drm/i915: Extract per-platform plane->check() functions

Split up intel_check_primary_plane() and intel_check_sprite_plane()
into per-platform variants. This way we can get a unified behaviour
between the SKL universal planes, and we stop checking for non-SKL
specific scaling limits for the "sprite" planes. And we now get
a natural place where to add more plarform specific checks.

v2: Split the .check_plane() calling convention change out (José)

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907152413.15761-10-ville.syrjala@linux.intel.com

drm/i915: Nuke plane->can_scale/min_downscale

We can easily calculate the plane can_scale/min_downscale on demand.
And later on we'll probably want to start calculating these dynamically
based on the cdclk just as skl already does.

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907152413.15761-9-ville.syrjala@linux.intel.com

drm/i915: s/int plane/int color_plane/

To reduce the confusion between a drm plane and the planes of
framebuffers let's desiginate the latter as "color plane".

Weak-Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907152413.15761-8-ville.syrjala@linux.intel.com

drm/i915: Store ggtt_view in plane_state

Stash the gtt_view structure into the plane state. This will become
useful when we do GTT remapping as the gtt_view will not come directly
from the fb anymore.

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907152413.15761-7-ville.syrjala@linux.intel.com

drm/i915: Store the final plane stride in plane_state

Let's store the final plane stride in the plane state. This avoids
having to pick between the normal vs. rotated stride during hardware
programming. And once we get GTT remapping the plane stride will
no longer match the fb stride so we'll need a place to store it
anyway.

v2: Keep checking fb->pitches[0] for cursor as later on we won't
    populate plane_state->color_plane[0].stride for invisible planes
    and we have been checking the cursor fb stride even for invisible
    planes
v3: s/betwen/between in commit msg (José)
v4: Check color_plane[0].stride instead of fb->pitches[0] in
    the skl_check_main_surface() X-tiling kludge

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180911150139.23922-1-ville.syrjala@linux.intel.com

drm/i915: Rename the plane_state->main/aux to plane_state->color_plane[]

Make the main/aux surface stuff a bit more generic by using an array
of structures. This will allow us to deal with both the main and aux
surfaces with common code.

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907152413.15761-5-ville.syrjala@linux.intel.com

drm/i915: Use pipe A primary plane .max_stride() as the global stride limit

Let's assume that the primary plane for pipe A has the highest max
stride of all planes, and we'll use that as the global limit when
creating a new framebuffer.

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907152413.15761-4-ville.syrjala@linux.intel.com

drm/i915: Add .max_stride() plane hook

Each plane may have different stride limitations. Let's add a new
plane function to retutn the maximum stride for each plane. There's
going to be some use for this outside the .atomic_check() stuff hence
the separate hook.

v2: Fix ilk+ x-tiled max stride to be 32k (José)

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907152413.15761-3-ville.syrjala@linux.intel.com

drm/i915: s/tile_offset/aligned_offset/ etc.

Rename some of the tile_offset() functions to aligned_offset() since
they operate on both linear and tiled functions. And we'll include
_plane_ in the name of all the variants that take a plane state.
Should make it more clear which function to use where.

v2: Pimp the patch subject a bit (José)

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907152413.15761-2-ville.syrjala@linux.intel.com

drm/i915: Reorder execobject[] to insert non-48b objects into the low 4G

If the caller supplies more than 4G of objects and than one that has to
be in the low 4G, it is possible for the low 4G to be full before we
attempt to find room for the last object that must be there. As we don't
reorder the two types, every pass hits the same problem and we fail with
ENOSPC. However, if we impose a little bit of ordering between the two
classes of objects, on the second pass we will be able to fit the
special object as we do it first. For setups that only use !48b objects,
we now reverse the order between passes, hopefully making the subsequent
passes more likely to succeed given that we are trying a different
order (rather than repeating the previous pass!)

v2: Quick one line explanation for the relative priorities given to
reservations.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180912101133.31377-1-chris@chris-wilson.co.uk

drm/i915/ringbuffer: Reload PDs harder on byt/bcs

Baytrail takes a little more convincing that it needs to actually reload
its Page Directoy (ppGTT) before the context switch, so repeat it until
it gets the message. Once again the arbitrary values here are
empirically derived.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107861
Testcase: igt/gem_exec_parallel/fds
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180910130808.10809-1-chris@chris-wilson.co.uk

drm/i915: Nuke struct_mutex from context_setparam

Userspace should be free to race against itself and shoot itself in
the foot if it so desires to adjust a parameter at the same time as
submitting a batch to that context. As such, the struct_mutex in context
setparam is only being used to serialise userspace against itself and
not for any protection of internal structs and so is superfluous.

v2: Separate user_flags from internal flags to reduce chance of
interference; and use locked bit ops for user updates.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180911132206.23032-1-chris@chris-wilson.co.uk

drm/i915/icl: Define T_INIT_MASTER registers

This patch defines DSI_T_INIT_MASTER register for DSI ports
0/1 which will be used in dphy programming.

Signed-off-by: Madhav Chauhan <madhav.chauhan@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1531215614-6828-5-git-send-email-madhav.chauhan@intel.com

drm/i915/overlay: Use the ioctl parameters directly

The user parameters to put_image are not copied back to userspace
(DRM_IOW), and so we can modify the ioctl parameters (having already been
copied to a temporary kernel struct) directly and use those in place,
avoiding another temporary malloc and lots of manual copying.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180906190144.1272-2-chris@chris-wilson.co.uk

drm/i915/overlay: Allocate physical registers from stolen

Given that we are now reasonably confident in our ability to detect and
reserve the stolen memory (physical memory reserved for graphics by the
BIOS) for ourselves on most machines, we can put it to use. In this
case, we need a page to hold the overlay registers.

On an i915g running MythTv, H Buus noticed that

commit 6a2c4232ece145d8b5a8f95f767bd6d0d2d2f2bb
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Nov 4 04:51:40 2014 -0800
drm/i915: Make the physical object coherent with GTT

introduced stuttering into his video playback. After discarding the
likely suspect of it being the physical cursor updates, we were left
with the use of the phys object for the overlay. And lo, if we
completely avoid using the phys object (allocated just once on module
load!) by switching to stolen memory, the stuttering goes away.

