Don't hide those small functions behind ifdefs.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are no ia_css_set_system_mode() nor
ia_css_is_system_mode_suspend_or_resume() functions at the driver.

So, get rid of the code that would try to call it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

It doesn't make any sense to change the number of parameters
for this function depending on the ISP version.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The css_trace header for ISP2401 also builds on older versions, and
seems to be compatible with all versions. So, remove all ifdefs
in favor of the CSP2401 version.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This patch addresses what it sounds to be a change at the
name of some ACPI registers on newer ACPI tables.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The CSS version returned by ISP2400 is different than the one
returned by ISP2401.

While we could return just one version for both, as this sounds
like just an informative string, for now, let's keep returning
different versions, as we don't know if this would affect
userspace.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

As namespaces aren't duplicated here, just remove the ifdefs.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Instead of packing parameters differently on ISP2400 and ISP2401,
use just one way of passing them for both.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The way atomisp_dfs_tables.h is defined, it ends by duplicating
all data structs there on both atomisp_v4l2.c and atomisp_cmd.c.

Change the logic in order to place the definitions only on a single
place.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This header is really version-independent. So, just get rid
of the macros from it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This function call has two parameters that are used only with
ISP2401, enclosed on some ugly ifdefs. Make the function independent,
passing NULL values for ISP2400.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Use the same struct for both ISP2400 and ISP2401.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are several #ifdefs checking for ISP version there. Some
of them are just two different ways to represent the same contants,
while 3 parameters are actually different, depending on the ISP
version.

Change the header in a way that it will be compatible with both
versions, and change dependent code to keep running, removing
ifdefs there only when possible.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are two ioctls that are only available with ISP2401. Yet,
at the compat level, we don't really need to take care, as
the native ioctl handler will already return an error code if
the ioctl doesn't exist.

So, let's just remove the ifdefs here.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Remove ifdefs that check ISP version from the code, switching
to specific ISP-dependent code at runtime.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The ifdefs there are meaningless. Just remove them for good.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Remove ISP-version-dependent ifdefs.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Add a runtime check to use the proper wdt timer init at runtime,
depending on the chipset revision.

For now, we can't get rid of the remaining version checks, as
the rest of the code is not prepared yet to detect the ISP
version on runtime.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Replace #ifdef occurrences there with runtime checks.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Those ifs can easily be removed without breaking the code.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There's a dependency on this header for the ISP model. While
this sounds really weird (as just one resolution needs it),
as we don't know what's the right value, let's just keep it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The logic there has lots of ifdef dependencies if the hardware
is either ISP2400 or ISP2041.

Replace them by runtime checks.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The atomisp supports two different chipsets: ISP2400 and ISP2401.
Right now, this is controlled by ugly #defines inside the driver.

Add a global bolean to identify the type of hardware. While this
is hacky, it would be a quick way to start removing the ugly
ifdefs.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are some uneeded defines there. Simplify it, and make
it independent of defines.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are several parts of atomisp that are meant to be
built on different environments, tested using ifdefs.

Remove some of them, as this code should build only on
Linux.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Use some auto-reformat tools to make the atomisp style
a little better. There are still lots of weird things there,
but this will hopefully reduce the number of pure coding
style patches submitted upstream.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Use checkpatch.pl --fix-inplace --strict to solve several
coding style issues, manually reviewing the produced code and
fixing some troubles caused by checkpatch.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Running checkpatch.pl codespell logic found several typos at atomisp
driver.

Fix them using --fix-inline.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This driver needs IOSF_MBI in order to talk with some PM
registers. Select it at compile time.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The atomisp driver builds again. So, remove depends on BROKEN.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There's a typo at the compat32 code, with forgot to get the
pointer address, causing the driver to not build.

Not sure why this didn't produce an error back when the
driver got removed.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The old ida API got replaced by a new one, with avoids
locking issues. As the old API was removed, start using the
new one, as defined by changeset b03f8e43 ("ida: Remove old API").

Fixes: b03f8e43 ("ida: Remove old API")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Those functions got renamed. Update them on atomisp driver.

Fixes: 67b06ba0 ("PM: QoS: Drop PM_QOS_CPU_DMA_LATENCY and rename related functions")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

The time 64 API patchset changed the ts stamp to u64.

Update this driver accordingly.

Fixes: 15a40b27 ("media: videobuf: use u64 for the timestamp internally")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This type was renamed in the past by a more meaningul
name. Change it on atomisp too.

Fixes: 238e4a5b ("media: rename VFL_TYPE_GRABBER to _VIDEO")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Fix the usage of totalram_pages, as this is now a function.

Fixes: ca79b0c2 ("mm: convert totalram_pages and totalhigh_pages variables to atomic")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

This macro had its first parameter lost. Remove it.

While on it, fix the alignments where this macro is used.

Fixes: 96d4f267 ("Remove 'type' argument from access_ok() function")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

From now on, I'll be maintaining the atomisp driver, and
Sakari will be reviewing it.
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

There are some interest on having this driver back, and I
can probably dedicate some time to address its issue. So,
let's ressurect it.

For now, the driver causes a recursive error and doesn't
build, so, make it depend on BROKEN.

This reverts commit 51b8dc51.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>

Parse device properties and register controls for them using the newly
introduced helpers.
Signed-off-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>