Add "updating" flag to mgr_priv_data, which is used to track whether the
manager is currently actively working, i.e. updating the screen.

For auto-update displays updating is always true when the display is
enabled, but for manual update displays the flag is true only during
frame transfers.

The flag is used to manage registers properly, for example when updating
is false and a register for a manager is written, there's no need to set
shadow_dirty flag nor set the GO bit, as the settings will be taken into
use by the HW automatically when the output is enabled.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Add dispc_mgr_get_framedone_irq() which returns the irq number for
FRAMEDONE for the given channel.

Note that the function returns always 0 for DIGIT channel, even if OMAP4
does have FRAMEDONE_TV interrupt. The reason for this is that this
function is currently used only to track manual updates, and thus
FRAMEDONE_TV is not needed.

If there's need in the future to also get the FRAMEDONE_TV this needs
revisiting. However, as FRAMEDONE_TV doesn't exist on OMAP2/3, the use
of this function for that purpose needs some extra code to handle the
OMAP2/3 case.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

The functions writing to the DISPC registers in apply.c are more complex
than needed. This patch cleans things up.

Two helper functions are added: need_isr(), which returns if an isr for
VSYNC is needed, and need_go(), which tells if there has been a write to
the shadow registers and we need to set the GO bit.

The functions that call the actual dispc write functions will set the
dirty and shadow_dirty flags themselves, instead of that being done on
the caller side. The writing functions also check if the dirty flag is
even set, and return immediately if not.

This allows us to clean up dss_write_regs().
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Overlays are currently enabled and disabled with a boolean in the struct
omap_overlay_info. The overlay info is set with ovl->set_overlay_info(),
and made into use with mgr->apply().

This doesn't work properly, as the enable/disable status may affect also
other overlays, for example when using fifo-merge. Thus the enabling and
disabling of the overlay needs to be done outside the normal overlay
configuration.

This patch achieves that by doing the following things:

1) Add function pointers to struct omap_overlay: enable(), disable() and
is_enabled(). These are used to do the obvious. The functions may block.

2) Move the "enabled" field from struct omap_overlay to ovl_priv_data.

3) Add a new route for settings to be applied to the HW, called
"extra_info". The status of the normal info and extra_info are tracked
separately.

The point here is to allow the normal info to be changed and
applied in non-blocking matter, whereas the extra_info can only be
changed when holding the mutex. This makes it possible to, for example,
set the overlay enable flag, apply it, and wait until the HW has taken
the flag into use.

This is not possible if the enable flag would be in the normal info, as
a new value for the flag could be set at any time from the users of
omapdss.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Add a new field to mgr_priv data, "busy", which is used to track the
status of GO bit with the manager. As manual update displays never use
the GO bit, the field is always false for managers used for manual
update displays.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

struct omap_overlay_manager contains "enabled"-field, used to track if
the manager is enabled or not. This field should be internal to apply.c.

This patch moves the field to mgr_priv_data, and applies the necessary
locking when accessing the field.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

The current code uses dsi_video_mode_enable/disable functions to
enable/disable DISPC output for video mode displays. For command mode
displays we have no notion in the DISPC side of whether the panel is
enabled, except when a dss_mgr_start_update() call is made.

However, to properly maintain the DISPC state in apply.c, we need to
know if a manager used for a manual update display is currently in use.

This patch achieves that by changing dsi_video_mode_enable/disable to
dsi_enable/disable_video_output, which is called by both video and
command mode displays. For video mode displays it starts the actual
pixel stream, as it did before. For command mode displays it doesn't do
anything else than mark that the manager is currently in use.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

The functions in apply.c, called mostly via function pointers in overlay
and overlay_manager structs, will be divided into two groups. The other
group will not sleep and can be called from interrupts, and the other
group may sleep.

The idea is that the non-sleeping functions may only change certain
settings in overlays and managers, and those settings may only affect
the particular overlay/manager. For example, set the base address of the
overlay.

The blocking functions, however, will handle more complex configuration
changes. For example, when an overlay is enabled and fifo-merge feature
is used, we need to do the enable in multiple steps, waiting in between,
and the change affects multiple overlays and managers.

apply.c already contains a spinlock, which has been used to protect
(badly) the dss_data. This patch adds locks/unlocks of the spinlock to
the missing places, and the lock should now properly protect dss_data.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

The functions in apply.c, called mostly via function pointers in overlay
and overlay_manager structs, will be divided into two groups. The first
group will not sleep and can be called from interrupts, and the second
group may sleep.

The idea is that the non-sleeping functions may only change certain
settings in overlays and managers, and those settings may only affect
the particular overlay/manager. For example, set the base address of the
overlay.

The blocking functions, however, will handle more complex configuration
changes. For example, when an overlay is enabled and fifo-merge feature
is used, we need to do the enable in multiple steps, waiting in between,
and the change affects multiple overlays and managers.

This patch adds the mutex which is used in the blocking functions to
have exclusive access to overlays and overlay managers.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

DSS currently tries to check that the given overlay and overlay manager
settings are acceptable, but the code does not work quite properly.
Things may change between the check and the actual use of the settings.

Furthermore, the following patches will rewrite how settings are
configured and managed, and trying to keep the (broken) settings
checking working during the rewrite would be very difficult.

