Most arches have an asm/gpio.h that merely includes linux/gpio.h.  The
others select ARCH_HAVE_CUSTOM_GPIO_H, and when that's selected,
linux/gpio.h includes asm/gpio.h.

Therefore, code should include linux/gpio.h instead of including asm/gpio.h
directly.

Remove includes of asm/gpio.h, adding an include of linux/gpio.h when
necessary.

This is a follow-on to 7563bbf8 ("gpiolib/arches: Centralise
bolierplate asm/gpio.h").
Signed-off-by: NBjorn Helgaas <bhelgaas@google.com>
Acked-by: NThomas Gleixner <tglx@linutronix.de>
Acked-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NAlexandre Courbot <acourbot@nvidia.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

parallel to mutex_{lock,unlock,trylock,is_locked,lock_nested},
inode_foo(inode) being mutex_foo(&inode->i_mutex).

Please, use those for access to ->i_mutex; over the coming cycle
->i_mutex will become rwsem, with ->lookup() done with it held
only shared.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

As seen by Julia, the initial allocation memory is not checked anymore
after commit "video: fbdev: pxafb: initial devicetree conversion".
Introduce back the removed test.
Reported-by: NJulia Lawall <julia.lawall@lip6.fr>
Signed-off-by: NRobert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

This makes the driver use the data pointer added to the gpio_chip
to store a pointer to the state container instead of relying on
container_of().

Cc: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Acked-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

As we want gpio_chip .get() calls to be able to return negative
error codes and propagate to drivers, we need to go over all
drivers and make sure their return values are clamped to [0,1].
We do this by using the ret = !!(val) design pattern.

Cc: Florian Tobias Schandinat <FlorianSchandinat@gmx.de>
Acked-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Now that omapfb has its own copy of omapdss and display drivers, we can
move omapdss and display drivers which omapdrm uses to omapdrm's
directory.

We also need to change the main drm Makefile so that omapdrm directory
is always entered, because omapdss has a file that can't be built as a
module.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Acked-by: NRob Clark <robdclark@gmail.com>

VRFB is only used by omapfb, so we can move it under omapfb's directory.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Acked-by: NRob Clark <robdclark@gmail.com>

omapfb's private copy of omapdss is now ready to be used.

This patch makes omapfb use its private omapdss and display drivers, and
also makes omap_vout (which uses omapfb) to depend on omapfb.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Acked-by: NRob Clark <robdclark@gmail.com>

We need to change the config symbols of omapfb's private copy of the
panel and encoder drivers so that we won't have config symbol conflicts.

This patch changes the symbols from the panel and encoder drivers using
simple replacement of DISPLAY_* to FB_OMAP2*.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Acked-by: NRob Clark <robdclark@gmail.com>

We need to change the config symbols of omapfb's private copy of
omapdss so that we won't have config symbol conflicts.

This patch changes the symbols from omapdss using simple replacement of
CONFIG_OMAP* to CONFIG_FB_OMAP*.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Acked-by: NRob Clark <robdclark@gmail.com>

This patch makes a copy of the omapdss driver and the omap panel &
encoder drivers for omapfb. The purpose is to separate omapdrm and
omapfb drivers from each other.

Note that this patch only does a direct copy of the files without any
other modifications. The files are not yet used.

The original files are in:

drivers/video/fbdev/omap2/dss/
drivers/video/fbdev/omap2/displays-new/

Here's a more detailed explanation about this and the following patches,
from the introduction mail of the patch series:

A short background on the current status. We have the following
entities:

* omapdss, located in drivers/video/fbdev/omap2/dss/. This is a driver for the
  display subsystem IPs used on OMAP (and related) SoCs. It offers only a
  kernel internal API, and does not implement anything for fbdev or drm.

* omapdss panels and encoders, located in
  drivers/video/fbdev/omap2/displays-new/. These are panel and external encoder
  drivers, which use APIs offered by omapdss driver. These also don't implement
  anything for fbdev or drm.

* omapdrm, located in drivers/gpu/drm/omapdrm/. This is a drm driver, which
  uses omapdss and the panel/encoder drivers to operate the hardware.