For lack of a better explanation, claim victory and kill two birds with
one stone.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107600
Fixes: 6a2c4232ece1 ("drm/i915: Make the physical object coherent with GTT")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180906190144.1272-1-chris@chris-wilson.co.uk

drm/i915/chv: Update csc coefficient matrix during modeset

During modeset, previously configured csc coefficient matrix,if any, will
not persist. This can result in blank screen as csc mode will be programmed
while loading LUT but csc coefficient matrix remains unprogrammed.

Changes since V1:
- Removed platform check

Signed-off-by: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1536589634-29680-1-git-send-email-raviraj.p.sitaram@intel.com

drm/i915/guc: Update GuC power domain states

We should update GuC power domain states also when GuC submission
is disabled, otherwise GuC might complain or ignore our requests.
This seems to be required for all currently released GuC firmwares.

v2: it is only needed by pre-Gen11 firmwares

Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: John Spotswood <john.a.spotswood@intel.com>
Cc: Anusha Srivatsa <anusha.srivatsa@intel.com>
Cc: Tomasz Lis <tomasz.lis@intel.com>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Anusha Srivatsa <anusha.srivatsa@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180910104150.101752-1-michal.wajdeczko@intel.com

drm/i915: Missed interrupt simulation is no more, tell the world

Using the guc, we cannot disable the user interrupt generation as we use
it for driving submission. And from Icelake, we no longer have the
ability to individually mask interrupt generation from each engine,
disabling our ability to fake missed interrupts.

In both cases, report back to userspace that the missed interrupt
generator is no longer available.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180907112856.28242-1-chris@chris-wilson.co.uk

drm/i915: Update DRIVER_DATE to 20180906

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

Merge tag 'gvt-next-2018-09-04' of https://github.com/intel/gvt-linux into drm-intel-next-queued

gvt-next-2018-09-04

- guest context shadow optimization for restore inhibit one (Yan)
- cmd parser optimization (Yan)
- W=1 warning fixes (Zhenyu)

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
# Conflicts:
# drivers/gpu/drm/i915/gvt/reg.h
From: Zhenyu Wang <zhenyuw@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180904030154.GG20737@zhen-hp.sh.intel.com

drm/i915: Update DRIVER_DATE to 20180906

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Update DRIVER_DATE to 20180906

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Move final cleanup of drm_i915_private to i915_driver_destroy

Introduce a complementary function to i915_driver_create() to undo all
that is created.

Suggested-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Reviewed-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180905140921.17467-2-chris@chris-wilson.co.uk

drm/i915: Attach the pci match data to the device upon creation

Attach our device_info to the our i915 private on creation so that it is
always available for inspection.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180905140921.17467-1-chris@chris-wilson.co.uk

drm/i915/bdw: Increase IPS disable timeout to 100ms

During IPS disabling the current 42ms timeout value leads to occasional
timeouts, increase it to 100ms which seems to get rid of the problem.

References: https://bugs.freedesktop.org/show_bug.cgi?id=107494
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107562
Reported-by: Diego Viola <diego.viola@gmail.com>
Tested-by: Diego Viola <diego.viola@gmail.com>
Cc: Diego Viola <diego.viola@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180905100005.7663-1-imre.deak@intel.com

drm/i915: Reduce context HW ID lifetime

Future gen reduce the number of bits we will have available to
differentiate between contexts, so reduce the lifetime of the ID
assignment from that of the context to its current active cycle (i.e.
only while it is pinned for use by the HW, will it have a constant ID).
This means that instead of a max of 2k allocated contexts (worst case
before fun with bit twiddling), we instead have a limit of 2k in flight
contexts (minus a few that have been pinned by the kernel or by perf).

To reduce the number of contexts id we require, we allocate a context id
on first and mark it as pinned for as long as the GEM context itself is,
that is we keep it pinned it while active on each engine. If we exhaust
our context id space, then we try to reclaim an id from an idle context.
In the extreme case where all context ids are pinned by active contexts,
we force the system to idle in order to recover ids.

We cannot reduce the scope of an HW-ID to an engine (allowing the same
gem_context to have different ids on each engine) as in the future we
will need to preassign an id before we know which engine the
context is being executed on.

v2: Improved commentary (Tvrtko) [I tried at least]

References: https://bugs.freedesktop.org/show_bug.cgi?id=107788
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180904153117.3907-1-chris@chris-wilson.co.uk

drm/i915: Be defensive and don't assume PSR has any commit to sync against

If the previous modeset commit has completed and is no longer part of
the crtc state, skip waiting for it.

Ville pointed out that, in fact, the commit is never removed after a
modeset so the only way we could see a NULL here should be if there was
never a commit attached. Nevertheless, we have the evidence it can be
NULL and it has been defended against elsewhere, for example commit
93313538c153 ("drm/i915: Pass idle crtc_state to intel_dp_sink_crc").

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107792
Fixes: c44301fce614 ("drm/i915: Allow control of PSR at runtime through debugfs, v6")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180904162902.2578-1-chris@chris-wilson.co.uk

drm/i915: Pull intel_uncore_arm_unclaimed_mmio_detection() under the spinlock

Elsewhere we manipulate uncore.unclaimed_mmio_check and
i915_param.mmio_debug under the irq lock (e.g. preserving the current
value across a user forcewake grab), but do not protect the manipulation
inside intel_uncore_arm_unclaimed_mmio_detection() from concurrent
access, even from itself. This is an issue as we do call
arm_unclaimed_mmio_detection from multiple threads without coordination.

Suggested-by: Mika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intelcom>
Link: https://patchwork.freedesktop.org/patch/msgid/20180904131207.17563-1-chris@chris-wilson.co.uk

drm/i915/icl: Fix context RPCS programming

There are two issues with the current RPCS programming for Icelake:

Expansion of the slice count bitfield has been missed, as well as the
required programming workaround for the subslice count bitfield size
limitation.

1)

Bitfield width for configuring the active slice count has grown so we need
to program the GEN8_R_PWR_CLK_STATE accordingly.

Current code was always requesting eight times the number of slices (due
writing to a bitfield starting three bits higher than it should). These
requests were luckily a) capped by the hardware to the available number of
slices, and b) we haven't yet exported the code to ask for reduced slice
configurations.

