Again just complicates gem init functions and makes a general mess out
of everything.

Good riddance!

v2: In my enthusiasm to start removing dri1/ums crud I went overboard a
bit and killed parts of hangcheck. Resurrect it.
Acked-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

We've killed ums support by now, it's time to reap the benefits. This
one here is getting in the way of doing some ring init cleanup.
Acked-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@intel.com>

With the deprecation of UMS, and by association DRI1, we have a tough
choice when updating the ring access routines. We either rewrite the
DRI1 routines blindly without testing (so likely to be broken) or take
the liberty of declaring them no longer supported and remove them
entirely. This takes the latter approach.

v2: Also remove the DRI1 sarea updates
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
[danvet: Fix rebase conflicts.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Up until now, we have pinned every logical ring context backing object
during creation, and left it pinned until destruction. This made my life
easier, but it's a harmful thing to do, because we cause fragmentation
of the GGTT (and, eventually, we would run out of space).

This patch makes the pinning on-demand: the backing objects of the two
contexts that are written to the ELSP are pinned right before submission
and unpinned once the hardware is done with them. The only context that
is still pinned regardless is the global default one, so that the HWS can
still be accessed in the same way (ring->status_page).

v2: In the early version of this patch, we were pinning the context as
we put it into the ELSP: on the one hand, this is very efficient because
only a maximum two contexts are pinned at any given time, but on the other
hand, we cannot really pin in interrupt time :(

v3: Use a mutex rather than atomic_t to protect pin count to avoid races.
Do not unpin default context in free_request.

v4: Break out pin and unpin into functions.  Fix style problems reported
by checkpatch

v5: Remove unpin_lock as all pinning and unpinning is done with the struct
mutex already locked.  Add WARN_ONs to make sure this is the case in future.

Issue: VIZ-4277
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Signed-off-by: NThomas Daniel <thomas.daniel@intel.com>
Reviewed-by: NAkash Goel <akash.goels@gmail.com>
Reviewed-by: Deepak S<deepak.s@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

When disabling the RPS interrupts there is a tricky dependency between
the thread disabling the interrupts, the RPS interrupt handler and the
corresponding RPS work. The RPS work can reenable the interrupts, so
there is no straightforward order in the disabling thread to (1) make
sure that any RPS work is flushed and to (2) disable all RPS
interrupts. Currently this is solved by masking the interrupts using two
separate mask registers (first level display IMR and PM IMR) and doing
the disabling when all first level interrupts are disabled.

This works, but the requirement to run with all first level interrupts
disabled is unnecessary making the suspend / unload time ordering of RPS
disabling wrt. other unitialization steps difficult and error prone.
Removing this restriction allows us to disable RPS early during suspend
/ unload and forget about it for the rest of the sequence. By adding a
more explicit method for avoiding the above race, it also becomes easier
to prove its correctness. Finally currently we can hit the WARN in
snb_update_pm_irq(), when a final RPS work runs with the first level
interrupts already disabled. This won't lead to any problem (due to the
separate interrupt masks), but with the change in this and the next
patch we can get rid of the WARN, while leaving it in place for other
scenarios.

To address the above points, add a new RPS interrupts_enabled flag and
use this during RPS disabling to avoid requeuing the RPS work and
reenabling of the RPS interrupts. Since the interrupt disabling happens
now in intel_suspend_gt_powersave(), we will disable RPS interrupts
explicitly during suspend (and not just through the first level mask),
but there is no problem doing so, it's also more consistent and allows
us to unify more of the RPS disabling during suspend and unload time in
the next patch.

v2/v3:
- rebase on patch "drm/i915: move rps irq disable one level up" in the
  patchset
Signed-off-by: NImre Deak <imre.deak@intel.com>
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

In sandybridge_pcode_read and sandybridge_pcode_write,
extend the mbox parameter from u8 to u32.

On Haswell and Sandybridge, bits 7:0 encode the mailbox
command and bits 28:8 are used for address control for
specific commands.

Based on suggestion from Ville Syrjälä.
Signed-off-by: NTom O'Rourke <Tom.O'Rourke@intel.com>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

On skylake, DPLL 1, 2 and 3 can be used for DP and HDMI. The shared dpll
framework allows us to share those DPLLs among DDIs when possible.