* omapfb, located in drivers/video/fbdev/omap2/omapfb/. This is an fbdev
  driver, which uses omapdss and the panel/encoder drivers to operate the
  hardware.

* omap_vout, located in drivers/media/platform/omap/. This is a v4l2 driver,
  which uses omapdss and omapfb to implement a v4l2 API for the video overlays.

So, on the top level, we have either omapdrm, or omapfb+omap_vout. Both of
those use the same low level drivers. Without going to the historical details
why the architecture is like that, I think it's finally time to change that.

The situation with omapfb+omap_vout is that it still works, but no new features
have been added for a long time, and I want to keep it working as it's still
being used.  At some point in the future I'd like to remove omapfb and
omap_vout altogether.

Omapdrm, on the other hand, is being actively developed. Sharing the low level
parts with omapfb makes that development more difficult than it should be. It
also "hides" half of the development, as everything happening in the low level
parts resides under fbdev directory, not in the drm directory.

I've been wanting to clean this up for a long time, but I haven't figured out a
very good way to do it. I still haven't, but here's the best way I have come up
with.

This series makes a full copy of the low level parts, omapdss and panel/encoder
drivers. Both omapfb+omap_vout and omapdrm will have their own versions. The
copy omapfb+omap_vout get is a new copy, and the copy that omapdrm gets is just
the current files moved. This way git will associate the omapdrm version with
the old files.

The omapfb+omap_vout versions won't be touched unless there are some big issues
there.

The omapdrm versions can be refactored and cleaned up, as the omapfb support
code is no longer needed. We can perhaps also merge omapdss and omapdrm into
the same kernel module.

This series only does the copy, and the absolutely necessary parts. No further
cleanups are done yet.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Acked-by: NRob Clark <robdclark@gmail.com>

At the moment omapfb and omapdrm can be compiled at the same time, if
both are modules. However, they can't be both loaded, as they use the
same hardware. This has been mostly for compile testing.

To make it clear that omapfb and omapdrm are mutually exclusive drivers,
this patch makes omapfb available only if omapdrm is disabled.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Acked-by: NRob Clark <robdclark@gmail.com>

gpiod_set_value_cansleep is a noop when the passed descriptor is NULL.
So there is no need to duplicate the check for NULL; just call the
function unconditionally instead.
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

This makes .disable operate in reverse order compared to .enable.
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

The WARN() macro has to take a condition.  The current code will just
print the stack trace and the function name instead of the intended
warning message.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Device node iterators perform an of_node_put on each iteration, so putting
an of_node_put before a continue results in a double put.

A simplified version of the semantic match that finds this problem is as
follows (http://coccinelle.lip6.fr):

// <smpl>
@@
expression root,e;
local idexpression child;
iterator i;
@@

 i(..., child, ...) {
   ... when != of_node_get(child)
*  of_node_put(child);
   ...
*  continue;
}
// </smpl>
Signed-off-by: NJulia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Boolean tests do not need explicit comparison to true or false.
Signed-off-by: NLuis de Bethencourt <luisbg@osg.samsung.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

DSI_IRQ_SYNC_LOST was ORed twice so we can remove one.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

All the output drivers check for 'out' being NULL, but it can never be
NULL. Remove the check.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>

A bunch of dispc functions are only used inside dispc, so we can make
them static.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>

dss_feat_get_supported_displays() and dss_feat_get_supported_outputs()
are not used outside omapdss, but are exported. We can thus remove the
export and move the declarations to the omapdss internal header.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>

The functions dispc_ovl_set_fifo_threshold and
dispc_ovl_compute_fifo_thresholds are exported, but not declared in
public headers, and thus are not used outside omapdss. So we can remove
the EXPORT_SYMBOL()s.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>

dispc_wb_setup() handles configuration only for mem-to-mem case. This
patch adds the necessary configuration to handle also display capture
mode.

