This driver no longer supports a "legacy" attach mechanism that searches
for a suitable PCI device and increments it's reference count, but since
the common "detach" handler `pci224_detach()` still has a left-over
`pci_dev_put()`, a matching `pci_dev_get()` is needed in the
"auto_attach" handler `pci224_auto_attach()`.  There is no longer any
reason to "get" and "put" the PCI device, so those calls can be removed.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Since the driver no longer supports options in its "legacy" attach
mechanism to describe the jumper settings (or any options beyond
restricting a PCI search to a particular bus and/or slot), there is no
need to retain this mechanism in the driver.  Remove the comedi driver
"attach" handler `pci224_attach()`, and the now unused
`pci224_find_pci_dev()`.  Also, remove the "wildcard" entry from the
board table `pci224_boards[]` as it is no longer needed.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

When attaching a PCI224 or PCI234 manually via the `COMEDI_DEVCONFIG`
ioctl, there are several options the user can supply that describe the
state of the hardware jumpers (LK1 for PCI224, LK1 thru LK5 for PCI234).
These options control how the driver sets up the AO range tables for the
device.  Those options are useless when the board is attached
automatically via the PCI driver probe function
`amplc_pci225_pci_probe()`, `comedi_pci_auto_config()`, and the
comedi driver "auto_attach" handler `pci224_auto_attach()`.

Rip out the range table selection options and use a single, static range
table per board type, containing all the software- and
hardware-selectable ranges for that board.  The PCI234 used to have a
per-channel `range_table_list` rather than an all-channel `range_table`,
as the jumpers selected different ranges for all channels.  Now that the
channels are using a unified range table, use an all-channel
`range_table` instead.

When checking the channel list for an asynchronous command in
`pci224_ao_check_chanlist()` make sure the ranges specified in the list
have compatible jumper settings.  We don't know how the jumpers are
actually set, but we can at least avoid conflicting settings.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove the "address-of" operator when the operand is a function.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

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

Remove some pairs of parentheses that don't really improve readability.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Since anything non-zero is logically "true", don't bother doing
"not-equal" comparisons with zero, except when testing for an explicit
number 0 (not as a result of bit tests for example).
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

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

Add a bit of whitespace to the initializer of `pci224_boards[]` for
aesthetic reasons.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix checkpatch issue: "CHECK: spinlock_t definition without comment".
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix checkpatch issue: "CHECK: multiple assignments should be avoided".
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix checkpatch issues: "CHECK: Blank lines aren't necessary before a
close brace '}'".
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

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

Reformat comments to fit in with the preferred coding style, including
the copyright and comedi driver description comments at the start of the
file.  Also, remove a boiler-plate comment for the comedi device private
data structure.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Fixed coding style issue "Missing a blank line after declarations"
detected by the 'checkpatch.pl' script.
Signed-off-by: Tobenna P. Igwe <ptigwe@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

There was a missing line between declarations and code in vbusdeviceinfo.h.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Clean up the format of the vmcall functions in iovmcall_gnuc.h. These functions
are rewritten for clarity and to correct the indention, without changing any
functionality.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

There are spacing errors in the functions in iovmcall_gnuc. Fix these.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The controlVM status file had comments going over the 80 character limit. These
are moved to the next line and the spacing is fixed.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Correct a couple of missing blank lines in uisutils.h.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove the uiscmpxchg64 macro from uisqueue.h and uisqueue.c.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Several macros in timskmod.h are unused. Remove them.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove all of the semaphore macros from timskmod.h and switch all uses of those
types to the correct function names.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Fix all whitespace formatting issues in timskmod.h.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Delete the unused common VMMIO types in timskmod.h.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Delete commontypes.h, and replace all of the places that #included it with
correct #includes for the types used in that file.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

All of the remaining macros in commontypes.h - the CHANNEL_*_MISMATCH macros
and UltraLogEvent() - are used only in channel.h. Move the entire set of them
to their new home.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The only place OFFSETOF was being used was in iochannel.h. Remove the macro
from commontypes.h and replace all uses of it in iochannel.h with offsetof()
instead.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The typedef for GUEST_PHYSICAL address is only used in two files, so remove it
from commontypes.h and add it to visorchipset/file.c and controlvmchannel.h.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

There was a macro for INLINE. Delete it and fix the one occurrance of it to
just use 'inline' instead, though that function looks awfully big for an
inline...
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove the MEMORYBARRIER and VolatileBarrier() defines from commontypes.h and
convert the spots that used this to mb(). Add comments to each use to indicate
that the barrier is used to ensure channel synchronization between guests.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes MEMCPY, MEMCPY_FROMIO, and MEMCPY_TOIO from commontypes.h,
and switches all use of these macros to the appropriate built in definition.
Signed-off-by: Benjamin Romer <benjamin.romer@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

* use dev_{warn,err} instead of pr_{warn,err}
* removes dgap_err() and just use pr_err(). pr_err() used in
dgap_parsefile() not dev_err() because if dgap_parsefile() is
failed, just one message is printed.
* removes "out of memory" messages.
Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

It is same manner with setting a board type.
After allocating a type of "MNODE", get a token value
set to "module.type".
Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

It is same manner with setting a board type.
For example of config file for concentrator,
    "conc ccon" or
    "conc epcon"

After allocating a type of "CNODE" then set a type of concentrator.
So remove cases in swith statement, just get a token from string
and set to "conc.type". And also it doesn't need to "conc.v_type".
Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Board types need to separate normal command like IO, MEM and so on.
And the board type will come after "board" string in config file normally.
(If it is not, dgap_gettok returns an error with zero)
After that, set a variable of a number which is matched with specific
a board number to "board.type". The dgap_gettok() returns that number so
just set to "board.type" and also "v_type" can be removed.

In case of boards of PCI type are set variables to zero. These can
be removed because "p" as cnode get memory from kzalloc so already
set to zero.
Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

If true_count is not same with brd->nasync, warning messages are
printed. But it has duplicated messages within if statement.
Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

These are already defined in dgap.h.
Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dgap_major_serial_registered and dgap_major_transparent_print_registered
could be checked whether a board is initialized.
But it doesn't need to check that variables becasue dgap module
isn't calling the dgap_cleanup_tty() without initializing
for a board completely.
Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

When a configration file is parsed with dgap_parsefile(),
makes nodes for saving configrations for board.

Making a node will allocate node memory and strings for saving
configrations with kstrdup().

So these are freed when dgap is unloaded or failed to initialize.
Signed-off-by: Daeseok Youn <daeseok.youn@gmail.com>
Tested-by: Mark Hounschell <markh@compro.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>