Upon driver load, reset the VBLANK machinery to off to reflect the
hardware state. Since the ->reset() callback is called from the initial
drm_mode_config_reset() call, move the latter after the VBLANK machinery
initialization by drm_vblank_init().
Signed-off-by: NThierry Reding <treding@nvidia.com>

Support low-active hotplug detect signals by storing the GPIO flags
parsed from device tree.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Use this macro to reduce some of the boilerplate.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Signed-off-by: NThierry Reding <treding@nvidia.com>

Use a sized unsigned 32-bit data type (u32) to store register contents.
The DPAUX registers are 32 bits wide irrespective of the architecture's
data width.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Pages allocated from shmemfs don't end up being cleared and flushed on
ARMv7, so they must be flushed explicitly. Use the DMA mapping API for
that purpose, even though it's not used for anything else.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The IOMMU may not always be able to address 2 GiB of memory. On Tegra20,
the GART supports 32 MiB starting at 0x58000000. Also the aperture on
Tegra30 and later is in fact the full 4 GiB, rather than just 2 GiB as
currently assumed.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The DPAUX read/write FIFO registers aren't sequential in the register
space, causing transfers larger than 4 bytes to cause accesses to non-
existing FIFO registers.

Fixes: 6b6b6042 ("drm/tegra: Add eDP support")
Signed-off-by: NThierry Reding <treding@nvidia.com>

Tegra would not only need a hardware vblank counter that
increments at leading edge of vblank, but also support
for instantaneous high precision vblank timestamp queries, ie.
a proper implementation of dev->driver->get_vblank_timestamp().

Without these, there can be off-by-one errors during vblank
disable/enable if the scanout is inside vblank at en/disable
time, and additionally clients will never see any useable
vblank timestamps when querying via drmWaitVblank ioctl. This
would negatively affect swap scheduling under X11 and Wayland.
Signed-off-by: NMario Kleiner <mario.kleiner.de@gmail.com>
Acked-by: NThierry Reding <treding@nvidia.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

At present, dma_buf_export() takes a series of parameters, which
makes it difficult to add any new parameters for exporters, if required.

Make it simpler by moving all these parameters into a struct, and pass
the struct * as parameter to dma_buf_export().

While at it, unite dma_buf_export_named() with dma_buf_export(), and
change all callers accordingly.
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@canonical.com>
Reviewed-by: NDaniel Thompson <daniel.thompson@linaro.org>
Acked-by: NMauro Carvalho Chehab <mchehab@osg.samsung.com>
Acked-by: NDave Airlie <airlied@redhat.com>
Signed-off-by: NSumit Semwal <sumit.semwal@linaro.org>

As there isn't a way for the firmware on the Nyan Chromebooks to hand
over the display to the kernel, and the kernel isn't redoing the whole
configuration at present.

With this patch, the SOR is brought to a known state and we get correct
display on every boot.
Signed-off-by: NTomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>

Name the fields of the SOR_SEQ_CTL register consistently.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Resets on Tegra are synchronous, so keep the clock enabled while
asserting the reset.
Signed-off-by: NThierry Reding <treding@nvidia.com>

This function is called by output drivers so should be documented. While
at it, move it to a more appropriate location.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The ->mode_set() and ->mode_set_base() callbacks are no longer used with
full atomic mode-setting drivers, so remove them.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The tegra_dc_setup_clock() function is unused after the conversion to
atomic mode-setting, so remove it.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Instead of duplicating the code, make use of the newly introduced atomic
state duplicate and destroy helpers. This allows changes to the base
atomic state handling to automatically propagate to the Tegra driver and
thereby prevent breakage resulting from both copies going out of sync.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The display controller on Tegra can use syncpoints to count VBLANK
events. syncpoints are 32-bit unsigned integers, so well suited as
VBLANK counters.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Signed-off-by: NThierry Reding <treding@nvidia.com>

Use a sized unsigned 32-bit data type (u32) to store register contents.
The SOR registers are 32 bits wide irrespective of the architecture's
data width.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The Tegra DRM driver uses a single IO virtual address space for buffer
mappings. Provide a table of the address space usage in debugfs.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Check that the desired parent clock is indeed a valid parent for the
display controller clock. This is purely cosmetic at this point since
the parent clocks are specified in DT and all the currently defined
parents are in fact valid parents of the display controller clock.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Use cases like rotation require these hooks to have some context so they
know how to prepare and cleanup the frame buffer correctly.

For i915 specifically, object backing pages need to be mapped differently
for different rotation modes and the driver needs to know which mapping to
instantiate and which to tear down when transitioning between them.

v2: Made passed in states const. (Daniel Vetter)

[airlied: add mdp5 and atmel fixups]
Signed-off-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: NRob Clark <robdclark@gmail.com>
Signed-off-by: NDave Airlie <airlied@redhat.com>

These names only make sense because of backwards compatability with
the order used by the crtc helper library. There's not really any real
requirement in the ordering here.

