In the future will have arbitrarily long video pipeline chains, instead
of the current two-entities-per-pipeline model.

This patch changes the affected get/find style functions so that they
properly go through the video pipeline chain, for example when getting
the overlay manager connected to a given display.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

In order to allow multiple display block in a video pipeline, we need to
give the drivers way to register themselves. For now we have
the omapdss_register_display() which is used to register panels, and
dss_register_output() which is used to register DSS encoders.

This patch makes dss_register_output() public (with the name of
omapdss_register_output), which can be used to register also external
encoders. The distinction between register_output and register_display
is that a "display" is an entity at the end of the videopipeline, and
"output" is something inside the pipeline.

The registration and naming will be made saner in the future, but the
current names and functions are kept to minimize changes during the dss
device model transition.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Disable overlay via ovl->disable() interface, which will
properly set flags in cache and GO bits for managers.
This allows overlay user to re-enable it on next frame,
thus recovering from FIFO underflows.
Signed-off-by: Sergey Kibrik <sergiikibrik@ti.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

It's not necessary to free memory allocated with devm_kzalloc
in a remove function and using kfree leads to a double free.
Signed-off-by: Emil Goode <emilgoode@gmail.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

DISPC needs to know the clock rate for DIGIT (i.e. TV) channel, and this
clock is provided by either VENC or HDMI modules. Currently DISPC will
call a function in VENC/HDMI, asking what the clock rate is. This means
we have a fixed dependency from DISPC to both VENC and HDMI.

To have a more generic approach, and in particular to allow adding OMAP5
HDMI driver, we need to remove this dependency. This patch makes
VENC/HDMI inform DISPC when the their clock changes, thus reversing the
dependency and removing the issue.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

The use of platform callbacks, backlight, DSI TE and reset gpio from the
struct omap_dss_device has been removed. We can thus remove the fields
from omap_dss_device.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

hpd_gpio is no longer used by the OMAP4 HDMI IP driver, and we can thus
remove the unnecessary code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

The TRM tells to set PHY to TXON only after getting LINK_CONNECT, and to
set PHY to OFF or LDOON after getting LINK_DISCONNECT, in order to avoid
damage to the PHY.

We don't currently do it quite like that. Instead of using the HDMI
interrupts, we use HPD signal. This works, but is not actually quite
correct, as HPD comes at a different time than LINK_CONNECT and
LINK_DISCONNECT interrupts. Also, the HPD GPIO is a property of the TPD
level shifter, not HDMI IP, so handling the GPIO in the HDMI driver is
wrong.

This patch implements the PHY power handling correctly, using the
interrupts.

There is a corner case that causes some additional difficulties: we may
get both LINK_CONNECT and LINK_DISCONNECT interrupts at the same time.
This is handled in the code by retrying: turning off the PHY, clearing
the interrupt status, and re-enabling the PHY. This causes a new
LINK_CONNECT interrupt to happen if a cable is connected.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Currently omapfb returns EPROBE_DEFER if no displays have been probed at
the time omapfb is probed. However, sometimes some of the displays have
been probed at that time, but not all. We can't return EPROBE_DEFER in
that case, because then one missing driver would cause omapfb to defer
always, preventing any display from working.

However, if the user has defined a default display, we can presume that
the driver for that display is eventually loaded. Thus, this patch
changes omapfb to return EPROBE_DEFER in case default display is not
found.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Now that omap_dss_output has been combined into omap_dss_device, we can
add ref counting for the relevant output functions also.

This patch adds omap_dss_get_device() calls to the various find_output()
style functions. This, of course, means that the users of those
find_output functions need to do a omap_dss_put_device() after use.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Setup the owner field for DSS output's omap_dss_device so that module
refcounting works.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

omap_dss_get_device() should be called for omap_dss_device before it is
used to increase its refcount. Currently we only increase the refcount
for the underlying device.

This patch adds managing the ref count to the underlying module also,
which contains the ops for the omap_dss_device.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Add struct module *owner field to omap_dss_device, which points to the
module containing the ops for this omap_dss_device. This will be used to
manage the ref count for the module.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

We currently have omap_dss_device, which represents an external display
device, sometimes an external encoder, sometimes a panel. Then we have
omap_dss_output, which represents DSS's output encoder.

