seq_write

as well as some helpers used by the previously modified functions (otherwise fixing the sparse warning add some new ones):
- lprocfs_write_frac_helper
- lprocfs_write_helper
- lprocfs_write_u64_helper

The patch also fixes one __user pointer direct dereference by strncmp in function fld_proc_hash_seq_write.

Signed-off-by: Tristan Lelong <tristan@lelong.xyz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: lustre: fld: fixed symbol was not declared

Fixes the sparse warning:
"warning: symbol 'fld_type_proc_dir' was not declared. Should it be static?"

Also removes initialization to NULL.

Signed-off-by: Brian Vandre <bvandre@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: ft1000-pcmcia: fix else close brace style

Fix the following checkpatch.pl errors

ERROR: else should follow close brace '}'
+                       }
+                       else {

ERROR: else should follow close brace '}'
+                       }
+                       else {

ERROR: else should follow close brace '}'
+                       }
+                       else {

ERROR: else should follow close brace '}'
+       }
+       else {

ERROR: else should follow close brace '}'
+                       }
+                       else {

Signed-off-by: Austin Kerbow <amkerbow@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: ft1000: ft1000-usb: ft1000_hw.c: Fix a potential memory leak.

Avoid allocate memory if we will exit the function.

Was found by 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: ft1000: ft1000-pcmcia: Fixed a coding style issue

Fixed a coding style issue for braces.

Signed-off-by: Dean Michael Ancajas <dbancajas@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gs_fpgaboot: io.c: Remove unused function

Remove the function bitswap() that is not used anywhere.

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

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

Staging: nvec: Fixed a typo

Fixed a typo in nvec.c file

Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: olpc_dcon: fix sparse symbol not declared warning

This patch gets rid of the following sparse warning:

drivers/staging/olpc_dcon/olpc_dcon.c:787:19: warning: symbol 'dcon_driver' was not declared. Should it be static?

Signed-off-by: Murilo Opsfelder Araujo <mopsfelder@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: NetLogic: Coding style correction

Misspelled comment corrected

Signed-off-by: tolga ceylan <tolga.ceylan@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: olpc_dcon: Fixed a typo

Fixed a typo in olpc_dcon_xo_1.c file.

Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: skein: Fixing single statement macro checkpatch warning

WARNING: Single statement macros should not use a do {} while (0) loop

Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Acked-by: Jason Cooper <jason@lakedaemon.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: skein: Fixing trailing whitespace error

Signed-off-by: Mathieu Poirier <mathieu.poirier@linaro.org>
Acked-by: Jason Cooper <jason@lakedaemon.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: dgnc: implement proper error handling in dgnc_start()

dgnc_start() ignores errors in class_create() and device_create()
and it does not deallocate resources if dgnc_tty_preinit() fails.

The patch implements proper error handling.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: octeon-usb: Fixed a typo

Fixed a typo in octeon-hcd.c file

Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: octeon: Fixed a typo

Fixed a typo in ethernet.c file.

Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: octeon: Add blank lines after declarations

There are 2 missing blank lines after variable declarations in
octeon/ethernet-rx.c. They have been added per the coding style.

Signed-off-by: Jamie Lawler <jamie.lawler@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: gdm724x: fix line limit coding style issue in gdm_lte.c

This is a patch to fix "WARNING: line over 80 characters" found by
checkpatch.pl in gdm_lte.c.

Signed-off-by: Ning Zhou <zhou.ning.gd@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: gdm724x: fix space before comma coding style issue in gdm_mux.c

This is a patch to the gdm_mux.c file that fixes a space before comma
coding style issue found by the checkpatch.pl tool.

Signed-off-by: Robert Nachlinger <robert.nachlinger@googlemail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: vt6655: Checkpatch fix: c99 comment headings

This patch fixes some of the following checkpatch.pl errors in mac.h
ERROR: do not use C99 // comments

This patch reformats all single line c99 style comments to the
preferred style.