So rename them to something more descriptive and update the kerneldoc
a bit. Motivated in a discussion with Laurent about how to restore
plane state for dpms for drivers with runtime pm.

v2: Squash in fixup from Stephen Rothwell to fix a conflict with
tegra.

Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Reviewed-by: NRob Clark <robdclark@gmail.com>
Acked-by: NLaurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

The code in tegra_crtc_prepare() really belongs in tegra_dc_init(), or
at least most of it. This fixes an issue with VBLANK handling because
tegra_crtc_prepare() would overwrite the interrupt mask register that
tegra_crtc_enable_vblank() had written to to enable VBLANK interrupts.
Tested-by: NTomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>

Store a pointer to the CRTC in its atomic state to make it easy for
state handling code to get at the CRTC.
Tested-by: NTomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>

Commit eab3bbef ("drm/atomic: Add drm_crtc_state->active") added the
field to track the DPMS state. However, the Tegra driver was in modified
in parallel and subclasses the CRTC atomic state, so needed to duplicate
the code in the atomic helpers. After the addition of the active_changed
field it became out of sync and doesn't reset it when duplicating state.

This causes a full modeset on things like page-flips, which will in turn
cause warnings due to the VBLANK machinery being disabled when it really
should remain on.
Tested-by: NTomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>

Recent changes in the clock framework have caused a behavioural change
in that clocks that have not had their rate set explicitly will now be
reset to their initial rate (or 0) when the clock is released. This is
triggered in the deferred probing path, resulting in the clock running
at a wrong frequency after the successful probe.

This can be easily fixed by setting the rate explicitly rather than by
relying on the implicit rate inherited by the parent.
Tested-by: NTomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>

When copying a relocation from userspace, copy the correct target
offset.
Signed-off-by: NDavid Ung <davidu@nvidia.com>
Fixes: 961e3bea ("drm/tegra: Make job submission 64-bit safe")
Cc: stable@vger.kernel.org
[treding@nvidia.com: provide a better commit message]
Signed-off-by: NThierry Reding <treding@nvidia.com>

For now only disable the KMS hotplug polling helper logic upon suspend
and re-enable it on resume.
Reviewed-by: NSean Paul <seanpaul@chromium.org>
Reviewed-by: NMark Zhang <markz@nvidia.com>
Signed-off-by: NThierry Reding <treding@nvidia.com>

Previously output drivers would enable continuous display mode and power
up the display controller at various points during the initialization.
This is suboptimal because it accesses display controller registers in
output drivers and duplicates a bit of code.

Move this code into the display controller driver and enable the display
controller as the final step of the ->mode_set_nofb() implementation.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Tracking these in the plane state allows them to be computed in the
->atomic_check() callback and reused when applying the configuration in
->atomic_update().
Signed-off-by: NThierry Reding <treding@nvidia.com>

Wrap struct drm_crtc_state in a driver-specific structure and add the
planes field which keeps track of which planes are updated or disabled
during a modeset. This allows atomic updates of the the display engine
at ->atomic_flush() time.

v2: open-code getting the state of the CRTC that the plane is being
    attached to (Daniel Vetter)
Signed-off-by: NThierry Reding <treding@nvidia.com>

All output drivers have now been converted to use the ->atomic_check()
callback, so the ->mode_fixup() callbacks are no longer used.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Provide a custom ->atomic_commit() implementation which supports async
commits. The generic atomic page-flip helper can use this to implement
page-flipping.
Signed-off-by: NThierry Reding <treding@nvidia.com>

Replace drm_crtc_helper_set_config() by drm_atomic_helper_set_config().
All drivers have now been converted to use ->atomic_check() to set the
atomic state, therefore the atomic mode setting helpers can be used.
Signed-off-by: NThierry Reding <treding@nvidia.com>

All clock state is now stored in the display controller's atomic state,
so the output drivers no longer need to call back into the display
controller driver to set up the clock. This is also required to make
sure no hardware changes are made before validating a configuration.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The implementation of the ->atomic_check() callback precomputes all
parameters to check if the given configuration can be applied. If so the
precomputed values are stored in the atomic state object for the encoder
and applied during modeset. In that way the modeset no longer needs to
perform any checking but simply program values into registers.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The implementation of the ->atomic_check() callback precomputes all
parameters to check if the given configuration can be applied. If so the
precomputed values are stored in the atomic state object for the encoder
and applied during modeset. In that way the modeset no longer needs to
perform any checking but simply program values into registers.
Signed-off-by: NThierry Reding <treding@nvidia.com>

The implementation of the ->atomic_check() callback precomputes all
parameters to check if the given configuration can be applied. If so the
precomputed values are stored in the atomic state object for the encoder
and applied during modeset. In that way the modeset no longer needs to
perform any checking but simply program values into registers.
Signed-off-by: NThierry Reding <treding@nvidia.com>