The most tricky part is to provide a DPLL state that can be easily
compared. DPLL_CRTL1 is shared by all the DPLLs, 6 bits each. The
per-dpll crtl1 field of the hw state is then normalized to be the same
value if 2 DPLLs do indeed have identical values for those 6 bits.

v2: Port the code to the shared DPLL infrastructure (Damien)

v3: Rebase on top of Ander's clock computation staging work for atomic (Damien)

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com> (v2)
Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v1)
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Adding structure/enum for SKL clocking implementation.

v2: Addressed Damien's comment
	- Removed internal structure from this header file

v3: Stove this into the generic intel_dpll_id enum and give them the established
DPLL_ID_ prefixes. (Daniel)

v4: - We'll only try to share DPLL1/2/3, leaving DPLL0 to eDP
    - Use SKL in the skylake shared DPLL names
    - Re-add the skl_dpll enum
    (Damien)

v5: Remove SKL_DPLL_NONE (Daniel)

v6: Modified as per review comments from Paulo
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: Satheeshakrishna M <satheeshakrishna.m@intel.com> (v2)
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> (v4,v5)
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> (v3)
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This is not used within the driver, and merely saving/restoring these
registers isn't going to do any good anyway. In fact, it's possible it's
actively harmful. Any code enabling the feature should handle this
completely in the regular platform specific enable/disable backlight
functions.
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

On VLV/CHV both pipes A and B have their own backlight control
registers. In order to correctly read out the current hardware state at
init we need to know which pipe is driving the eDP port. Pass that
information down from the eDP init code into the backlight code.

To determine the correct pipe we first look at which pipe is currently
configured in the port control register, if that look invalid we look
at which pipe's PPS is currently controlling the port, and if that
too looks invalid we just assume pipe A.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Currently objects for which the hardware needs a contiguous physical
address are allocated a shadow backing storage to satisfy the contraint.
This shadow buffer is not wired into the normal obj->pages and so the
physical object is incoherent with accesses via the GPU, GTT and CPU. By
setting up the appropriate scatter-gather table, we can allow userspace
to access the physical object via either a GTT mmaping of or by rendering
into the GEM bo. However, keeping the CPU mmap of the shmemfs backing
storage coherent with the contiguous shadow is not yet possible.
Fortuituously, CPU mmaps of objects requiring physical addresses are not
expected to be coherent anyway.

This allows the physical constraint of the GEM object to be transparent
to userspace and allow it to efficiently render into or update them via
the GTT and GPU.

v2: Fix leak of pci handle spotted by Ville
v3: Remove the now duplicate call to detach_phys_object during free.
v4: Wait for rendering before pwrite. As this patch makes it possible to
render into the phys object, we should make it correct as well!
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NRodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We need the HPLL frequency when calculating cdclk. Currently we read
that out from the hardware every single time, which isn't going to fly
very well if the device is runtime suspended. So cache the HPLL
frequency in dev_priv and use the cached value.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reference: https://bugs.freedesktop.org/show_bug.cgi?id=82939Reviewed-by: NImre Deak <imre.deak@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

So that it can be used by the flip code to wait for rendering without
holding any locks.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Implement common forcewake functions shared by Gen9 features.

v2: Make the focewake_{get,put} functions static (Mika)
    Small coding style fix in the function definition (Damien)
Reviewed-by: NMika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: Zhe Wang <zhe1.wang@intel.com> (v1)
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> (v2)
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

To correctly flush the new DDB allocation we need to know about the pipe
allocation layout inside the DDB in order to sequence the re-allocation
to not cause a newly allocated pipe to fetch from a space that was
previously allocated to another pipe.

This patch preserves the per-pipe (start,end) allocation to be used in
the flush.
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Ville suggested that we should use the same semantics as C arrays to
reduce the number of those pesky +1/-1 in the allocation code.