Signed-off-by: Emrys Bayliss <emrys@paradise.net.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: vt6655: comment headings formatting

This patch makes the formatting of the comments in mac.h more
consistent.

* Where a heading takes up three comment lines it is reduced to one.
* A newline always separates column headings

Signed-off-by: Emrys Bayliss <emrys@paradise.net.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: vt6655: Checkpatch fix: lines longer than 80 columns

This patch fixes the following checkpatch.pl errors in mac.h:
WARNING: line over 80 characters

Signed-off-by: Emrys Bayliss <emrys@paradise.net.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: vt6655: Checkpatch fix: c99 comments after statements

This patch fixes some c99 errors in mac.h:
ERROR: do not use C99 // comments

All instances where the c99 comment occurs after a statement.

Signed-off-by: Emrys Bayliss <emrys@paradise.net.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: vt6655: comment after statement formatting

This patch makes the formatting of the comments in mac.h more
consistent.

* Where a comment follows a statement, it is separated by a single
  space.

Signed-off-by: Emrys Bayliss <emrys@paradise.net.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: vt6655: Checkpatch fix: empty c99 comments

This patch fixes some of the following checkpatch.pl errors in mac.h:
ERROR: do not use C99 // comments

This patch truncates statements which are followed by empty c99
comments.

Signed-off-by: Emrys Bayliss <emrys@paradise.net.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: vt6655: Checkpatch fix: unnecessary whitespace

This patch fixes the following checkpatch.pl errors in mac.h:
ERROR: space prohibited before that ',' (ctx:WxW)

Signed-off-by: Emrys Bayliss <emrys@paradise.net.nz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Staging: comedi: kcomedilib: Fixed a typo

Fixed a typo in kcomedilib_main.c file.

Signed-off-by: Gustavo A. R. Silva <garsilva@embeddedor.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

staging: comedi: pcl812: use common function to setup dma

The pcl812_ai_setup_dma() and pcl812_ai_setup_next_dma() functions are similar
other than the buffer switch and the inclusion of the "unread_samples" in
pcl818_ai_setup_next_dma() when calculating the dma size.

Merge these two functions by initializing the 'dma->cur_dma' in the callers
and passing '0' for the "unread_samples" when first starting the DMA.

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: pcl812: simplify the dma->size calculations

The dma->size calculations are overly complicated in this driver and the
'prealloc_bufsz' sanity checks are unnecessary.

Currently this driver determines the number of DMA "runs" needed and the size
of the "last" DMA transfer in order to perform a command. As long as there are
more "runs" required, the dma->size is set to the buffer maxsize. On the last
"run" the buffer is set to the "last" size.

Refactor the driver to use the comedi core helpers to determine the DMA size
based on the buffer maxsize and the number of samples remaining in the command.

This allows removing the 'dma_runs_to_end' and 'last_dma_run' mamebers from
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: pcl816: use common function to setup dma

THe pcl816_ai_setup_dma() and pcl816_ai_setup_next_dma() functions are similar
other than the buffer switch and the inclusion of the "unread_samples" in
pcl818_ai_setup_next_dma() when calculating the dma size.

Merge these two functions by initializing the 'dma->cur_dma' in the callers
and passing '0' for the "unread_samples" when first starting the DMA.

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: pcl816: simplify the dma->size calculations

Currently this driver determines the number of DMA "runs" needed and the size
of the "last" DMA transfer in order to perform a command. As long as there are
more "runs" required, the dma->size is set to the buffer maxsize. On the last
"run" the buffer is set to the "last" size.

Refactor the driver to use the comedi core helpers to determine the DMA size
based on the buffer maxsize and the number of samples remaining in the command.

This allows removing the 'dma_runs_to_end' and 'last_dma_run' mamebers from
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: pcl818: use common function to setup dma

THe pcl818_ai_setup_dma() and pcl818_ai_setup_next_dma() functions are similar
other than the buffer switch and the inclusion of the "unread_samples" in
pcl818_ai_setup_next_dma() when calculating the dma size.