In the future with new display device model, we construct a video
pipeline from the display blocks. To accomplish this, all the blocks
need to be presented by the same entity.

Thus, this patch combines omap_dss_output into omap_dss_device. Some of
the fields in omap_dss_output are already found in omap_dss_device, but
some are not. This means we'll have DSS output specific fields in
omap_dss_device, which is not very nice. However, it is easier to just
keep those output specific fields there for now, and after transition to
new display device model is made, they can be cleaned up easier than
could be done now.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

The omap_dss_start_device() and omap_dss_stop_device(), called by the
DSS output drivers, are old relics. They originally did something
totally else, but nowadays they increase the module ref count for panels
that are enabled.

This model is quite broken: the panel modules may be used even before
they are enabled. For example, configuring the panel requires calls to
functions located in the panel modules.

In the following patches we try to improve the ref count management for
the modules and display devices. The first step, however, is to remove
the omap_dss_start/stop_device() totally.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

We are about to remove the dss bus support, which also means that the
omap_dss_device won't be a real device anymore. This means that the
embedded "dev" struct needs to be removed from omap_dss_device.

After we've finished the removal of the dss bus, we see the following
changes:

- struct omap_dss_device won't be a real Linux device anymore, but more
  like a "display entity".
- struct omap_dss_driver won't be a Linux device driver, but "display
  entity ops".
- The panel devices/drivers won't be omapdss devices/drivers, but
  platform/i2c/spi/etc devices/drivers, whichever fits the control
  mechanism of the panel.
- The panel drivers will create omap_dss_device and omap_dss_driver,
  fill the required fields, and register the omap_dss_device to
  omapdss.
- omap_dss_device won't have an embedded dev struct anymore, but a
  dev pointer to the actual device that manages the omap_dss_device.

The model described above resembles the model that has been discussed
with CDF (common display framework).

For the duration of the conversion, we temporarily have two devs in the
dssdev, the old "old_dev", which is a full embedded device struct, and the
new "dev", which is a pointer to the device. "old_dev" will be removed
in the future.

For devices belonging to dss bus the dev is initialized to point to
old_dev. This way all the code can just use the dev, for both old and
new style panels.

Both the new and old style panel drivers work during the conversion, and
only after the dss bus support is removed will the old style panels stop
to compile.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

We aim to remove the custom omapdss bus totally, as it's quite a strange
construct and won't be compatible with common display framework. One
problem on the road is that we have sysfs files for each display, and
they depend on the omapdss bus.

This patch creates the display sysfs files independent of the omapdss
bus. This gives us backwards compatibility without using the omapdss bus
for the sysfs files.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

We have support functions to suspend and resume all the displays that
are used with system suspend. These functions use the dss bus to iterate
the display devices.

As we aim to remove the custom dss bus totally, this patch removes the
explicit use of dss bus from these functions. Instead the
for_each_dss_dev() macro is used to go through the devices.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

omap_dss_get_next_device() uses the dss bus to iterate over the
displays. This patch changes omap_dss_get_next_device() to use the new
panel list instead.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

We currently use the omapdss bus (which contains all the available
displays) to iterate the displays. As the omapdss bus is on its way out,
this needs to be changed.

Instead of using the dss bus to iterate displays, this patch adds our
own list of displays which we manage. The panels on the dss bus are
automatically added to this new list.

An "alias" field is also added to omap_dss_device. This field is
set to "display%d", the same way as omap_dss_device's dev name is set.
This alias is later used to keep backward compatibility, when the
embedded dev is no longer used.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

In the future the "dssdev" parameter passed to output drivers will
change its meaning. Instead of being a pointer to the panel device, it's
a pointer to the output instance.

To make the transition easier, some of the uses for this dssdev
parameter can be easily removed.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Add helper functions to convert between omapdss specific video timings
and the common videomode.

Eventually omapdss will be changed to use only the common video timings,
and these helper functions will make the transition easier.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Clean up the VENC driver's regulator init to remove the (unused)
omap_dss_device parameter, renaming the function to a more sensible
name, and making the code slightly clearer.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

SDI requires a regulator to operate. This regulator is, for some reason,
currently attached to the virtual omapdss device, instead of the SDI
device. This does not work for DT, as the regulator mappings need to be
described in the DT data, and the virtual omapdss device is not present
there.