Due both of the above there was no impact from this incorrect programming
but we should still fix it.

2)

Due subslice count bitfield being only three bits wide and furthermore
capped to a maximum documented value of four, special programming
workaround is needed to enable more than four subslices.

With this programming driver has to consider the GT configuration as
2x4x8, while the hardware internally translates this to 1x8x8.

A limitation stemming from this is that either a subslice count between
one and four can be selected, or a subslice count equaling the total
number of subslices in all selected slices. In other words, odd subslice
counts greater than four are impossible, as are odd subslice counts
greater than a single slice subslice count.

This also had no impact in the current code base due breakage from 1)
always reqesting more than one slice.

While fixing this we also add some asserts to flag up any future bitfield
overflows.

v2:
 * Use a local in all branches for clarity. (Lionel)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Bspec: 12247
Reported-by: tony.ye@intel.com
Suggested-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: tony.ye@intel.com
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903113007.2643-1-tvrtko.ursulin@linux.intel.com

drm/i915/ringbuffer: Move double invalidate to after pd flush

Continuing the fun of trying to find exactly the delay that is
sufficient to ensure that the page directory is fully loaded between
context switches, move the extra flush added in commit 70b73f9ac113
("drm/i915/ringbuffer: Delay after invalidating gen6+ xcs") to just
after we flush the pd. Entirely based on the empirical data of running
failing tests in a loop until we survive a day (before the mtbf is 10-30
minutes).

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107769
References: 70b73f9ac113 ("drm/i915/ringbuffer: Delay after invalidating gen6+ xcs")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180904063802.13880-1-chris@chris-wilson.co.uk

drm/i915: Double check we didn't miss an unclaimed register access

Currently, if the user has enabled mmio-debug around each register
access, we presume that we have then checked them all. However, it is
still possible through omission (raw register access) or external
interaction that the unclaimed access was not highlighted.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180904111732.24266-1-chris@chris-wilson.co.uk

drm/i915: Fix ICL+ HDMI clock readout

Copy the 38.4 vs. 19.2 MHz ref clock exception from the dpll
mgr into the clock readout function as well.

v2: Refactor the code into a common function
    s/is_icl/gen11+/ (Rodrigo)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107722
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903142841.14627-1-ville.syrjala@linux.intel.com
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

drm/i915: Use a cached mapping for the physical HWS

Older gen use a physical address for the hardware status page, for which
we use cache-coherent writes. As the writes are into the cpu cache, we use
a normal WB mapped page to read the HWS, used for our seqno tracking.

Anecdotally, I observed lost breadcrumbs writes into the HWS on i965gm,
which so far have not reoccurred with this patch. How reliable that
evidence is remains to be seen.

v2: Explicitly pass the expected physical address to the hw
v3: Also remember the wild writes we once had for HWS above 4G.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903152304.31589-2-chris@chris-wilson.co.uk

drm/i915: Combine cleanup_status_page()

Pull the physical status page cleanup into a common
cleanup_status_page() for caller simplicity.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903152304.31589-1-chris@chris-wilson.co.uk

drm/i915: Fix up FORCE_GPU_RELOC (debug) to flush CPU write domains

We currently assert that if the target is in a CPU write domain, we use
a CPU reloc path rather than the GPU reloc path. However, we have a debug
override to force the GPU path and that unfortunately hits the assert.
Include the async clflush under the debug option to ensure correct
behaviour even when debugging, and strict when not.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903150216.19965-1-chris@chris-wilson.co.uk

drm/i915: Update DRIVER_DATE to 20180903

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Forcibly flush unwanted requests in drop-caches

Add a mode to debugfs/drop-caches to flush unwanted requests off the GPU
(by wedging the device and resetting). This is very useful if a test
terminated leaving a long queue of hanging batches that would ordinarily
require a round trip through hangcheck for each.

It reduces the inter-test operation to just a write into drop-caches to
reset driver/GPU state between tests.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903083337.13134-5-chris@chris-wilson.co.uk

drm/i915: Early rejection of buffer allocations larger than RAM

We currently try to pin and allocate the whole buffer at a time. If that
object is larger than RAM, we will try to pin the whole of physical
memory, force the machine into oom, and then still fail the allocation.

If the request is obviously too large, error out early. We opt to do
this in the backend to make it easy to use alternate paths that do not
require the entire object pinned, or may easily handle proxy objects
that are larger than physical memory.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903083337.13134-4-chris@chris-wilson.co.uk

drm/i915: Force the slow path after a user-write error

If we fail to write the user relocation back when it is changed, force
ourselves to take the slow relocation path where we can handle faults in
the write path. There is still an element of dubiousness as having
patched up the batch to use the correct offset, it no longer matches the
presumed_offset in the relocation, so a second pass may miss any changes
in layout.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903083337.13134-3-chris@chris-wilson.co.uk

drm/i915: Flag any possible writes for a GTT fault

We do not explicitly mark the PTE for the user's GTT mmap as being
wrprotect, so we don't get a refault when we would need to change a
read-only mmapping into read-write. As such, we must presume that if the
vma has PROT_WRITE it may be written to, although this is supposed to be
indicated by set-domain there are cases (e.g. after swap) where
userspace may not be aware of the implicit domain change.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903083337.13134-2-chris@chris-wilson.co.uk

drm/i915: Do a full device reset after being wedged

We only call unset_wedged on the global reset path (since it's a global
operation), so if we are terminally wedged and wish to reset, take the
full device reset path rather than the quicker individual engine resets.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180903083337.13134-1-chris@chris-wilson.co.uk

drm/i915: Determine uses-full-ppgtt from context for execbuf

Rather than inspect the global module parameter for whether full-ppgtt
maybe enabled, we can inspect the context directly as to whether it has
its own vm.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Bob Paauwe <bob.j.paauwe@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180901092451.7233-1-chris@chris-wilson.co.uk

drm/i915: Explicitly mark Global GTT address spaces

So far we have been relying on vm->file pointer being NULL to declare
something GGTT.

This has the unfortunate consequence that the default kernel context is
also declared GGTT and interferes with the following patch which wants to
instantiate VMA's and execute requests against the kernel context.