This patch removes the checking code, and a working version of checking
will be added back after the settings rewrite has been done.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

apply.c will handle the management of dispc in the future patches. This
patch moves overlay manager related functions to apply.c so that they
will have access to the private data and functions of apply.c.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

apply.c will handle the management of dispc in the future patches. This
patch moves overlay related functions to apply.c so that they will have
access to the private data and functions of apply.c.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

dss_cache struct contains private data used to manage dispc. "cache" is
not a good word for it, so rename it to dss_data.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

dss_cache struct contains a spinlock used to protect the struct. A more
logical place for the spinlock is outside the struct that it is
protecting. So move it there.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

manager_cache_data is not a suitable name for the struct. It is more of
a private data for the manager.

Rename the struct to mgr_priv_data, and add a function,
get_mgr_priv(mgr), to get a pointer to the data.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

overlay_cache_data is not a suitable name for the struct. It is more of
a private data for the overlay.

Rename the struct to ovl_priv_data, and add a function,
get_ovl_priv(ovl), to get a pointer to the data.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Make configure_overlay() and configure_manager() take overlay/manager
pointer as an argument, instead of the ovl/mgr id. This will be useful
with the future patches.

Also rename the functions to be a bit more informative:
dss_ovl_write_regs, dss_mgr_write_regs, dss_write_regs.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Use dispc_mgr_get_vsync_irq() to get the interrupt numbers for vsync,
instead of hardcoding the values depending on the display type.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Add dispc_mgr_get_vsync_irq() which returns the irq number for vsync on
the given channel.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Create separate functions for the vsync isr register/unregister code for
cleaner code.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Current way of handling overlay-manager links is a bit strange: each
manager has a static array, containing pointers to all the overlays
(even those used by other managers). The overlays contain a pointer to
the manager being used.

This patch makes the system a bit saner: each manager has a linked list
of overlays, and only the overlays linked to that manager are in the
list.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Overlay managers are stored in a linked list. There's no need for this
list, as an array would do just as fine.

This patch changes the code to use an array for overlay managers.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Overlays are stored in a linked list. There's no need for this list, as
an array would do just as fine.

This patch changes the code to use an array for overlays.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

dss_mgr_start_update() has a loop, of which sole purpose is to find the
manager used for this update. The whole loop is extra, as we already
know the manager.

Remove the loop.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

There's no need to register the vsync ISR and configure the hardware if
the overlay manager is disabled, so this patch adds a check for disabled
managers to the omap_dss_mgr_apply() function.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

The mechanism to cache manager and overlay settings and configure them
into the HW registers in VSYNC is meant only for auto-update displays,
as it doesn't make sense (and doesn't work) for manual-update displays.

This patchs adds a check so that we skip the above for manual update
displays.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Add "enabled" field to struct omap_overlay_manager, which tells if the
output is enabled or not. This will be used in apply.c in the following
patches.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

omap_overlay_manager struct contains enable() and disable() functions.
However, these are only meant to be used from inside omapdss, and thus
it's bad to expose the functions.

This patch adds dss_mgr_enable() and dss_mgr_disable() functions to
apply.c, which handle enabling and disabling the output.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Create a new file, apply.c, and move code about handling the
apply-mechanism and configuration of the managers and overlays from
manager.c to apply.c.

Not all related code is moved in this patch, but only the core
apply/configure functions. The later patches move rest of the code from
overlay.c and manager.c, adding necessary locking at the same time.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

omap_dss_mgr_apply currently applies settings to all overlays and
overlay managers. The reason for this was to support cases where
configuration changes affecting multiple managers are made. However, the
current code doesn't support changing such configurations, so the
functionality is not needed.

Change the apply to affect only the manager given as an argument, and
the overlays attached to that manager.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Split omap_dss_mgr_apply() into smaller functions for clarity.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Add helper functions ovl_manual_update() and mgr_manual_update() which
return whether the overlay or manager is used with a manual update
display.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

There's an unnecessary wrapper function, dss_ovl_wait_for_go(), which
does nothing else than call dss_mgr_wait_for_go_ovl(). The
dss_ovl_wait_for_go() function can be removed.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Remove unused functions:

dispc_mgr_get_default_color
dispc_mgr_get_trans_key
dispc_mgr_trans_key_enabled
dispc_mgr_alpha_fixed_zorder_enabled
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Change manager configuration to be similar to overlay configuration by
creating dispc_mgr_setup() which takes omap_overlay_manager_info as
parameter.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Overlay channel is configured with ovl_setup, with all the other overlay
attriutes. This patch separates overlay channel setup so that we can
later configure the channel only when needed.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Overlay FIFO thresholds are configured with ovl_setup, with all the
other overlay attributes. This patch separates FIFO threshold setup so
that we can later configure FIFO thresholds only when needed.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Move the configuration of interlace and replication from
omap_dss_mgr_apply() to configure_overlay(). This removes the need to
store the values into the cache data.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

Overlay out_width/height are set to 0 when scaling is not used by the
users of omapdss. Currently ovl_setup() expects the caller of
ovl_setup() to convert those zero values to width or height.

This patch makes ovl_setup() accept zero values for out_width/height,
making calling ovl_setup() a bit simpler as the overlay_info can be
just passed to this function without modifications.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>

dss_start_update() takes currently the dss device as a parameter. Change
the parameter to ovl manager, as that is what the dss_start_update()
actually needs. Change the name of the function to
dss_mgr_start_update() to reflect the change.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>