We need to set CAPTUREMODE to 0 (continuous capture), and WBDELAYCOUNT
according to the vertical timings of the display, so that the WB FIFO
has time to flush before the next frame starts.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

In mem-to-mem mode WB CAPTUREMODE needs to be set to 1 (capture one
frame).
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

The color conversion coefficients for RGB to YUV conversion (used with
writeback) don't result in the correct result. This patch sets the
correct coefficients.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

When doing mem-to-mem writeback, there's no pixelclock. However, the
code that calculates scaling factors check that there is a pixel clock.
We can just skip the check when doing mem-to-mem writeback.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Mflag thresholds for all pipelines are set in dispc_init_mflag(), but we
are missing that for WB pipeline. Add WB configuration.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Fifo thresholds for all pipelines are set in dispc_init_fifos(), but we
are missing it for WB pipeline. Add that.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Burst size for all pipelines is set in dispc_configure_burst_sizes(),
but we are missing WB pipe. So add that.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Add handling of OMAP_DSS_CHANNEL_WB to dispc_ovl_set_channel_out() and
dispc_ovl_get_channel_out().
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Refactor dispc_ovl_get_channel_out() to a bit cleaner form, which makes
it easier to add support for writeback in the following patch.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

At the moment we have a function to get the number of writeback
pipelines supported. However, the function is used in a wrong way,
causing a wrong pipeline to get configured when trying to use WB. Also,
we only have a single writeback pipeline on any of the DSS versions.

To fix and simplify this, create a 'has_writeback' flag into the dispc
driver, and after checking the flag, use OMAP_DSS_WB pipeline
explicitly when calling the configuration functions.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

OMAP5+ DSS has a writeback pipeline, but this was not specified in the
features list for OMAP5 DSS.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Add dumping of the writeback registers to the register dump function.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

DISPC_MFLAG_THRESHOLD_OFFSET() is missing the offset for WB. Add the
offset.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

By default DISPC asserts hsync and vsync sequentially, i.e. there's
first hsync and that is immediately followed by vsync. This is the only
available behaviour on OMAP2/3, and default behaviour on OMAP4+.

OMAP4+ has ALIGN bit in POL_FREQ register, which makes DISPC assert both
syncs at the same time.

It has been observed that some panels don't like sequential syncs (AM5
EVM's panel). After studying the datasheets for multiple panels and
encoders, and MIPI DPI spec, it looks like there is no standard way to
handle this.

Sometimes the datasheets don't mention the required syncs behaviour at
all, sometimes the datasheets have images that hint towards simultaneous
syncs, and sometimes it is explicitly mentioned that simultaneous syncs
are needed. No panels or encoders requiring sequential sync was found.

It thus seems to be safe to default to simultaneous syncs when the ALIGN
bit is available. This fixed AM5 EVM's panel, and no side effects have
been observed on other panels or encoders.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>

Update omapdss driver to set the state of the pins to:
- "default on resume
- "sleep" on suspend

By optionally putting the pins into sleep state in the suspend callback
we can accomplish two things.
- minimize current leakage from pins and thus save power,
- prevent the IP from driving pins output in an uncontrolled manner,
which may happen if the power domain drops the domain regulator.
Signed-off-by: NDave Gerlach <d-gerlach@ti.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

This patch brings a first support of pxa framebuffer devices to a
devicetree pxa platform, as was before platform data.

There are restrictions with this port, the biggest one being the lack of
support of smart panels. Moreover the conversion doesn't provide a way
to declare multiple framebuffer configurations with different bits per
pixel, only the LCD hardware bus width is used.

The patch was tested on both pxa25x, pxa27x and pxa3xx platform (namely
lubbock, mainstone and zylonite).
Signed-off-by: NRobert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

In order to prepare the transition to a mixed platform data and
device-tree initialization, remove all the platform data references all
over the driver.

Copy the platform data into the internal structure of the pxafb, and
only use this afterward.
Signed-off-by: NRobert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

There's no point in having support for framebuffer notifications
is CONFIG_FB is disabled. This commit adds the necessary stubs
for code to link properly when CONFIG_FB=n and moves fb-notify.o
to be built only when CONFIG_FB=y.
Signed-off-by: NEzequiel Garcia <ezequiel@vanguardiasur.com.ar>
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>