Sync objects are new toplevel drm object, that contain a
pointer to a fence. This fence can be updated via command
submission ioctls via drivers.

There is also a generic wait obj API modelled on the vulkan
wait API (with code modelled on some amdgpu code).

These objects can be converted to an opaque fd that can be
passes between processes.

v2: rename reference/unreference to put/get (Chris)
fix leaked reference (David Zhou)
drop mutex in favour of cmpxchg (Chris)
v3: cleanups from danvet, rebase on drm_fops rename
check fd_flags is 0 in ioctls.
v4: export find/free, change replace fence to take a
syncobj. In order to support lookup first, replace
later semantics which seem in the end to be cleaner.
Reviewed-by: Sean Paul <seanpaul@chromium.org>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>

Merge tag 'exynos-drm-next-for-v4.13' of git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos into drm-next

Summary:
- Rework vblank handling
  . This patch series adds frame counter callback and removes
    unnecessary pipe relevnt fields and simplifies event handling.
- clean up and fix up sw-trigger relevant code
  . This patch series moves TE relevant code from Panel and HDMI
    to DECON driver to fix a race between interrupt handlers and
    DECON disable, and to fix timeout issue at wait-for-vblank.
  . It removes unnecessary flags and check code specific to Exynos driver.

* tag 'exynos-drm-next-for-v4.13' of git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos: (27 commits)
  drm/exynos/decon5433: remove useless check
  drm/exynos/decon5433: kill BIT_SUSPENDED flag
  drm/exynos/decon5433: kill BIT_WIN_UPDATED flag
  drm/exynos/decon5433: kill BIT_CLKS_ENABLED flag
  drm/exynos/decon5433: kill BIT_IRQS_ENABLED flag
  drm/exynos/decon5433: move TE handling to DECON
  dt-bindings: exynos5433-decon: add TE interrupt binding
  dt-bindings: exynos5433-decon: fix interrupts bindings
  drm/exynos/decon5433: always do sw-trigger when vblanks enabled
  drm/exynos: mixer: document YCbCr magic numbers
  drm/exynos: mixer: simplify mixer_cfg_rgb_fmt()
  drm/exynos/dsi: fix bridge_node DT parsing
  drm/exynos/hdmi: fix pipeline disable order
  drm/exynos/decon5433: simplify shadow protect code
  drm/exynos/decon5433: kill BIT_IRQS_ENABLED
  drm/exynos/decon5433: kill DECON_UPDATE workaround
  drm/exynos: kill mode_set_nofb callback
  drm/exynos: kill pipe field from drivers contexts
  drm/exynos: set plane possible_crtcs in exynos_plane_init
  drm/exynos: kill exynos_drm_private::pipe
  ...

Core Changes:
- Stop proliferation of drm_vblank_cleanup by adding to the docs and deleting
  boilerplate (Daniel)
- Roll out and use mode_valid hooks across crtc/encoder/bridge (Jose)
- Add drm_vblank.[hc] to isolate vblank code from optional irq helpers (Daniel)

Driver Changes:
- Replace drm_for_each_connector with drm_for_each_connector_iter (Gustavo)
- A couple misc driver fixes

Cc: Gustavo Padovan <gustavo.padovan@collabora.com>
Cc: Jose Abreu <Jose.Abreu@synopsys.com>
Cc: Daniel Vetter <daniel.vetter@intel.com>

* tag 'drm-misc-next-2017-06-02' of git://anongit.freedesktop.org/git/drm-misc: (34 commits)
  drm/vc4: Mark the device as active when enabling runtime PM.
  drm: remove writeq/readq function definitions
  drm/atmel-hlcdc: Use crtc->mode_valid() callback
  drm/exynos: Drop drm_vblank_cleanup
  drm/hdlcd|mali: Drop drm_vblank_cleanup
  drm/doc: Polish irq helper documentation
  drm: Extract drm_vblank.[hc]
  drm/vc4: Fix comment in vc4_drv.h
  drm/pl111: fix warnings without CONFIG_ARM_AMBA
  drm/atomic: Consitfy mode parameter to drm_atomic_set_mode_for_crtc()
  drm/arcgpu: Drop drm_vblank_cleanup
  drm/atmel: Drop drm_vblank_cleanup
  drm/imx: Drop drm_vblank_cleanup
  drm/meson: Drop drm_vblank_cleanup
  drm/stm: Drop drm_vblank_cleanup
  drm/sun4i: Drop drm_vblank_cleanup
  drm: better document how to send out the crtc disable event
  drm: Use vsnprintf extension %ph
  drm/doc: move printf helpers out of drmP.h
  drm/pl111: select DRM_PANEL
  ...

