All the display related blocks on Tegra require contiguous memory. Using
the DMA API, there is no knowing at import time which device will end up
using the buffer, so it's not known whether or not an IOMMU will be used
to map the buffer.

Move the check for non-contiguous buffers/mappings to the tegra_dc_pin()
function which is now the earliest point where it is known if a DMA BUF
can be used by the given device or not.

v2: add check for contiguous buffer/mapping in tegra_dc_pin()
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

Buffers that are imported from a DMA-BUF don't have pages allocated with
them. At the same time an SG table for them can't be derived using the
DMA API helpers because the necessary information doesn't exist. However
there's already an SG table that was created during import, so this can
simply be duplicated.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The Tegra DRM never actually took that lock, so the driver was broken
until generic locking was added in commit:

	commit b962a120
	Author: Rob Clark <robdclark@gmail.com>
	Date:   Mon Oct 22 14:31:22 2018 +0200

	    drm/atomic: integrate modeset lock with private objects

It's now no longer necessary to take that lock, so drop the check.

v2: add rationale for why it is now safe to drop the check (Daniel)
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Thierry Reding <treding@nvidia.com>

Currently configurations can be generated where IOMMU_SUPPORT is
disabled but IOMMU_IOVA is built as a module and DRM_TEGRA as built-in.
In such a case, the symbols guarded by IOMMU_IOVA will not be available
when linking the Tegra DRM driver and cause a linking failure.

Simplify this by unconditionally selecting IOMMU_IOVA, which makes sure
that it will be forced to =y if DRM_TEGRA=y. Technically we can now get
IOMMU_IOVA code built-in even if we don't use it (Tegra DRM only uses it
when IOMMU_SUPPORT is also enabled), but such configuration are of a
mostly academic nature. In all practical configurations we want IOMMU
support anyway.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Currently configurations can be generated where IOMMU_SUPPORT is
disabled but IOMMU_IOVA is built as a module and HOST1X as built-in. In
such a case, the symbols guarded by IOMMU_IOVA will not be available
when linking the host1x driver and cause a linking failure.

Simplify this by unconditionally selecting IOMMU_IOVA, which makes sure
that it will be forced to =y if HOST1X=y. Technically we can now get
IOMMU_IOVA code built-in even if we don't use it (host1x only uses it
when IOMMU_SUPPORT is also enabled), but such configuration are of a
mostly academic nature. In all practical configurations we want IOMMU
support anyway.
Signed-off-by: Thierry Reding <treding@nvidia.com>

If a client is already attached to an IOMMU domain that is not the
shared domain, don't try to attach it again. This allows using the
IOMMU-backed DMA API.

Since the IOMMU-backed DMA API is now supported and there's no way
to detach from it on 64-bit ARM, don't bother to detach from it on
32-bit ARM either.
Signed-off-by: Thierry Reding <treding@nvidia.com>

If a display controller is not attached to an explicit IOMMU domain,
which usually means that it's connected to an IOMMU domain controlled by
the DMA API, make sure to map the framebuffer to the display controller
address space. This allows us to transparently handle setups where the
display controller is attached to an IOMMU or setups where it isn't. It
also allows the driver to work with a DMA API that is backed by an
IOMMU.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Rename paddr -> iova and vaddr -> virt to make it clearer how these
addresses are used. This is important for a subsequent patch that makes
a distinction between the physical address (physical address of the
system memory from the CPU's point of view) and the IOVA (physical
address of the system memory from the device's point of view).
Signed-off-by: Thierry Reding <treding@nvidia.com>

Having to provide allocator hooks to the Falcon library is somewhat
cumbersome and it doesn't give the users of the library a lot of
flexibility to deal with allocations. Instead, remove the notion of
Falcon "operations" and let drivers deal with the memory allocations
themselves.
Signed-off-by: Thierry Reding <treding@nvidia.com>

If the Tegra DRM clients are backed by an IOMMU, push buffers are likely
to be allocated beyond the 32-bit boundary if sufficient system memory
is available. This is problematic on earlier generations of Tegra where
host1x supports a maximum of 32 address bits for the GATHER opcode. More
recent versions of Tegra (Tegra186 and later) have a wide variant of the
GATHER opcode, which allows addressing up to 64 bits of memory.

If host1x itself is behind an IOMMU as well this doesn't matter because
the IOMMU's input address space is restricted to 32 bits on generations
without support for wide GATHER opcodes.

However, if host1x is not behind an IOMMU, it won't be able to process
push buffers beyond the 32-bit boundary on Tegra generations that don't
support wide GATHER opcodes. Restrict the DMA mask to 32 bits on these
generations prevents buffers from being allocated from beyond the 32-bit
boundary.
Signed-off-by: Thierry Reding <treding@nvidia.com>

If host1x_bo_pin() returns an SG table, create a DMA mapping for the
buffer. For buffers that the host1x client has already mapped itself,
host1x_bo_pin() returns NULL and the existing DMA address is used.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Currently when the gather buffers are copied, they are copied to a
buffer that is allocated for the host1x client that wants to execute the
command streams in the buffers. However, the gather buffers will be read
by the host1x device, which causes SMMU faults if the DMA API is backed
by an IOMMU.

Fix this by allocating the gather buffer copy for the host1x device,
which makes sure that it will be mapped into the host1x's IOVA space if
the DMA API is backed by an IOMMU.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Add direction flags to host1x relocations performed during job pinning.
These flags indicate the kinds of accesses that hardware is allowed to
perform on the relocated buffers.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The debugfs files created for host1x are never removed, causing these
files to be left dangling in debugfs. This results in a crash when any
of these files are accessed after the host1x driver has been removed,
as well as a failure to create the debugfs entries when they are added
again on driver probe.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The host1x_bo_pin() and host1x_bo_unpin() APIs are used to pin and unpin
buffers during host1x job submission. Pinning currently returns the SG
table and the DMA address (an IOVA if an IOMMU is used or a physical
address if no IOMMU is used) of the buffer. The DMA address is only used
for buffers that are relocated, whereas the host1x driver will map
gather buffers into its own IOVA space so that they can be processed by
the CDMA engine.