Fix the issue by acquiring the regulator in the SDI device. To retain
compatibility with the current board files, the old method of getting
the regulator is kept. The old method can be removed when the board
files have been changed to pass the regulator to SDI.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Clean up the SDI driver's regulator init to remove the (unused)
omap_dss_device parameter, renaming the function to a more sensible
name, and making the code slightly clearer.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Separate regulator init code into its own function for clarity.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

On some platforms DPI requires a regulator to be enabled to power up the
output pins. This regulator is, for some reason, currently attached to
the virtual omapdss device, instead of the DPI device. This does not
work for DT, as the regulator mappings need to be described in the DT
data, and the virtual omapdss device is not present there.

Fix the issue by acquiring the regulator in the DPI device. To retain
compatibility with the current board files, the old method of getting
the regulator is kept. The old method can be removed when the board
files have been changed to pass the regulator to DPI.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Split regulator and DSI PLL init code to their own functions for
clarity.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Separate the regulator initialization code to its own function, removing
duplicate code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Use devm_regulator_get() instead of regulator_get() to simplify code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

We currently have two steps in panel initialization and startup: probing
and enabling. After the panel has been probed, it's ready and can be
configured and later enabled.

This model is not enough with more complex display pipelines, where we
may have, for example, two panels, of which only one can be used at a
time, connected to the same video output.

To support that kind of scenarios, we need to add new step to the
initialization: connect.

This patch adds support for connecting and disconnecting panels. After
probe, but before connect, no panel ops should be called. When the
connect is called, a proper video pipeline is established, and the panel
is ready for use. If some part in the video pipeline is already
connected (by some other panel), the connect call fails.

One key difference with the old style setup is that connect() handles
also connecting to the overlay manager. This means that the omapfb (or
omapdrm) no longer needs to figure out which overlay manager to use, but
it can just call connect() on the panel, and the proper overlay manager
is connected by omapdss.

This also allows us to add back the support for dynamic switching
between two exclusive panels. However, the current panel device model is
not changed to support this, as the new device model is implemented in
the following patches and the old model will be removed. The new device
model supports dynamic switching.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Currently omapdrm creates crtcs, which map directly to DSS overlay
managers, only on demand at init time. This would make it difficult to
manage connecting the display entities in the future, as the code cannot
just search for a suitable overlay manager.

We cannot fix this the sane way, which would be to create crtcs for each
overlay manager, because we need an overlay for each crtc. With limited
number of overlays, that's not possible.

So the solution for now is to detach the overlay manager from the crtc.
crtcs are still created on demand at init time, but all overlay managers
are always initialized by the omapdss.

This way we can create and connect whole display pipelines from the
overlay manager to the display, regardless of which crtcs omapdrm would
create.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Split the function that creates overlay manager structs into two: one
that creates just the structs, and one that creates the sysfs files for
the manager.

This will help us use the overlay manager structs with omapdrm in the
following patches, while still leaving the sysfs files out.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Add two helper functions that can be used to find either the DSS output
or the overlay manager that is connected to the given display.

This hides how the output and the manager are actually connected, making
it easier to change the connections in the future.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Minor cleanup for the dss_[ovl|mgr]_get_device() functions to make them
more readable.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

When using DT, dss device does not have platform data. However,
dss_get_ctx_loss_count() uses dss device's platform data to find the
get_ctx_loss_count function pointer.

To fix this, dss_get_ctx_loss_count() needs to be changed to get the
platform data from the omapdss device, which is a "virtual" device and
always has platform data.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Add a support function to find a DSS output by given DT node. This is
used in later patches to link the panels to DSS outputs.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Add a support function to find a DSS output by given name. This is used
in later patches to link the panels to DSS outputs.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

omapdss output drivers always read the platform data. This crashes when
there's no platform data when using DT.

Add a check to read the platform data only if it exists.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

We can currently set the default display (i.e. the initial display) in
the omapdss platform data by using a pointer to the default
omap_dss_device. Internally omapdss uses the device's name to resolve
the default display.

As it's difficult to get the omap_dss_device pointer in the future,
after we've changed the omapdss device model, this patch adds a new way
to define the default display, by using the name of the display.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Archit Taneja <archit@ti.com>