Change the is_ggtt test to use an explicit flag in struct address_space to
solve this issue.

Note that the bit used is free since there is an alignment hole in the
struct.

v2:
 * Mark mock ggtt.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180831143643.12366-1-tvrtko.ursulin@linux.intel.com

drm/i915/dp_mst: Fix enabling pipe clock for all streams

commit afb2c4437dae ("drm/i915/ddi: Push pipe clock enabling to encoders")
inadvertently stopped enabling the pipe clock for any DP-MST stream
after the first one. It also rearranged the pipe clock enabling wrt.
initial MST payload allocation step (which may or may not be a
problem, but it's contrary to the spec.).

Fix things by making the above commit truly a non-functional change.

Fixes: afb2c4437dae ("drm/i915/ddi: Push pipe clock enabling to encoders")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107365
Reported-by: Lyude Paul <lyude@redhat.com>
Reported-by: dmummenschanz@web.de
Tested-by: dmummenschanz@web.de
Tested-by: Lyude Paul <lyude@redhat.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: dmummenschanz@web.de
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180831174739.30387-1-imre.deak@intel.com

drm/i915/perf: reuse intel_lrc ctx regs macro

Abstract the context image access a bit.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180813080218.28994-3-tvrtko.ursulin@linux.intel.com

drm/i915/perf: simplify configure all context function

We don't need any special treatment on error so just return as soon as
possible.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180813080218.28994-2-tvrtko.ursulin@linux.intel.com

drm/i915/intel_csr.c Fix DMC FW Loading issue on ICL.

This patch resolves the DMC FW loading issue.
Earlier DMC FW package have only one DMC FW for one stepping. But as such
there is no such restriction from Package side.
For ICL icl_dmc_ver1_07.bin binary package has DMC FW for 2 steppings.
So while reading the dmc_offset from package header, for 1st stepping
offset used to come 0x0 and was working fine till now.
But for second stepping and other steppings, offset is non zero number
and is in dwords. So we need to convert into bytes to fetch correct DMC
FW from correct place.

v2 : Added check for DMC FW max size for various gen. (Imre Deak)
v3 : Corrected naming convention for various gen. (Imre Deak)
v4 : Initialized max_fw_size to 0
v5 : Corrected DMC FW MAX_SIZE for various gen. (Imre Deak)
v6 : Fixed the typo issues.

Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Jyoti Yadav <jyoti.r.yadav@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1535695223-4648-1-git-send-email-jyoti.r.yadav@intel.com

drm/i915/selftests: Add a simple exerciser for suspend/hibernate

Although we cannot do a full system-level test of suspend/hibernate from
deep with the kernel selftests, we can exercise the GEM subsystem in
isolation and simulate the external effects (such as losing stolen
contents and trashing the register state).

v2: Don't forget to hold rpm
v3: Suspend the GTT mappings, and more rpm!

References: https://bugs.freedesktop.org/show_bug.cgi?id=96526
References: 5ab57c702069 ("drm/i915: Flush logical context image out to memory upon suspend")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jakub Bartmiński <jakub.bartminski@intel.com>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Jakub Bartmiński <jakub.bartminski@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180830134806.21939-1-chris@chris-wilson.co.uk

drm/i915: Keep physical cursors pinned while in use

The optimisation inherent in commit 6a2c4232ece1 ("drm/i915: Make the
physical object coherent with GTT") relies on that once we allocated a
cursor we would have coherent, zero overhead access to the scanout plane
holding the cursor. That is we could then do the very frequent cursor
updates X enjoys with no indirection or kernel involvement. However,
that all hinges on the GGTT mmap of the cursor being pinned and not
require refaulting on each access -- handling such a page fault likely
requires the busy GGTT to be rearranged causing a stall. A very simple
fix is then to handle the physical cursor exactly like other cursors and
keep its vma pinned while active.

References: https://bugs.freedesktop.org/show_bug.cgi?id=107600
References: 6a2c4232ece1 ("drm/i915: Make the physical object coherent with GTT")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180817082405.755-1-chris@chris-wilson.co.uk

drm/i915/ringbuffer: Delay after invalidating gen6+ xcs

During stress testing of full-ppgtt (on Baytrail at least), we found
that the invalidation around a context/mm switch was insufficient (writes
would go astray). Adding a second MI_FLUSH_DW barrier prevents this, but
it is unclear as to whether this is merely a delaying tactic or if it is
truly serialising with the TLB invalidation. Either way, it is
empirically required.

v2: Avoid the loop for readability;

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107715
References: https://bugs.freedesktop.org/show_bug.cgi?id=107759
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180830161042.29193-1-chris@chris-wilson.co.uk

drm/i915: clear error registers after error capture

We need to clear the register in order to get correct value after the
next potential hang.

v2: Centralize error register clearing in i915_irq.c (Chris)

v3: Don't read gen8 register on < gen6 (Chris)

v4: Don't swap gen8+ & gen6+ code... (Chris)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180830132424.21940-1-lionel.g.landwerlin@intel.com

drm/i915: Update DRIVER_DATE to 20180830

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915: Update DRIVER_DATE to 20180830

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

drm/i915/execlists: Flush tasklet directly from reset-finish

On finishing the reset, the intention is to restart the GPU before we
relinquish the forcewake taken to handle the reset - the goal being the
GPU reloads a context before it is allowed to sleep. For this purpose,
we used tasklet_flush() which although it accomplished the goal of
restarting the GPU, carried with it a sting in its tail: it cleared the
TASKLET_STATE_SCHED bit. This meant that if another CPU queued a new
request to this engine, we would clear the flag and later attempt to
requeue the tasklet on the local CPU, breaking the per-cpu softirq
lists.

Remove the dangerous tasklet_kill() and just run the tasklet func
directly as we know it is safe to do so (the tasklets are internally
locked to allow mixed usage from direct submission).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180828152702.27536-1-chris@chris-wilson.co.uk

drm/i915: Don't check power domains state in intel_power_domains_init_hw()

During power domains initialization we acquire power well references for
power wells in the INIT power domain. The rest of power wells - which
BIOS could have left enabled - we can only acquire references as needed
during display HW readout and so must defer sanitization until then
(also implying that we must always do HW readout to cleanup unused power
wells).

