Without CONFIG_PCI, config space reads never return any data,
leading to undefined behavior that gcc warns about:

platform/intel-mid/intel_mid_pcihelpers.c: In function 'intel_mid_msgbus_read32_raw':
platform/intel-mid/intel_mid_pcihelpers.c:66:9: error: 'data' is used uninitialized in this function [-Werror=uninitialized]
platform/intel-mid/intel_mid_pcihelpers.c: In function 'intel_mid_msgbus_read32_raw_ext':
platform/intel-mid/intel_mid_pcihelpers.c:84:9: error: 'data' is used uninitialized in this function [-Werror=uninitialized]
platform/intel-mid/intel_mid_pcihelpers.c: In function 'intel_mid_msgbus_read32':
platform/intel-mid/intel_mid_pcihelpers.c:137:9: error: 'data' is used uninitialized in this function [-Werror=uninitialized]

With a dependency on CONFIG_PCI, we don't get this warning. This seems
safe as PCI config space accessors should always return something
when PCI is enabled.

Fixes: a49d2536 ("staging/atomisp: Add support for the Intel IPU v2")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The driver fails to build if this is disabled, so we need an explicit
Kconfig dependency:

drivers/staging/media/atomisp/pci/atomisp2/./atomisp_cmd.c:6085:48: error: 'struct v4l2_subdev_fh' has no member named 'pad'

Fixes: a49d2536 ("staging/atomisp: Add support for the Intel IPU v2")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

I ran into a build warning on my randconfig build box:

drivers/staging/media/atomisp/pci/atomisp2/css2400/sh_css_params.c: In function 'ia_css_lace_statistics_free':
drivers/staging/media/atomisp/pci/atomisp2/css2400/sh_css_params.c:2845:64: error: parameter 'me' set but not used [-Werror=unused-but-set-parameter]

It turns out that not only the parameter is unused but the entire function has no
caller. Let's just remove it.

Fixes: a49d2536 ("staging/atomisp: Add support for the Intel IPU v2")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

ia_css_dequeue_param_buffers does not have an arguement type, causing a warning:

drivers/staging/media/atomisp/pci/atomisp2/css2400/sh_css_params.c: In function 'ia_css_dequeue_param_buffers':
drivers/staging/media/atomisp/pci/atomisp2/css2400/sh_css_params.c:3728:6: error: old-style function definition [-Werror=old-style-definition]

This adds a 'void' keywork to silence the warning.

Fixes: a49d2536 ("staging/atomisp: Add support for the Intel IPU v2")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Defining a debug function to nothing causes a warning with an empty block
after if()/else():

drivers/staging/media/atomisp/i2c/ov2680.c: In function 'ov2680_s_stream':
drivers/staging/media/atomisp/i2c/ov2680.c:1208:55: error: suggest braces around empty body in an 'else' statement [-Werror=empty-body]

This changes the empty debug statement to dev_dbg(), which by default also
does nothing, but avoids this warning and also checks the format string.
As a side-effect, we can now use dynamic debugging to turn on the
output at runtime.

Fixes: a49d2536 ("staging/atomisp: Add support for the Intel IPU v2")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Two dependencies were missing to build atomisp drivers:

_ MEDIA_CONTROLLER: to use the entity field of v4l2_subdev structure. Since
every atomisp driver needs MEDIA_CONTROLLER has a dependency, let's add it
to INTEL_ATOMISP

_ EFI: to use efivar_entry_get:
drivers/built-in.o: In function `gmin_get_config_var':
(.text+0xe062b): undefined reference to `efivar_entry_get'
Signed-off-by: Jérémy Lefaure <jeremy.lefaure@lse.epita.fr>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

REGMAP_I2C should be enabled to build the driver ap1302 because it uses
regmap functions.
Signed-off-by: Jérémy Lefaure <jeremy.lefaure@lse.epita.fr>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

There are some error paths in atomisp_css_frame_allocate() which don't
initialize "res" so it could lead us to try release random memory.

Fixes: a49d2536 ("staging/atomisp: Add support for the Intel IPU v2")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

We call this function with the lock held and should also return with the
lock held as well.  This one error path is not-consistent because we
should return without the lock held.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The first of these checks the PCI identifier in order to decide what to do so needs no
ifdef. The other is simply a variation on what is dumped for debug - so favour dumping the
most.

Signed-off-by Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This was reported by Dan Carpenter. When we error with an IMX 227 we don't release
the lock and the sensor would then hang.
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This is a do nothing function so we can replace it with nothing and eliminate it entirely.
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Our driver only ever uses one set of routines for the allocators used by the CSS layer to
manage memory and the memory management on the ISP. We can thus remove the function vectors
and simply call the intended routines directly.

These routines in turn are simply wrappers around another layer of code so remove this
second layer of wrappers and call the hrt methods directly. In time we can remove this layer
of indirection as well.
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

We don't need this layer of indirection and the debugging information is not used. With
this removed we can then go on to try and remove the abstraction layer entirely.
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The aa kernel is used but it consists of nothing more than a set of wrappers
for a memset and an assignment. Replace these at the calling points with the
memset and assignment.

Keep the structures for now - those should disappear as the next layer up
gets unwrapped.
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

While this is included and the headers pulled in nothing actually uses this
functionality in the driver, so remove it.
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In an effort to create a more uniform coding style within the Unisys
s-Par driver set, this patch adjusts the formatting of all #define
directives within this source file to match the following template,
and thereby eliminate irregular usage of whitespace:
Reviewed-by: Tim Sell <timothy.sell@unisys.com>

The amount of whitespace used between the <token> and the <value> is
dependent on what is needed to make the surrounding #define directives
as uniform as possible.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes the #pragma directive from visorchannel.h
and adds the __packed keyword to all structs to suppress structure
padding.
Signed-off-by: Jon Frisch <jon.frisch@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes inline functions in visornic_main.c.
Signed-off-by: Jon Frisch <jon.frisch@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes comments that are no longer relevant
in channel.h.
Signed-off-by: Jon Frisch <jon.frisch@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes the unused function declarations
in channel.h.
Signed-off-by: Jon Frisch <jon.frisch@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch removes the unused macro lengthof.
Signed-off-by: Jon Frisch <jon.frisch@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Removes unnecessary cast of to __iomem of a non-io-mapped pointer, thereby
eliminating the following sparse warning:

visorchannel.c:159:17: warning: cast adds address space to expression (<asn:2>)
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

The function spar_check_channel_client shouldn't need to do
readq's, it is referencing a local copy of the channel
header. Simplify it to just access the fields directly.
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Adds identifier to function definition arguments to satisfy checkpatch
warnings:

WARNING: function definition argument 'struct visor_driver *' should also
	have an identifier name
WARNING: function definition argument 'struct visor_driver *' should also
	have an identifier name
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove local variable on stack by directly returning the value in the
array.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Renames __IOMONINTF_H__ to __VMCALLINTERFACE_H__, which more closely
reflects the naming conventions of the code.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Move function add_physinfo_entries() to visornic_main.c, which is the
only function where it is used.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Replace references to virtpci to visornic in iochannel.h.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Rename #define VMCALL_IO_CONTROLVM_ADDR to VMCALL_CONTROLVM_ADDR, as this
vmcall can be used by any partition, not just the IO partition.
Signed-off-by: David Binder <david.binder@unisys.com>
Signed-off-by: David Kershner <david.kershner@unisys.com>
Reviewed-by: Tim Sell <timothy.sell@unisys.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>