This approach has a couple of issues. On one hand it's not very useful
to return a DMA address for the buffer if host1x doesn't need it. On the
other hand, returning the SG table of the buffer is suboptimal because a
single SG table cannot be shared for multiple mappings, because the DMA
address is stored within the SG table, and the DMA address may be
different for different devices.

Subsequent patches will move the host1x driver over to the DMA API which
doesn't work with a single shared SG table. Fix this by returning a new
SG table each time a buffer is pinned. This allows the buffer to be
referenced by multiple jobs for different engines.

Change the prototypes of host1x_bo_pin() and host1x_bo_unpin() to take a
struct device *, specifying the device for which the buffer should be
pinned. This is required in order to be able to properly construct the
SG table. While at it, make host1x_bo_pin() return the SG table because
that allows us to return an ERR_PTR()-encoded error code if we need to,
or return NULL to signal that we don't need the SG table to be remapped
and can simply use the DMA address as-is. At the same time, returning
the DMA address is made optional because in the example of command
buffers, host1x doesn't need to know the DMA address since it will have
to create its own mapping anyway.
Signed-off-by: Thierry Reding <treding@nvidia.com>

All the devices that make up the DRM device are now part of the same
IOMMU group. This simplifies the handling of the IOMMU attachment and
also avoids exhausting the number of IOMMUs available on early Tegra
SoC generations.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The ->load() and ->unload() drivers are midlayers and should be avoided
in modern drivers. Fix this by moving the code into the driver ->probe()
and ->remove() implementations, respectively.

v2: kick out conflicting framebuffers before initializing fbdev
v3: rebase onto drm/tegra/for-next
Tested-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>

In order to support different modes (DP in addition to HDMI), split out
the audio setup/teardown into callbacks.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The code to enable audio support is split into two parts, one being
generic for the SOR and another part that is specific whether the SOR is
in HDMI mode or in DP mode. Split out the common part in preparation for
reusing the code in DP mode.
Signed-off-by: Thierry Reding <treding@nvidia.com>

When the SOR is disabled in DP mode as part of an unplug event, do not
attempt to power the DP link down. Powering down the link requires the
DPAUX to transmit AUX messages which only works if there's a connected
sink.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The SOR0 on Tegra210 does, contrary to what was previously assumed, in
fact support DisplayPort. The difference between SOR0 and SOR1 is that
the latter supports audio and HDCP over DP, whereas the former doesn't.

The code for eDP and DP is now almost identical and the differences can
easily be parameterized based on the presence of a panel. There is no
need any longer to duplicate the code.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The correct I/O pad needs to be powered up before DP can be used. Make
sure the correct default is set for Tegra generations where the I/O pad
cannot be derived from the SOR instance.
Signed-off-by: Thierry Reding <treding@nvidia.com>

With the clocks modelled consistently across SoC generations, the clock
setup for eDP, HDMI and DP can now be unified.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Reuse parameters from earlier generations to support DisplayPort on
Tegra194.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The connector type detection code is duplicated in two places. Keeping
both places in sync is an extra maintenance burden that can be avoided
by comparing the connector type operations that are set upon the first
detection.
Signed-off-by: Thierry Reding <treding@nvidia.com>

So far the pad clock was only needed on the second SOR instance. The
clock does exist for all SOR instances, though, so make sure it is
always implemented. This prepares for further unification of the code
in subsequent patches.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The device tree bindings for the Tegra210 SOR don't require the
controller instance to be defined, since the instance can be derived
from the compatible string. The index is never used on Tegra210, so we
got away with it not getting set. However, subsequent patches will
change that, so make sure the proper index is used.
Signed-off-by: Thierry Reding <treding@nvidia.com>

It turns out that SOR1 is just another instance of the same block as the
SOR0, so there is no need to distinguish them.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Add support for regular DisplayPort on Tegra210 and Tegra186.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The SOR found on Tegra SoCs does not support all the rates potentially
advertised by eDP 1.4. Make sure that the rates that are not supported
are filtered out.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Rework eDP code to correspond more closely to what's documented. This
also improves the reliability of modesets.
Signed-off-by: Thierry Reding <treding@nvidia.com>

This is necessary for the output abstraction to retrieve a list of valid
modes from the EDID of a connected panel/monitor. This will be useful in
conjunction with DisplayPort support that will be added in a subsequent
patch, so that the driver can read EDID via the AUX channel.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Make use of the DP link training helpers to implement full and fast link
training. While at it, refactor some of the code and remove various code
sequences that are not necessary.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Add a helper that will perform link training as described in the
DisplayPort specification.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Parses additional link rates from DPCD if the sink supports eDP 1.4.
Signed-off-by: Thierry Reding <treding@nvidia.com>

This helper chooses an appropriate configuration, according to the
bitrate requirements of the video mode and the capabilities of the
DisplayPort sink.
Signed-off-by: Thierry Reding <treding@nvidia.com>

If the sink is eDP and supports the alternate scrambler reset, enable
it.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Make use of ANSI 8B/10B channel coding if the DisplayPort sink supports
it.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Store the AUX read interval from DPCD, so that it can be used to wait
for the durations given in the specification during link training.
Signed-off-by: Thierry Reding <treding@nvidia.com>

If the sink supports eDP, read the eDP revision from it's DPCD.
Signed-off-by: Thierry Reding <treding@nvidia.com>