The "edge detection interrupt" subdevice supports the
`INSN_CONFIG_CHANGE_NOTIFY` comedi instruction which is only supported
by one other driver. The `INSN_CONFIG_DIGITAL_TRIG` comedi instruction
is more flexible as it supports both edge and level detection, but is
not currently supported by this driver.
Add partial support for `INSN_CONFIG_DIGITAL_TRIG`, but only for edge
detection. Make use of the `ni6527_set_edge_detection()` used for
`INSN_CONFIG_CHANGE_NOTIFY`, but add a parameter holding a mask of the
rising and falling edges to be updated and preserve the unmasked edges
when updating.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This driver currently inverts the outputs for the DO (digital output)
subdevice for some of the boards it supports, indicated by the
`invert_outputs` member in the board-specific data being initialized to
1. It seems this driver shouldn't really be inverting outputs for these
boards at all, but has done so since the driver was first written back
in October 2006. I've had confirmation that for the PCI-6515 at least,
the output voltage levels are opposite to the values set by the user
program.
The driver by Jon Grierson originally supported only PCI-6514 and
PXI-6514 (and was originally called "ni_6514"). The driver was based on
"ni_6527", which is where the inversion of outputs appears to have come
from. Over a period of a few days, the driver was enhanced by Frank
Mori Hess to support other boards. Some of these plainly didn't require
inverted outputs and some guesswork was used to decide which boards
should have inverted outputs. Some of the boards in question are
described in the manual as having "Sink Outputs" and others are
described as having "Source Outputs", but this does not correspond in
any way with which boards are marked as having inverted outputs, so the
criterion that Frank used is a bit of a mystery!
Change the driver so it doesn't invert the outputs of these by boards by
default, but add a module parameter, "legacy_invert_outputs", that can
be set to 'true' to restore the old behaviour. Also rename the
`invert_outputs` member of `struct ni_65xx_board` to `legacy_invert`.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A 'line over 80 characters' fixed by removing comment to the previous
line.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A 'line over 80 characters' fixed by removing comment to
previous line.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This parameter is never used by any of the comedi drivers that provide a
(*buf_change) callback. If the 'new_size' is needed in the callback it can
be found from the 's->async->prealloc_bufsz' as done in the ni_pcidio driver.
Remove the unused parameter.
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 'do_lock' and 'do_unlock' callbacks are not used be any of the comedi
drivers or the comedi core. Just 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>
According to Ian Abbott, this code was added in comedi-0.7.6 back in June 1999
and has always been disabled.
The 'lock_f' and 'unlock' members don't even exist in the comedi_subdevice.
Just remove the disabled 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>
Sparse reports that MimeThread is not used. Actually, it can be used
if THREAD is defined. By enclosing the MimeThread function into the
same #ifdef as the caller of MimeThread, this fixes the sparse
warnings.
Signed-off-by: Guillaume Clement <gclement@baobob.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The IEEE11hbMgrRxAction is not exported and never used. Deleting it
allows to delete other functions that were only used by
IEEE11hbMgrRxAction.
This allows to fix several warnings reported by sparse.
Signed-off-by: Guillaume Clement <gclement@baobob.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some variables are used only in the context of their .c file, which
gives warnings with sparse.
Signed-off-by: Guillaume Clement <gclement@baobob.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There is one function in aes_ccmp.c which is exported, but sparse sees
it unexported because it doesn't include the header that exports it.
Signed-off-by: Guillaume Clement <gclement@baobob.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
PortOffset was an unsigned long, but used as an pointer to io
memory. Sometimes it was not properly cast before use, which caused
many warning by sparse.
By updating its type to void __iomem *, and reflecting the changes
where it is needed, this removes most of those warnings.
Signed-off-by: Guillaume Clement <gclement@baobob.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Declare several functions and variables as static in order
to address the following sparse warnings
drivers/staging/lustre/lnet/lnet/acceptor.c:49:3: warning: symbol 'lnet_acceptor_state' was not declared. Should it be static?
drivers/staging/lustre/lnet/lnet/acceptor.c:79:1: warning: symbol 'lnet_acceptor_get_tunables' was not declared. Should it be static?
drivers/staging/lustre/lnet/lnet/acceptor.c:211:1: warning: symbol 'lnet_accept' was not declared. Should it be static?
drivers/staging/lustre/lnet/lnet/acceptor.c:333:1: warning: symbol 'lnet_acceptor' was not declared. Should it be static?