Merge these two functions by initializing the 'dma->cur_dma' in the callers
and passing '0' for the "unread_samples" when first starting the DMA.

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: pcl818: simplify the dma->size calculations

Currently this driver determines the number of DMA "runs" needed and the size
of the "last" DMA transfer in order to perform a command. As long as there are
more "runs" required, the dma->size is set to the buffer maxsize. On the last
"run" the buffer is set to the "last" size.

Refactor the driver to use the comedi core helpers to determine the DMA size
based on the buffer maxsize and the number of samples remaining in the command.

This allows removing the 'dma_runs_to_end' and 'last_dma_run' mamebers from
the private data. Also remove the 'ai_data_len' member which is set but never
used.

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: ni_labpc_isadma: use comedi_isdma module

The labpc driver has already had the depends on ISA_DMA_API limitation removed
by moving all the DMA support code into the labpc_isadma module.

For aesthetics, use the comedi_isadma module to provide the actual ISA DMA
support.

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: ni_at_a2150: remove depends on ISA_DMA_API limitation

Use the new comedi_isadma module to provide the ISA DMA support. This removes
all the ISA_DMA_API code from this driver and allows the driver to be used on
platforms that don't support the ISA_DMA_API.

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 depends on ISA_DMA_API limitation

Use the new comedi_isadma module to provide the ISA DMA support. This removes
all the ISA_DMA_API code from this driver and allows the driver to be used on
platforms that don't support the ISA_DMA_API.

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: das1800: remove depends on ISA_DMA_API limitation

Use the new comedi_isadma module to provide the ISA DMA support. This removes
all the ISA_DMA_API code from this driver and allows the driver to be used on
platforms that don't support the ISA_DMA_API.

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: das16: remove depends on ISA_DMA_API limitation

Use the new comedi_isadma module to provide the ISA DMA support. This removes
all the ISA_DMA_API code from this driver and allows the driver to be used on
platforms that don't support the ISA_DMA_API.

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: pcl812: remove depends on ISA_DMA_API limitation

Use the new comedi_isadma module to provide the ISA DMA support. This removes
all the ISA_DMA_API code from this driver and allows the driver to be used on
platforms that don't support the ISA_DMA_API.

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: pcl816: remove depends on ISA_DMA_API limitation

Use the new comedi_isadma module to provide the ISA DMA support. This removes
all the ISA_DMA_API code from this driver and allows the driver to be used on
platforms that don't support the ISA_DMA_API.

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: pcl818: remove depends on ISA_DMA_API limitation

Use the new comedi_isadma module to provide the ISA DMA support. This removes
all the ISA_DMA_API code from this driver and allows the driver to be used on
platforms that don't support the ISA_DMA_API.

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: comedi_isadma: introduce helper module for ISA DMA

Introduce a new helper module to consolidate all the ISA DMA code. This will
allow removing the "depends on ISA_DMA_API" from the legacy drivers that can
use DMA for async command support.

This module is complete based on the various uses of ISA DMA in the comedi
drivers.

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: das1800: tidy up das1800_ai_transfer_size()

For aesthetics, pass the fill time 'ns' as a parameter to this function.

Refactor this function to calculate the transfer size in 'samples' instead
of 'bytes'. This removes the need to constantly multiply the values by the
'sample_size'. It also helps avoid any possible integer overflow issues.

Use the comedi_nsamples_left() helper to limit the samples when cmd->stop_src
is TRIG_COUNT.

Use comedi_samples_to_bytes() to return the final DMA size in bytes.

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: das1800: remove hard coded 'sample_size'

Use the comedi_bytes_per_sample() helper to remove the hard coded sample_size
in suggest_transfer_size().

The helper function needs to comedi_subdevice pointer. Change the parameters
passed to suggest_transfer_size() and setup_dma() so that this pointer is
available. 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: ni_labpc_isadma: introduce labpc_isadma_program()

Introduce a helper function to program the ISA DMA controller. Program
the ISA DMA as described in Documentation/DMA-ISA-LPC.txt.

