Initialize and power the 3D unit on Tegra20, Tegra30 and Tegra114 and
register a channel with the Tegra DRM driver so that the unit can be
used from userspace.
Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Signed-off-by: Thierry Reding <treding@nvidia.com>

Command stream submissions are the same across all devices that expose
a channel to userspace, so move the code into a generic function.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Instead of using magic numbers for the registers which contain memory
addresses in the firewall table, using symbolic names.
Signed-off-by: Thierry Reding <treding@nvidia.com>

A connector's DPMS mode isn't initialized by default, therefore using a
default of 0 (DRM_MODE_DPMS_ON). This can cause problems in that the DRM
core won't explicitly turn on a connector because it thinks that it is
already on.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The VDD regulator used to be enabled only at tegra_output_hdmi_enable,
which is called after a sink is detected. However, the HDMI hotplug pin
works by returning the voltage supplied by the VDD pin, so this meant
that the hotplug pin was never asserted and the sink was not detected
unless the VDD regulator was set to be always on.

This patch moves the enable to the tegra_hdmi_init() function to make
sure the regulator will get enabled and therefore ensure proper hotplug
detection.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>

These seem to show up when building for architectures other than ARM,
which I guess will never happen. The reason why the kbuild test bot ran
into these was a missing dependency which has hence been fixed. Still it
doesn't hurt to fix them anyway.
Reported-by: kbuild test bot <fengguang.wu@intel.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>

Use EDID data to determine whether the display supports HDMI or DVI
only. The HDMI output used to assume to be connected to HDMI displays,
but that broke support for DVI displays that don't understand the
interspersed audio/other data.

To be on the safe side, default to DVI if no EDID data is available.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
[treding@nvidia.com: move detection to separate function]
Signed-off-by: Thierry Reding <treding@nvidia.com>

Tegra114 TMDS configuration requires a new peak_current field and the
driver current override bit has changed position.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>

Use a structure to parameterize the code to handle differences between
the HDMI hardware on various SoC generations. This removes the need to
clutter the code with checks for individual compatible values.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Everything related to Tegra uses Tegra20 and Tegra30 instead of Tegra2
and Tegra3, respectively. Rename the TMDS arrays in the HDMI driver for
consistency.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Tegra114 uses a slightly updated version of host1x with an additional
syncpoint.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Since the .init() and .exit() functions are executed whenever the DRM
driver is loaded or unloaded, care must be taken not to use them for
resource allocation. Otherwise deferred probing cannot be used, since
the .init() and .exit() are not run at probe time. Similarly the code
that frees resources must be run at .remove() time. If it is run from
the .exit() function, it can release resources multiple times.

To handle this more consistently, rename the tegra_output_parse_dt()
function to tegra_output_probe() and introduce tegra_output_remove()
which can be used to free output-related resources.
Signed-off-by: Thierry Reding <treding@nvidia.com>

When the DRM driver is unloaded, all the associated resources must be
cleaned up and zeroed out. This is necessary because of the architecture
of the Tegra DRM driver, where not all subdrivers are unloaded along
with the DRM driver. Therefore device-managed managed won't be freed and
memory cannot be assumed to have been cleared (because it hasn't been
reallocated using kzalloc()) by the time the DRM driver is reloaded. It
is therefore necessary to zero out the structures to prevent strange
errors (such as slab corruptions) from occurring.
Signed-off-by: Thierry Reding <treding@nvidia.com>

In order to make subsystem-wide changes easier, move the Tegra DRM
driver back into the DRM tree.
Signed-off-by: Thierry Reding <treding@nvidia.com>

This is slightly safer than adding -Idrivers/gpu/host1x to cflags-y.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The Tegra DRM driver currently uses some infrastructure to defer the DRM
core initialization until all required devices have registered. The same
infrastructure can potentially be used by any other driver that requires
more than a single sub-device of the host1x module.

Make the infrastructure more generic and keep only the DRM specific code
in the DRM part of the driver. Eventually this will make it easy to move
the DRM driver part back to the DRM subsystem.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Expose the buffer objects, syncpoint and channel functionality in the
public public header so that drivers can use them.
Signed-off-by: Thierry Reding <treding@nvidia.com>

This structure derives from host1x_client. DRM-specific fields are moved
from host1x_client to this structure, so that host1x_client can remain
agnostic of DRM.
Signed-off-by: Thierry Reding <treding@nvidia.com>

In preparation to support host1x clients other than DRM, move this
header into a public location.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Rename the host1x_to_drm_bo() macro to host1x_to_tegra_bo() for
consistency and fixup various stylistic issues.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Other drivers use the tegra- prefix in their names, so add it to this
driver's name as well for consistency.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Rework the address table code for the host1x firewall. The previous
implementation allocated a bitfield but didn't check for a valid pointer
so it could potentially crash. Instead, embed a static bitmap within the
gr2d structure to avoid the allocation and use the Linux bitmap API to
reduce code complexity.

Don't annotate the driver's .remove() function __exit. Even if built in
the driver can be unloaded via sysfs, so .remove() needs to stick around
after initialization. Also remove the explicit initialization of the
driver's .owner field to THIS_MODULE because that's now handled by the
driver core.