Thus during initialization these latter power wells can have a refcount
of 0 while still being enabled. To avoid the false-positive state
mismatch error this causes remove the check from
intel_power_domains_init_hw() and rely on the state check in
intel_power_domains_enable() which follows the HW readout.

v2:
- Add comment to log and code clarifying how unused power wells get
  disabled. (Chris)

Fixes: 6dfc4a8f134f ("drm/i915: Verify power domains after enabling them")
Cc: Chris Wilson <chris@chris-wilson.co.uk>
References: https://bugs.freedesktop.org/show_bug.cgi?id=107411
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180828122231.14336-1-imre.deak@intel.com

drm/i915/dsc: Fix PPS register definition macros for 2nd VDSC engine

This patch fixes the PPS4 and PPS5 register definition macros that were
resulting into an incorect MMIO address.

Fixes: 2efbb2f099fb ("i915/dp/dsc: Add DSC PPS register definitions")
Cc: Anusha Srivatsa <anusha.srivatsa@intel.com>
Signed-off-by: Manasi Navare <manasi.d.navare@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Anusha Srivatsa <anusha.srivatsa@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180824014807.14681-1-manasi.d.navare@intel.com

drm/i915: Do not advertize support for NV12 on ICL yet.

ICL requires two planes for scanning out a NV12 framebuffer. Do
not advertize support for creating NV12 framebuffers until required
plane programming is implemented.

v2: Do not allow adding buffers.
    Check inside skl_plane_has_planar (Ville)

Bspec: Plane Planar YUV programming (18566)
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180824203856.17700-2-dhinakaran.pandiyan@intel.com

drm/i915: Clean up skl_plane_has_planar()

skl_plane_has_planar is hard to read, simplify the logic by checking for
support in the order of platform, pipe and plane.

No change in functionality intended.
v2: Fix logic for primary plane (Ville)

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180827225624.4912-1-dhinakaran.pandiyan@intel.com

drm/i915: Reject compressed Y/Yf with interlaced modes

Y/Yf tiling can't be used with IF-ID. We already reject uncompressed
Y/Yf but we should also reject them when compressed.

Cc: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180828142707.31583-2-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Mahesh Kumar <mahesh1.sh.kumar@gmail.com>

drm/i915: Don't pass plane to .check_plane()

.check_plane() already gets the plane state, so we can dig out the plane
from there if needed. No need in passing it separately.

Cc: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180828142707.31583-1-ville.syrjala@linux.intel.com
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

drm/i915: Fix gtt_view asserts

gcc is too smart for us and doesn't evaluate BUILD_BUG_ON()s in
unused static inlines. Collect them up in one static inline and
actually call it to make sure gcc sees it.

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Suggested-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180828133723.18505-1-ville.syrjala@linux.intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>

drm/i915: Use dp_to_i915 on intel_psr.c

Now that we have a generic caller let's simplify it and
clean up the intel_psr.c code a bit.

Cc: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180827223021.7145-2-rodrigo.vivi@intel.com

drm/i915: introduce dp_to_i915() helper

No functional change. But let's get first i915 pointer
directly from intel_dp so we can clean up a lot of code
later.

Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180827223021.7145-1-rodrigo.vivi@intel.com

drm/i915/selftests: ring all doorbells in igt_guc_doorbells

We currently verify that all doorbells can be registered with GuC and
HW but don't check that all works as expected after a db ring.

Do a nop ring of all doorbells to make sure we haven't misprogrammed
any WQ or stage descriptor data. This will also help validating
upcoming changes in the db programming flow.

Cc: Michel Thierry <michel.thierry@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Michel Thierry <michel.thierry@intel.com>
Acked-by: Katarzyna Dec <katarzyna.dec@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180827223614.22789-1-daniele.ceraolospurio@intel.com

drm/i915/psr: Rewrite comments in intel_psr_wait_for_idle()

Added bspec reference, aligned text and documented the function.

Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180824230844.12428-2-dhinakaran.pandiyan@intel.com

drm/i915/psr: Remove wait_for_idle() for PSR2

CI runs show PSR2 does not go to IDLE with selective update enabled on
all PSR exit triggers. Specifically, logs indicate the hardware enters
"SLEEP Selective Update" and not "IDLE Reset state', like the kernel
expects, when vblank interrupts are enabled. This check was added for PSR1
but incorrectly extended to PSR2, remove the check as it breaks tests
and prints out misleading error messages.

v2: Split out non-code changes (Rodrigo)

Cc: Tarun Vyas <tarun.vyas@intel.com>
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Fixes: c43dbcbbcc8c ("drm/i915/psr: Lockless version of psr_wait_for_idle")
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180824230844.12428-1-dhinakaran.pandiyan@intel.com

drm/i915: Re-apply "Perform link quality check, unconditionally during long pulse"

This re-applies the workaround for "some DP sinks, [which] are a
little nuts" from commit 1a36147bb939 ("drm/i915: Perform link
quality check unconditionally during long pulse").
It makes the secondary AOC E2460P monitor connected via DP to an
acer Veriton N4640G usable again.

This hunk was dropped in commit c85d200e8321 ("drm/i915: Move SST
DP link retraining into the ->post_hotplug() hook")

Fixes: c85d200e8321 ("drm/i915: Move SST DP link retraining into the ->post_hotplug() hook")
[Cleaned up commit message, added stable cc]
Signed-off-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Jan-Marek Glogowski <glogow@fbihome.de>
Cc: stable@vger.kernel.org
Link: https://patchwork.freedesktop.org/patch/msgid/20180825191035.3945-1-lyude@redhat.com

drm/i915: move lookup_power_well() up

There's no need for that forward declaration.

Cc: Imre Deak <imre.deak@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180820233139.11936-4-paulo.r.zanoni@intel.com

drm/i915: use for_each_power_well in lookup_power_well()

Use the nice helper function to make the implementation simpler.

v2: Rebase.