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: ni_labpc_isadma: move dma programming out of labpc_drain_dma()

The external caller of labpc_drain_dma() does not enable the DMA transfer. Only
the call from handle_isa_dma() results in the programmed DMA operation getting
enabled.

For aesthetics, move the dma programming to handle_isa_dma(). Make sure the
DMA operation would actually do something (dma->size != 0) before programming
it to avoid enabling the DMA at the end of a stop_src == TRIG_COUNT 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: ni_labpc_isadma: tidy up labpc_drain_dma()

Tidy up the code that determines the number of samples to read for the
current DMA transfer and how many samples are needed for the next DMA,

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: ni_labpc_isadma: introduce labpc_isadma_disable()

Introduce a helper function to disable and the ISA DMA controller and
return the residue as described in Documentation/DMA-ISA-LPC.txt.

The DMA will always be disabled when labpc_setup_dma() is called. Remove
the unnecessary disable in that 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>

staging: comedi: ni_labpc_isadma: convert 'dma_buffer_size' to a define

For aesthetics, convert this statis const global variable to a define.

Cleanup the maximum 'size' calc in labpc_suggest_transfer_size(). The modulo
operation will always result in '0'.

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: ni_labpc_isadma: remove hard coded samples size

Use the comedi_bytes_per_sample() helper and remove the hard coded sample
size.

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: ni_at_a2150: introduce a2150_isadma_disable()

Introduce a helper function to disable and the ISA DMA controller and
return the residue as described in Documentation/DMA-ISA-LPC.txt.

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: ni_at_a2150: introduce a2150_isadma_program()

Introduce a helper function to program the ISA DMA controller. Program
the ISA DMA as described in Documentation/DMA-ISA-LPC.txt.

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: pcl812: introduce pcl812_isadma_disable()

According to Documentation/DMA-ISA-LPC.txt, the DMA lock needs to be
claimed before using any of the ISA DMA routines. Introduce a helper
function to disable the ISA DMA controller and add the necessary
locking calls.

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: pcl812: introduce pcl812_isadma_program()

Introduce a helper function to program the ISA DMA controller. Program
the ISA DMA as described in Documentation/DMA-ISA-LPC.txt.

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: das16: introduce das16_isadma_disable()

Introduce a helper function to disable and the ISA DMA controller and
return the residue as described in Documentation/DMA-ISA-LPC.txt.

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: das16: introduce das16_isadma_program()

Introduce a helper function to program the ISA DMA controller. Program
the ISA DMA as described in Documentation/DMA-ISA-LPC.txt.

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: das16: move dma transfer 'size' to dma descriptor

For asethetics, move the variable that holds the dma transfer 'size' into
the dma descriptor.

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: das1800: introduce das1800_isadma_disable()

Introduce a helper function to disable and the ISA DMA controller and
return the residue as described in Documentation/DMA-ISA-LPC.txt.

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: das1800: introduce das1800_isadma_program()

Introduce a helper function to program the ISA DMA controller. Program
the ISA DMA as described in Documentation/DMA-ISA-LPC.txt.

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: das1800: move dma transfer 'size' to dma descriptor

For asethetics, move the variable that holds the dma transfer 'size' into
the dma descriptor.

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: das1800: pass dma descriptor to das1800_flush_dma_channel()

Instead of passing the dma descriptor 'chan' and 'virt_addr' just pass the
dma descriptor.

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: introduce dt282x_isadma_disable()

According to Documentation/DMA-ISA-LPC.txt, the DMA lock needs to be
claimed before using any of the ISA DMA routines. Introduce a helper
function to disable the ISA DMA controller and add the necessary
locking calls.

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: introduce dt282x_isadma_program()

Introduce a helper function to program the ISA DMA controller. Program
the ISA DMA as described in Documentation/DMA-ISA-LPC.txt.