Furthermore make an error message more consistent with other subdrivers,
index the syncpts array for better readability, remove a gratuituous
newline and reorder some variable declarations to make the code easier
to read.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The same code sequence is used in various places to validate a register
access in the command stream. This can be refactored into a separate
function.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The value stored in this field is a pointer to a command buffer, not an
ID. Avoid some confusion by reflecting that in the field's name.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Arguments on subsequent lines should be aligned with the first argument.
This one occurrence went unnoticed during code review.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The device can be unbound from the driver via sysfs, so regardless of
whether the driver is builtin or a module, its .remove() function needs
to stick around.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Most of the included files are either not required or already included
by some other header file.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The structure represents a context associated with a particular process
that has opened the Tegra DRM device and requested a channel. This is a
very DRM-specific notion and has nothing to do with host1x. Rename the
structure to more clearly mark the boundaries between the two.
Signed-off-by: Thierry Reding <treding@nvidia.com>

This structure extends drm_file with Tegra DRM specific fields and has
nothing to do with host1x. Rename the structure to more clearly mark the
boundaries between host1x and Tegra DRM.

While at it, move the structure definition out of the header. It's never
used outside of the drm.c source file, so it can be defined within that.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The host1x and Tegra DRM drivers are currently tightly coupled. Renaming
the structure marks the boundary more clearly.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Remove the unused host1x field from the structure and group the fields
more logically.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Some of the fields in struct host1x_drm haven't been used for a while,
so remove them.
Signed-off-by: Thierry Reding <treding@nvidia.com>

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

The num_relocs count are passed to the kernel per job, not per gather.

For multi-gather jobs, we would previously fail if there were relocs in
other gathers aside from the first one.

Fix this by simply moving the check until all gathers have been
consumed.
Signed-off-by: Erik Faye-Lund <kusmabite@gmail.com>
Reviewed-by: Arto Merilainen <amerilainen@nvidia.com>
Acked-By: Terje Bergstrom <tbergstrom@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>

Fix a typo (iotcl -> ioctl) in the debug message when an unknown IOCTL
is encountered.
Acked-by: David Airlie <airlied@linux.ie>
Signed-off-by: Thierry Reding <treding@nvidia.com>

When userspace removes the active framebuffer using DRM_IOCTL_MODE_RMFB,
or explicitly disables the CRTC (by calling drmModeSetCrtc(..., NULL)
for example), a NULL framebuffer will be passed to the .set_config()
implementation of a CRTC. The drm_crtc_helper_set_config() helper will
decide to disable a CRTC when that happens.

To do so, it calls drm_crtc_helper_disable(), which in turn will iterate
over all encoders and decouple them from their connectors and finally
call drm_helper_disable_unused_functions() to clean up and call the
.disable() or .dpms() implementation for each encoder. However, at no
point during this sequence does it track the DPMS mode of a connector,
so it will usually remain on after this.

When a connector is enabled again, drm_helper_connector_dpms() will not
notice that the DPMS mode actually changed and won't do anything, which
causes the connector to stay disabled indefinitely.

To prevent this from happening, explicitly set the connector's DPMS mode
to off when the CRTC is disabled. That way it reflects the correct state
and can be enabled again.

This solves an issue observed when terminating an X server running on
the xf86-video-modesetting driver. Without this patch, the connector
would not be enabled properly and the screen would stay dark.
Acked-by: David Airlie <airlied@linux.ie>
Signed-off-by: Thierry Reding <treding@nvidia.com>

This will make the next patch to change how this works a lot cleaner.
Signed-off-by: Dave Airlie <airlied@redhat.com>

New feature pile for 3.12! Highlights:
- Stereo/3d support for hdmi from Damien, both the drm core bits and
  the i915 integration.
- Manual boost/deboost logic for gpu turbo (Chris)
- Fixed up clock readout support for vlv (Chris).
- Tons of little fixes and improvements for vlv in general (Chon Minng
  Lee and Jesse Barnes).
- Power well support for the legacy vga plane (Ville).
- DP impromevents from Jani.
- Improvements to the Haswell modeset sequence (Ville+Paulo).
- Haswell DDI improvements, using the VBT for some tuning values and
  to check the configuration (Paulo).
- Tons of other small improvements and fixups.

* 'drm-intel-next' of git://people.freedesktop.org/~danvet/drm-intel: (92 commits)
  drm/i915: Use adjusted_mode in the fastboot hack to disable pfit
  drm/i915: Add a more detailed comment about the set_base() fastboot hack
  drm/i915/vlv: Turn off power gate for BIOS-less system.
  drm/i915/vlv: reset DPIO on load and resume v2
  drm/i915: Simplify PSR debugfs
  drm/i915: Tweak RPS thresholds to more aggressively downclock
  drm/i915: Boost RPS frequency for CPU stalls
  drm/i915: Fix __wait_seqno to use true infinite timeouts
  drm/i915: Add some missing steps to i915_driver_load error path
  drm/i915: Clean up the ring scaling calculations
  drm/i915: Don't populate pipe_src_{w,h} multiple times
  drm/i915: implement the Haswell mode set sequence workaround
  drm/i915: Disable/enable planes as the first/last thing during modeset on HSW
  i915/vlv: untangle integrated clock source handling v4
  drm/i915: fix typo s/PatherPoint/PantherPoint/
  drm/i915: Make intel_resume_power_well() static
  drm/i915: destroy connector sysfs files earlier
  drm/i915/dp: do not write DP_TRAINING_PATTERN_SET all the time
  drm/i915/dp: retry i2c-over-aux seven times on AUX DEFER
  drm/i915/vlv: reduce GT FIFO error info to a debug message
  ...

The conflict in intel_drv.h tripped me up a bit since a patch in dinq
moves all the functions around, but another one in drm-next removes a
single function. So I'ev figured backing this into a backmerge would
be good.

i915_dma.c is just adjacent lines changed, nothing nefarious there.

Conflicts:
	drivers/gpu/drm/i915/i915_dma.c
	drivers/gpu/drm/i915/intel_drv.h
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

It's not really needed, rather just adds another place to hold
intermediate values that could go wrong, and it's not clear that the
training pattern set or training lane set should be written at this
point at all.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>