Merge tag 'omapdrm-4.13-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba/linux into drm-next

omapdrm changes for v4.13

* dmabuf fence support
* TILER rotation fixes
* big pile of various cleanups and refactorings

* tag 'omapdrm-4.13-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/tomba/linux: (64 commits)
  drm/omap: fix tiled buffer stride calculations
  drm/omap: fix YUV422 90/270 rotation with mirroring
  drm/omap: fix YUV422 rotation with TILER
  drm/omap: pass rotation to dispc
  drm/omap: DRM_MODE_REFLECT_* instead of mirror boolean
  drm/omap: use DRM_MODE_ROTATE_* instead of OMAP_DSS_ROT_*
  drm/omap: remove omap_drm_win
  drm/omap: add drm_rotation_to_tiler helper()
  drm/omap: rename color_mode to fourcc
  drm/omap: cleanup formats array
  drm/omap: remove omap_framebuffer_get_formats()
  drm/omap: use DRM_FORMAT_* instead of OMAP_DSS_COLOR_*
  drm/omap: use u32 instead of enum omap_color_mode
  drm/omap: change supported_modes to an array
  drm/omap: remove unneeded prototypes
  drm/omap: add format_is_yuv() helper
  drm/omap: cleanup offset calculation
  drm/omap: remove dma & vrfb rotation
  drm/omap: ratelimit OCP error
  drm/omap: remove CLUT
  ...

Failing to do so meant that we got a resume() callback on first use of
the device, so we would leak the bin BO that we allocated during
probe.
Signed-off-by: Eric Anholt <eric@anholt.net>
Fixes: 553c942f ("drm/vc4: Allow using more than 256MB of CMA memory.")
Link: http://patchwork.freedesktop.org/patch/msgid/20170515171615.10168-1-eric@anholt.netReviewed-by: Boris Brezillon <boris.brezillon@free-electrons.com>

Instead of rewriting write/readq, use linux/io-64-nonatomic-lo-hi.h which
already have them.
Signed-off-by: Corentin Labbe <clabbe.montjoie@gmail.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: http://patchwork.freedesktop.org/patch/msgid/20170602112510.17544-1-clabbe.montjoie@gmail.com

Now that we have a callback to check if crtc supports a given mode
we can use it in atmel-hlcdc so that we restrict the number of probbed
modes to the ones we can actually display.

Also, remove the mode_fixup() callback as this is no longer needed
because mode_valid() will be called before.
Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Link: http://patchwork.freedesktop.org/patch/msgid/95fd6c06c58bd0b957e36a8d7068e6a74b581304.1495720737.git.joabreu@synopsys.com

omap_gem uses page alignment for buffer stride. The related calculations
are a bit off, though, as byte stride of 4096 gets aligned to 8192,
instead of 4096.

This patch changes the code to use DIV_ROUND_UP(), which fixes those
calculations and makes them more readable.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

When rotating 90/270 + mirroring with YUV422, the end result will have
adjacent pixels swapped. The problem is that
dispc_ovl_set_rotation_attrs() has wrong rotation values for these
cases.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

TILER rotation with YUV422 pixelformats does not work at the moment. All
other pixel formats work, because the pixelformat's pixel size is equal
to tiler unit size (e.g. XR24's pixel size is 32 bits, and the TILER
unit size that has to be used is 32 bits).

For YUV422 formats this is not the case, as the TILER unit size has to
be 32 bits, but the pixel size is 16 bits. The end result is OCP errors
and sync losts.

This patch adds the code to adjust the variables for YUV422 formats.