Tested by compilation only.
Signed-off-by: Matt Kurz <matt@ninezulu.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
sparse reported that gfp_mask was of the wrong type to store gfp flags. The
variable is not used so it can be removed.
Signed-off-by: Guillaume Morin <guillaume@morinfr.org>
Suggested-by: gregkh@linuxfoundation.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
An issue in 'line over 80 characters' fixed by removing over-lined
comment to the previous line.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
An issue in 'line over 80 characters' fixed by removing over-lined comment
to the previous line.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
An issue in 'line over 80 characters' fixed by taking back over-lined comment
a few characters.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
An issue in 'line over 80 characters' fixed and over-lined comment
removed to the previous line.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A 'line over 80 characters' fixed by removing over-lined comment
to previous line.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
By removing a comment to previous line, 'line over 80' issue
fixed.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A 'line over 80' issue fixed, by removing comment to previous line. And
also, by this, 'a blank line' isssue resolved.
Signed-off-by: Sam Asadi <asadi.samuel@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Some of the legacy comedi drivers have a *_{IO,MEM}SIZE define that is
only passed to comedi_request_region() to specify the size of the region.
For aesthetics, remove these defines.
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>
Most of the comedi drivers use the dev->board_name for the id string
passed to request_irq(). Fix the couple that still pass something
else.
Also, propogate the errno from request_irq().
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 usually only used to initialize the driver_name
member of the comedi_driver struct. Some of the drivers also use
the define as part of dev_{level}() messages and/or when doing a
request_irq().
The addition of the DRIVER_NAME to the messages is just added noise
and the dev->board_name is typically used in the request_irq().
Just remove all the DRIVER_NAME defines.
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>
Some of the legacy comedi drivers have a *_SIZE define that is only
passed to comedi_request_region() to specify the size of the region.
Some of the pnp drivers (pci, etc.) also have a *_SIZE define which
is unused.
For aesthetics, remove these defines.
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 only used once in the pci_device_id tables. Remove
them 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>
Move check_channel_list() to remove the need for the forward declaration.
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 private (*io) callback used by this driver is identical to the
default one provided by the 8255 module. Use the default one instead.
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>
Move some of the functions to remove the need for the forward
declarations.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move some of the functions to remove the need for the forward
declarations.
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Most of the comedi drivers that support async commands have some sort
of timer to control the acquisition timing. For these drivers, Step 4
of the (*do_cmdtest) operations calls a ns_to_timer() function that
converts the desired ns time of the command into a value used to set
the timer. These ns_to_timer() functions also typically pass the
cmd->flags in order to determine the desired rounding mode.
Some of the drivers mask the cmd->flags with TRIG_ROUND_MASK when
calling the ns_to_timer() functions. Move all the masking into the
ns_to_timer() functions and just pass the cmd->flags directly.
The cmd->flags member is an unsigned int, change the parameter type
in the the ns_to_timer() functions to match.
For aesthetics, rename the parameter in all the ns_to_timer()
functions to 'flags'.
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 patch adds enable_ints entry to virthba directory
Signed-off-by: Erik Arfvidson <erik.arfvidson@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch adds the rqwait entry into the virthba debugfs directory
Signed-off-by: Erik Arfvidson <erik.arfvidson@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch adds virthba debugfs directory and info entry
Signed-off-by: Erik Arfvidson <erik.arfvidson@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch removes all proc entries, directories, and functions
Signed-off-by: Erik Arfvidson <erik.arfvidson@unisys.com>
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The SIMPLE_DEV_PM_OPS macro already uses the const qualifier, so there's
no need to repeat it.
Signed-off-by: Thierry Reding <treding@nvidia.com>
Acked-by: Marc Dietrich <marvin24@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes the errors report by the checkpatch script.
Most of them are use of C99 comments.
The checkpatch doesn't report any errors after the clean up.
However, please not that there are still ton of warnings left.
Signed-off-by: Petar Dimitrijevic <petar.dimitrijevic@vorteksed.com.mk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