This patch leaves the debugfs file as is, showing the internal DDB
allocation structure, not the values written in the registers.

v2: Remove the test on ->end in skl_ddb_entry_size() (Ville)
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Suggested-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

v2: Don't check DDB on pre-SKL platforms
    Don't check DDB state on disabled pipes

v3: Squash "Expose skl_ddb_get_hw_state()"
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This patch implements the watermark algorithm and its necessary
functions. Two function pointers skl_update_wm and
skl_update_sprite_wm are provided. The skl_update_wm will update
the watermarks for the crtc provided as an argument and then
checks for change in DDB allocation for other active pipes and
recomputes the watermarks for those Pipes and planes as well.
Finally it does the register programming for all dirty pipes.
The trigger of the Watermark double buffer registers will have
to be once the plane configurations are done by the caller.

v2: fixed the divide-by-0 error in the results computation func.
    Also reworked the PLANE_WM register values computation func to
    make it more compact. Incorporated all other review comments
    from Damien.

v3: Changed the skl_compute_plane_wm function to now return success
    or failure. Also the result blocks and lines are computed here
    instead of in skl_compute_wm_results function.

v4: Adjust skl_ddb_alloc_changed() to the new planes/cursor split
    (Damien)

v5: Reworked the affected functions to implement new plane/cursor
    split.

v6: Rework the logic that triggers the DDB allocation and WM computation
    of skl_update_other_pipe_wm() to not depend on non-computed DDB
    values.
    Always give a valid cursor_width (at boot it's 0) to keep the
    invariant that we consider the cursor plane always enabled.
    Otherwise we end up dividing by 0 in skl_compute_plane_wm()
    (Damien Lespiau)

v7: Spell out allocation
    skl_ddb_ functions should have the ddb as first argument
    Make the skl_ddb_alloc_changed() parameters const
    (Damien)

v8: Rebase on top of the crtc->primary changes

v9: Split the staging results structure to not exceed the 1Kb stack
    allocation in skl_update_wm()

v10: Make skl_pipe_pixel_rate() take a pointer to the pipe config
     Add a comment about overflow considerations for skl_wm_method1()
     Various additions of const
     Various use of sizeof(variable) instead of sizeof(type)
     Various move of variable definitons to a narrower scope
     Zero initialize some stack allocated structures to make sure we
     don't have garbage in case we don't write all the values
     (Ville)

v11: Remove non-necessary default number of blocks/lines when the plane
     is disabled (Ville)
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NPradeep Bhat <pradeep.bhat@intel.com>
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We now need to allocate space in the DDB for planes being scanned out
ourselves. The data structure to represent an allocation mirrors what
we'll need to write in the registers later on: (start, end).

We add that allocation datat to the skl_wm_values structure as part of
the values to program the hardware with.

v2: Split planes and cursor for consistency.

v3: Make the skl_ddb_entry_size() parameter const
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This patch defines the structures needed for computation of
watermarks of pipes and planes for SKL.

v2: Incorporated Damien's review comments and removed unused fields
    in structs for future features like rotation, drrs and scaling.
    The skl_wm_values struct is now made more generic across planes
    and cursor planes for all pipes.

v3: implemented the plane/cursor split.

v4: Change the wm union back to a structure (Ville, Daniel)
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NPradeep Bhat <pradeep.bhat@intel.com>
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This patch reads the memory latency values for all the 8 levels for
SKL. These values are needed for the Watermark computation.

v2: Incorporated the review comments from Damien on register
    indentation.

v3: Updated the code to use the sandybridge_pcode_read for reading
    memory latencies for GEN9.

v4: Don't put gen 9 in the middle of an ordered list of ifs
    (Damien)

v5: take the rps.hw_lock around sandybridge_pcode_read() (Damien)

v6: Use gen >= 9 in the pcode_read() function for data1.
    Move the defines near the gen6 ones and prefix them with PCODE.
    Remove unused timeout define (the pcode_read() code has a larger
    timeout already).
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NPradeep Bhat <pradeep.bhat@intel.com>
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The function was removed in:

  commit 037bde19
  Author: Chris Wilson <chris@chris-wilson.co.uk>
  Date:   Thu Mar 27 08:24:19 2014 +0000

      Revert "drm/i915: Disable/Enable PM Intrrupts based on the current freq."
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Chris removed the code using it in:

  commit be2d599b
  Author: Chris Wilson <chris@chris-wilson.co.uk>
  Date:   Wed Sep 10 19:52:18 2014 +0100

      drm/i915: Remove dead code, i915_gem_verify_gtt
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

There's no users left after the conversion to calculate clocks before
disabling crtcs during mode set.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

It is possible for a mode set to fail if there aren't shared DPLLS that
match the new configuration requirement or other errors in clock
computation. If that step is executed after disabling crtcs, in the
failure case the hardware configuration is changed and needs to be
restored. Doing those things early will allow the mode set to fail
before actually touching the hardware.