We could make the code more generic by passing around the pixel format,
rotation type, angle and the tiler unit size, which would allow us to do
calculations without special case for YUV422. However, this would make
the code more complex, and at least for now this is much more easier to
handle with these two special cases for YUV422.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

The omapdrm driver has not passed the rotation value to the dispc
driver. This doesn't affect RGB formats, but YUV formats don't work
without dispc knowing the orientation.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Change dispc driver to use the DRM_MODE_REFLECT flags instead of a mirror
boolean.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

At the moment the dispc driver uses a custom enum for rotation. Change
it to use the DRM's DRM_MODE_ROTATE_*.

Note that mirroring is at the moment handled as a separate boolean in
the dispc driver, so we only use the DRM_MODE_ROTATE_* values.

Note, DSS HW uses clockwise rotation, DRM counter-clockwise.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

struct omap_drm_window is only used to pass plane setup data to
omap_framebuffer_update_scanout(). This can as well be accomplished by
just passing the DRM state.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Add a helper function to convert DRM rotation to TILER rotation.

Also drop a error print that can never happen, as the DRM framework
makes sure the rotation is valid.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Now that we use fourccs, we can also rename the 'color_mode' variables
to 'fourcc'.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

omap_fb.c has a table with DSS and DRM formats, used to convert between
them. This is no longer needed, so we can change the array to a plain
array of DRM_FORMAT_* values which contain all possible pixel formats
supported by any DSS IP version.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

We now get a fourcc array from dispc when asking for a plane's supported
pixel formats, so we can drop omap_framebuffer_get_formats() which was
used to convert between DSS and DRM pixel formats.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

This patch changes omapdrm to use DRM_FORMAT_* values instead of
OMAP_DSS_COLOR_* enum.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

In this step we drop 'enum omap_color_mode', and use u32 instead.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

enum omap_color_mode is a bitmask, so at the moment we present the
supported color modes as mask. To be able to move to fourccs, we need to
use an array to present the supported color modes.

As a first step towards fourccs, this patch changes the code to use an
array to store the enums.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

omapdss.h contains prototypes for three functions, which are also
defined in dss_features.h. Remove the extra prototypes from omapdss.h.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

In a few places the dispc driver needs to know whether the pixel format
is an YUV format. Add a helper to figure that out.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

The code to calculate offset in dispc's calc_offset() is overly complex.
Simplify it.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

We have three rotation methods supported by the SoCs with DSS: DMA,
VRFB and TILER.

DMA rotation works in theory on all DSS platforms, but in practice it's
unusable due to the huge amount of memory bandwidth it uses, and has
never really been used.

VRFB is available on OMAP3, but is not supported by omapdrm, even though
we have some code for it in the dispc driver.

TILER is supported on OMAP4/OMAP5/DRA7/AM5 platforms, but has some
driver bugs.

To clean up the driver to help fixing the TILER issues, this patch drops
the DMA and VRFB rotation support, leaving only TILER rotation.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Use dev_err_ratelimited() when an OCP error happens, to slightly easen
the flood.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

DSS IP versions 2 and 3 support CLUT modes (color lookup table), but the
driver has never supported those. We still have had some code for CLUT
modes. As the newer DSS IP versions have dropped CLUT support, we might
as well clean up the driver by removing the CLUT related code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

The code that sets and clears DOUBLESTRIDE is only ran when using NV12.
This is not correct, as we might first set the bith when using NV12, but
never clear it when using other formats.

Fix it so that when the bit is available (when the HW supports NV12) we
always either set or clear the bit.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

The omap_dss_find_output() function is unused, remove it.
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>

Structs omap_overlay and omap_overlay_manager are not used, remove them.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Remove unused fields in struct omap_dss_device.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

get_resolution is not used, remove it.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

get_dimensions function pointer is not used, remove it.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

The RFBI driver has not worked nor compiled for many years. There are
very few boards out there that use RFBI, and no one has stepped up to
fix it.

So let's remove the RFBI code that doesn't even compile.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Remove DPI's legacy data_lines code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

recommended_bpp hasn't been used for ages, remove the code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Non-DT booting is no longer supported, so remove legacy code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Non-DT booting is no longer supported, so remove legacy code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Remove extra includes related to non-DT boot.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>

Non-DT booting is no longer supported, so remove legacy code.
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>