Cc: Imre Deak <imre.deak@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com> (v1)
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180820233139.11936-3-paulo.r.zanoni@intel.com

drm/i915: WARN() if we can't lookup_power_well()

None of the current lookup_power_well() callers are actually checking
for NULL return values, they all just use the pointer right away.  The
first idea was to replace these theoretical segfaults with a BUG()
since this would at least make our code a little more explicit to the
reader. It was suggested that just converting the BUG() to a WARN()
and returning any power well would probably be better since it would
still keep the system running while at the same time exposing the
driver bug.

We can only hit this NULL/BUG()/WARN() condition if we try to lookup a
power well that isn't defined on a given platform. If that ever
happens, we have to fix our code, making it lookup the correct power
well. Because of this, I don't think it's worth trying to implement
error checking in every caller. Improving our CI system will be a
better use of our time once a bug is found in the wild.

v2: Avoid the BUG() with a WARN() return a random PW (Michal).

Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180820233139.11936-2-paulo.r.zanoni@intel.com

drm/i915/icl: implement the tc/legacy HPD {dis,}connect flows

Unlike the other ports, TC ports are not available to use as soon as
we get a hotplug. The TC PHYs can be shared between multiple
controllers: display, USB, etc. As a result, handshaking through FIA
is required around connect and disconnect to cleanly transfer
ownership with the controller and set the type-C power state.

This patch implements the flow sequences described by our
specification. We opt to grab ownership of the ports as soon as we get
the hotplugs in order to simplify the interactions and avoid surprises
in the user space side. We may consider changing this in the future,
once we improve our testing capabilities on this area.

v2:
 * This unifies the DP and HDMI patches so we can discuss everything
   at once so people looking at random single patches can actually
   understand the direction.
 * I found out the spec was updated a while ago. There's a small
   difference in the connect flow and the patch was updated for that.
 * Our spec also now gives a good explanation on what is really
   happening. As a result, comments were added.
 * Add some more comments as requested by Rodrigo (Rodrigo).
v3:
 * Downgrade a DRM_ERROR that shouldn't ever happen but we can't act
   on in case it does (Chris).

BSpec: 21750, 4250.
Cc: Animesh Manna <animesh.manna@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180801173441.9789-1-paulo.r.zanoni@intel.com

drm/i915: Free write_buf that we allocated with kzalloc.

We use kzalloc to allocate the write_buf that we use for
i2c transfer on hdcp write. But it seems that we are forgetting
to free the memory that is not needed after i2c transfer is
completed.

Reported-by: Brian J Wood <brian.j.wood@intel.com>
Fixes: 2320175feb74 ("drm/i915: Implement HDCP for HDMI")
Cc: Ramalingam C <ramalingam.c@intel.com>
Cc: Sean Paul <seanpaul@chromium.org>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: <stable@vger.kernel.org> # v4.17+
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180823205136.31310-1-rodrigo.vivi@intel.com

drm/i915: Simplify condition to keep DMC active during S0ix

For S0ix we want to deinit power domains (and so deactivate the DMC
firmware) exactly when the platform supports the DC9 state. To reach
S0ix we need DC9 on these platforms (for which the DMC FW needs to be
deactivated) while to reach S0ix on the rest of the DMC platforms we
need DC6 (which needs the DMC FW to stay active).

Simplify the condition accordingly so it will be automatically
correct for upcoming DC9 platforms like ICL.

Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180822112602.27543-1-imre.deak@intel.com

drm/i915: Rename PLANE_CTL_DECOMPRESSION_ENABLE

Rename PLANE_CTL_DECOMPRESSION_ENABLE to resemble the bpsec name -
PLANE_CTL_RENDER_DECOMPRESSION_ENABLE

Suggested-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180822015053.1420-2-dhinakaran.pandiyan@intel.com

drm/i915: Add a small wrapper to check for CCS modifiers.

Code looks cleaner with modifiers hidden inside this wrapper.
v2: Remove const qualifier (Ville)

Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180822193827.6341-1-dhinakaran.pandiyan@intel.com

drm/i915/psr: Add PSR mode/revision to debugfs

Log the PSR mode/revision (PSR1 or PSR2) in the debugfs file
i915_edp_psr_status.

Suggested-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: Azhar Shaikh <azhar.shaikh@intel.com>
Reviewed-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1534958628-193724-1-git-send-email-azhar.shaikh@intel.com

drm/i915: Fix glk/cnl display w/a #1175

The workaround was supposed to look at the plane destination
coordinates. Currently it's looking at some mixture of src
and dst coordinates that doesn't make sense. Fix it up.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180719182214.4323-2-ville.syrjala@linux.intel.com
Fixes: 394676f05bee (drm/i915: Add WA for planes ending close to left screen edge)
Reviewed-by: Imre Deak <imre.deak@intel.com>

drm/i915/psr: Mask PSR irq bits when re-enabling interrupts.

gen8_de_irq_postinstall() wasn't masking the IRQ bit before passing the
debug flag to psr_irq_control(). This check was missed when new debug bits
were defined in  'commit c44301fce614 ("drm/i915: Allow control of PSR at
runtime through debugfs, v6")'. Instead of ANDing the irq bit in all the
callers, move it to the callee.

v2: Rebased.

Fixes: c44301fce614 ("drm/i915: Allow control of PSR at runtime through
debugfs, v6")
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180821221156.2442-3-dhinakaran.pandiyan@intel.com

drm/i915/psr: Add missing check for I915_PSR_DEBUG_IRQ bit

We print the last attempted entry and last exit timestamps only when
IRQ debug is requested. This check was missed when new debug flags were
added in 'commit c44301fce614 ("drm/i915: Allow control of PSR at
runtime through debugfs, v6")

Fixes: c44301fce614 ("drm/i915: Allow control of PSR at runtime through
debugfs, v6")
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180821221156.2442-2-dhinakaran.pandiyan@intel.com

drm/i915/psr: Print PSR_STATUS when PSR idle wait times out.

Knowing the status of the PSR HW state machine is useful for debug,
especially since we are seeing errors with PSR2 in CI.

Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Dhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180821221156.2442-1-dhinakaran.pandiyan@intel.com

drm/i915: Correct CSB probing for engine state dumper

Since we no longer maintain our read position in the CSB pointers
register, it always returns 0 and not where we last read up to. As a
result the CSB probing in the state dumper starts from 0, either missing
entries or showing stale one.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180821101138.15822-1-chris@chris-wilson.co.uk

drm/i915/icl: Get DDI clock for ICL for MG PLL and TBT PLL

PLLs are the source clocks for the DDIs so in order to determine the
ddi clock we need to check the PLL configuration.

For MG PHy Ports (C - F), depending on whether it is a TBT PLL or MG
PLL the link lock can be obtained from the the PLL divisors based on
the specification.

v2 (from Paulo):
 * Make the algorithm look more like what's in the spec, also document
   where we differ form the spec and why.
 * Make the code a little more consistent with our coding style.

Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Manasi Navare <manasi.d.navare@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180817215209.29133-2-paulo.r.zanoni@intel.com

drm/i915/icl: Implement HSDIV_RATIO of MG_CLKTOP2_HSCLKCTL_PORT reg as separate divider value defines

The register value of Divider Ratio for high speed divider
(hsdiv_ratio) in MG_CLKTOP2_HSCLKCTL_PORT register is not same as the
actual numerical value of the divider. So this patch implements
separate divider value defines for that field.
icl_mg_pll_find_divisors() can use these defines instead of magic
register values.

The new defines are going to be used in the next patch.

v2 (from Paulo):
 * Rebase.
 * Make it look a little more like the rest of our code.
v3 (from Paulo):
 * Make hsdiv u32 now that it's a bit field (José).

Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Suggested-by: James Ausmus <james.ausmus@intel.com>
Signed-off-by: Manasi Navare <manasi.d.navare@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180817215209.29133-1-paulo.r.zanoni@intel.com

drm/i915/audio: Hook up component bindings even if displays are disabled

If the display has been disabled by modparam, we still want to connect
together the HW bits and bobs with the associated drivers so that we can
continue to manage their runtime power gating.

Fixes: 108109444ff6 ("drm/i915: Check num_pipes before initializing audio component")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Takashi Iwai <tiwai@suse.de>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Elaine Wang <elaine.wang@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180817100241.4628-1-chris@chris-wilson.co.uk

drm/i915: Increase LSPCON timeout

100 ms is not enough time for the LSPCON adapter on Intel NUC devices to
settle. This causes dropped display modes at boot or screen reconfiguration.
Empirical testing can reproduce the error up to a timeout of 190 ms. Basic
boot and stress testing at 200 ms has not (yet) failed.

Increase timeout to 400 ms to get some margin of error.

Changes from v1:
The initial suggestion of 1000 ms was lowered due to concerns about delaying
valid timeout cases.
Update patch metadata.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107503
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1570392
Fixes: 357c0ae9198a ("drm/i915/lspcon: Wait for expected LSPCON mode to settle")
Cc: Shashank Sharma <shashank.sharma@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: <stable@vger.kernel.org> # v4.11+
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Shashank Sharma <shashank.sharma@intel.com>
Signed-off-by: Fredrik Schön <fredrik.schon@gmail.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180817200728.8154-1-fredrik.schon@gmail.com

drm/i915: Verify power domains after enabling them

After
commit 2cd9a689e97b ("drm/i915: Refactor intel_display_set_init_power() logic")
it makes more sense to check the power domain/well refcounts after
enabling the power domains functionality. Before that it's guaranteed
that most power wells (in the INIT domain) will have a reference held,
so not an interesting state.

While at it also add the check after the init_hw/fini_hw, disable and
suspend/resume steps. Make the test optional on a Kconfig option since
it may add substantial overhead: on VLV/CHV the corresponding PUNIT reg
access for each power well may take up to 20ms.

v2:
- Add the state check to more spots. (Chris)

v3:
- During suspend check the state before deiniting display core.
  Afterwards DC states are disabled (and so the dc_off power well is
  enabled) even though we don't hold a reference on it.
- Do the test conditionally based on a new Kconfig option. (Chris)

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
[Add DRM_I915_DEBUG_RUNTIME_PM to welcome messages]
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180817145837.26592-1-imre.deak@intel.com

drm/i915: Do not redefine the has_csr parameter.

Let us reuse the already defined has_csr check and not
redefine it.

The main difference is that in effect this will flip .has_csr to 1
(via GEN9_FEATURES which GEN11_FEATURES pulls in).

Suggested-by: Imre Deak <imre.deak@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Anusha Srivatsa <anusha.srivatsa@intel.com>
Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=107382
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1534527210-16841-1-git-send-email-anusha.srivatsa@intel.com

drm/i915/execlists: Include reset depth in traces

Show the reset depth (the tasklet disable count) in the GEM_TRACE to
indicate when we might not expect tasklets to be flushed.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180815135827.25869-1-chris@chris-wilson.co.uk

drm/i915: remove confusing GPIO vs PCH_GPIO

Instead of defining all registers twice, define just a PCH_GPIO_BASE
that has the same address as PCH_GPIO_A and use that to calculate all
the others. This also brings VLV and !HAS_GMCH_DISPLAY in line, doing
the same thing.

v2: Fix GMBUS registers to be relative to gpio base; create GPIO()
    macro to return a particular gpio address and move the enum out of
    i915_reg.h (suggested by Jani)

v3: Move base offset inside the GPIO() macro so the GMBUS defines don't
    actually need to be changed (suggested by Daniel/Ville)

v4: Move definition of i915_gpio to intel_display.h and remove
    GMBUS/GPIO handling from gvt since now they have their own
    defines.

Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180727193647.8639-3-lucas.demarchi@intel.com

drm/i915/gvt: use its own define for gpio

The definition on i915_reg.h is going to change to depend on
dev_priv->gpio_mmio_base being properly initialized. Define our own
macros since init_generic_mmio_info() is called before than
gpio_mmio_base being set.

Cc: intel-gvt-dev@lists.freedesktop.org
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180727193647.8639-2-lucas.demarchi@intel.com

drm/i915: make PCH_GMBUS* definitions private to gvt

This is the only place that they are being used - the others use the
GMBUS* macros that rely on dev_priv being already properly initialized.