Follow up patches will convert different platforms to use the new
infrastructure.

v2: Keep pll->new_config valid only during mode set (Ville)
    Use kmemdup() in i915_shared_dpll_start_config() (Ville)
    Restore old pll config if something fails before commit (Ville)
    Don't set compute_clock hooks since dev_priv is kzalloc()'d (Ville)
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The new struct will be used in a follow up patch to allow a current and
a staged config to exist for the same shared DPLL.

v2: Rebase on by mask_to_refcount()->hweight32() change. (Damien)
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This will be used in a follow up patch to properly release shared DPLLs
without relying on the shared_dpll field in pipe_config.

v2: Fix white space error (Ville)
    Use hweight32() (Ville)
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

When looking at the bug report logs with triggered
WARN_ON, the person doing bug triaging will have to
find exact kernel source and match file/line.

Attach the condition that triggered the WARN_ON
to kernel log. In most cases the context is self
evident and this way we can save developer time.

The drawback is ~16kbytes bigger i915.ko
Signed-off-by: NMika Kuoppala <miku@iki.fi>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Introduce functions to enable/disable the audio codec, incorporating the
ELD setup within enable. The disable is initially limited to HSW,
covering exactly what was done previously.

The only functional difference is that ELD valid is no longer set if
there is no connector with ELD, which should be the right thing to do
anyway. Otherwise the sequence remains the same, with warts and all, in
preparation for applying more sanity.

v2: add kernel doc.
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This will simplify things later on. No functional changes.
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

If these flags are on the object level it will be more difficult to allow
for multiple VMAs per object.

v2: Simplification and cleanup after code review comments (Chris Wilson).
Signed-off-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

By now we handle switcheroo and legacy suspend/resume the same way, so
no need to keep separate functions for them.

No functional change.
Signed-off-by: NImre Deak <imre.deak@intel.com>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Some machines (like MBAs) might use a tiled framebuffer but not enable
display swizzling at boot time.  We want to preserve that configuration
if possible to prevent a boot time mode set.  On IVB+ it shouldn't
affect performance anyway since the memory controller does internal
swizzling anyway.

For most other configs we'll be able to enable swizzling at boot time,
since the initial framebuffer won't be tiled, thus we won't see any
corruption when we enable it.

v2: preserve swizzling if BIOS had it set (Daniel)
v3: preserve swizzling only if we inherited a tiled framebuffer (Daniel)
    check display swizzle setting in detect_bit_6_swizzle (Daniel)
    use gen6 as cutoff point (Daniel)
v4: fixup swizzle preserve again, had wrong init order (Daniel)
Reported-by: NKristian Høgsberg <hoegsberg@gmail.com>
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

If we build the workaround list in ring initialization
and decouple it from the actual writing of values, we
gain the ability to decide where and how we want to apply
the values.

The advantage of this will become more clear when
we need to initialize workarounds on older gens where
it is not possible to write all the registers through ring
LRIs.

v2: rebase on newest bdw workarounds

Cc: Arun Siluvery <arun.siluvery@linux.intel.com>
Cc: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: NArun Siluvery <arun.siluvery@linux.intel.com>
[danvet: Resolve tiny conflict in comments and ocd alignments a bit.]
[danvet2: Remove bogus force_wake_get call spotted by Paulo and QA.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

For consistency, since that's the rule followed for internal functions.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

For consistency, since that's the rule followed for internal functions.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Just a couple more macros that assume that they were being passed a
struct drm_device when they want a struct drm_i915_private. Use our
magic macro to ease transitioning over to using drm_i915_privates
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>