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: pcl816: introduce pcl816_isadma_disable()

According to Documentation/DMA-ISA-LPC.txt, the DMA lock needs to be
claimed before using any of the ISA DMA routines. Introduce a helper
function to disable the ISA DMA controller and add the necessary
locking calls.

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: pcl816: introduce pcl816_isadma_program()

Introduce a helper function to program the ISA DMA controller. Program
the ISA DMA as described in Documentation/DMA-ISA-LPC.txt.

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: pcl818: introduce pcl818_isadma_disable()

According to Documentation/DMA-ISA-LPC.txt, the DMA lock needs to be
claimed before using any of the ISA DMA routines. Introduce a helper
function to disable the ISA DMA controller and add the necessary
locking calls.

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: pcl818: introduce pcl818_isadma_program()

Introduce a helper function to program the ISA DMA controller. Program
the ISA DMA as described in Documentation/DMA-ISA-LPC.txt.

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: das16: board is not a PCI device

The DAS16 board is an ISA device not a PCI device. For aesthetics, use
dma_{alloc,free}_coherent() to allocate and free the DMA buffers instead
of the PCI versions.

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: das16: introduce das16_free_dma()

For aesthetics, move the freeing of the DMA channel and the buffers to
a helper 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>

staging: comedi: das16 introduce das16_alloc_dma()

DMA is optional with this driver. Introduce a helper function to request
the DMA channel and allocate the buffers. Don't fail the driver attach
if the user passed an invalid DMA channel or the channel cannot be
requested.

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: das16: introduce struct das16_dma_desc

For aesthetics, introduce a struct to hold the DMA descriptor 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: dt282x: remove VIRT_TO_BUS dependancy

Use dma_{alloc,free}_coherent() to allocate and free the DMA buffers.
This removes the dependancy on VIRT_TO_BUS.

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: introduce struct dt282x_dma_desc

For aesthetics, introduce a struct to hold the DMA descriptor 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: dt282x: introduce dt282x_alloc_dma()

The IRA and DMA are optional with this driver but both are required to
support async commands. Introduce a helper function to request the IRQ
and DMA channel and allocate the buffers. Don't fail the driver attach
if the user passed an invalid IRQ or DMA channel or they cannot be
requested.

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: pcl812: remove VIRT_TO_BUS dependancy

Use dma_{alloc,free}_coherent() to allocate and free the DMA buffers.
This removes the dependancy on VIRT_TO_BUS.

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: pcl812: introduce struct pcl812_dma_desc

For aesthetics, introduce a struct to hold the DMA descriptor 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: pcl812: introduce pcl812_free_dma()

For aesthetics, move the freeing of the DMA channel and the buffers to
a helper 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>

staging: comedi: pcl812: introduce pcl812_alloc_dma()

DMA is optional with this driver. Introduce a helper function to request
the DMA channel and allocate the buffers. Don't fail the driver attach
if the user passed an invalid DMA channel or the channel cannot be
requested.

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: pcl816: fix short DMA transactions

When the cmd->stop_src == TRIG_COUNT the last DMA transfer might be smaller
than the buffer size. This results in invalid data being added to the async
buffer.

Add a 'size' member to the DMA descriptor and initialize it with the
actual size of the DMA transfer. Use that in interrupt and ai subdevice
(*poll) function to return the proper number of samples. Use the
comedi_bytes_to_samples() helper to convert the byte size to comedi
samples in the interrupt handler.

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: pcl816: remove VIRT_TO_BUS dependancy

Use dma_{alloc,free}_coherent() to allocate and free the DMA buffers.
This removes the dependancy on VIRT_TO_BUS.

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: pcl816: introduce struct pcl816_dma_desc

For aesthetics, introduce a struct to hold the DMA descriptor 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: pcl816: introduce pcl816_free_dma()

For aesthetics, move the freeing of the DMA channel and the buffers to
a helper 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>