Cc: intel-gvt-dev@lists.freedesktop.org
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180727193647.8639-1-lucas.demarchi@intel.com

drm/i915: Refactor intel_display_set_init_power() logic

The device global init_power_on flag is somewhat arbitrary and makes
debugging power refcounting problems difficult. Instead arrange things
so that all display power domain get has a corresponding put call. After
this change we have the following sequences:

driver loading:
intel_power_domains_init_hw();
<other init steps>
intel_power_domains_enable();

driver unloading:
intel_power_domains_disable();
<other uninit steps>
intel_power_domains_fini_hw();

system suspend:
intel_power_domains_disable();
<other suspend steps>
intel_power_domains_suspend();

system resume:
intel_power_domains_resume();
<other resume steps>
intel_power_domains_enable();

at other times while the driver is loaded:
intel_display_power_get();
...
intel_display_power_put();

Suggested-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180816123757.3286-2-imre.deak@intel.com

drm/i915: Introduce intel_runtime_pm_disable to pair intel_runtime_pm_enable

Currently, we cancel the extra wakeref we have for !runtime-pm devices
inside power_wells_fini_hw. However, this is not strictly paired with
the acquisition of that wakeref in runtime_pm_enable (as the fini_hw may
be called on errors paths before we even call runtime_pm_enable). Make
the symmetry more explicit and include a check that we do release all of
our rpm wakerefs.

v2: Fixup transfer of ownership back to core whilst keeping our wakeref
count balanced.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Imre Deak <imre.deak@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180816123757.3286-1-imre.deak@intel.com

drm/i915: Stop holding a ref to the ppgtt from each vma

The context owns both the ppgtt and the vma within it, and our activity
tracking on the context ensures that we do not release active ppgtt. As
the context fulfils our obligations for active memory tracking, we can
relinquish the reference from the vma.

This fixes a silly transient refleak from closed vma being kept alive
until the entire system was idle, keeping all vm alive as well.

Reported-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Testcase: igt/gem_ctx_create/files
Fixes: 3365e2268b6b ("drm/i915: Lazily unbind vma on close")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Tested-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180816073448.19396-1-chris@chris-wilson.co.uk

drm/i915: Remove useless error return from intel_init_mocs_engine()

As the only error is for a programming error in constructing the static
tables describing the register values, replace the error code
propagation with an assert.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180815184251.5850-1-chris@chris-wilson.co.uk

drm/i915: Only skip connector output for disable_display

We want to add no connectors, encoders or crtcs if the display is
disabled, but we still need to hook up any existing HW so that we can
power it down.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180815201207.2203-1-chris@chris-wilson.co.uk

drm/i915: Fix PM refcounting w/o DMC firmware

The case where the firmware isn't specified for a platform (although
runtime PM works only with DMC on this platform) is the same case where
the firmware is specified but can't be loaded for some reason. Hence we
need to get a display init power domain ref in the first case too to
keep the refcount bookkeeping in balance.

Also convert the related log message to be a debug one, since it's a
valid scenario for a new platform, where we need to have
dev_info->has_csr=1 set, but add support for actually loading the
firmware only later.

v2:
- In addition to the debug log, WARN on non-alpha support platforms,
  since then the first case isn't valid scenario. (Chris)

References: https://bugs.freedesktop.org/show_bug.cgi?id=107382
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Anusha Srivatsa <anusha.srivatsa@intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180815131038.24446-1-imre.deak@intel.com

drm/i915: Clear stop-engine for a pardoned reset

If we pardon a per-engine reset, we may leave the STOP_RING bit asserted
in RING_MI_MODE resulting in the engine hanging. Unconditionally clear
it on the per-engine exit path as we know that either we skipped the
reset and so need the cancellation, or the reset was successful and the
cancellation is a no-op, or there was an error and we will follow up
with a full-reset or wedging (both of which will stop the engines again
as required).

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107188
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106560
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180814171857.24673-1-chris@chris-wilson.co.uk

drm/i915: Disable runtime-pm using lowlevel functions if !HAS_RC6

If we cannot setup rc6, we cannot let the GPU suspend itself as it
cannot save its state (to a powercontext). As such, we must disable
runtime-pm, but we should do so using the low-level pm-runtime function
which leaves our own debugging functions intact (and continue to detect
errors in our runtime-pm handling should we ever be able to enable rc6).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180812223642.24865-3-chris@chris-wilson.co.uk

drm/i915: set DP Main Stream Attribute for color range on DDI platforms

Since Haswell we have no color range indication either in the pipe or
port registers for DP. Instead, there's a separate register for setting
the DP Main Stream Attributes (MSA) directly. The MSA register
definition makes no references to colorimetry, just a vague reference to
the DP spec. The connection to the color range was lost.

Apparently we've failed to set the proper MSA bit for limited, or CEA,
range ever since the first DDI platforms. We've started setting other
MSA parameters since commit dae847991a43 ("drm/i915: add
intel_ddi_set_pipe_settings").

Without the crucial bit of information, the DP sink has no way of
knowing the source is actually transmitting limited range RGB, leading
to "washed out" colors. With the colorimetry information, compliant
sinks should be able to handle the limited range properly. Native
(i.e. non-LSPCON) HDMI was not affected because we do pass the color
range via AVI infoframes.

Though not the root cause, the problem was made worse for DDI platforms
with commit 55bc60db5988 ("drm/i915: Add "Automatic" mode for the
"Broadcast RGB" property"), which selects limited range RGB
automatically based on the mode, as per the DP, HDMI and CEA specs.

After all these years, the fix boils down to flipping one bit.

[Per testing reports, this fixes DP sinks, but not the LSPCON. My
 educated guess is that the LSPCON fails to turn the CEA range MSA into
 AVI infoframes for HDMI.]

Reported-by: Michał Kopeć <mkopec12@gmail.com>
Reported-by: N. W. <nw9165-3201@yahoo.com>
Reported-by: Nicholas Stommel <nicholas.stommel@gmail.com>
Reported-by: Tom Yan <tom.ty89@gmail.com>
Tested-by: Nicholas Stommel <nicholas.stommel@gmail.com>
References: https://bugs.freedesktop.org/show_bug.cgi?id=100023
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107476
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=94921
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: <stable@vger.kernel.org> # v3.9+
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180814060001.18224-1-jani.nikula@intel.com