Revert "drm/radeon/dpm: drop clk/voltage dependency filters for SI"

This reverts commit 186b1b2ba2a0684e3d2d3703427a993a3b35b16d.

There are still some stability problems on some SI boards so bring
this back.

Revert "drm/radeon: drop btc_get_max_clock_from_voltage_dependency_table"

This reverts commit fc9dfeb1383287631ad5c5a676a2558b799db6e9.

There are still some stability problems on some SI boards so bring
this back.

drm/mst: rework payload table allocation to conform better.

The old code has problems with the Dell MST monitors due to some
assumptions I made that weren't true.

I initially thought the Virtual Channel Payload IDs had to be in
the DPCD table in ascending order, however it appears that assumption
is bogus.

The old code also assumed it was possible to insert a member
into the table and it would move other members up, like it does
when you remove table entries, however reality has shown this
isn't true.

So the new code allocates VCPIs separate from entries in the payload
tracking table, and when we remove an entry from the DPCD table,
I shuffle the tracking payload entries around in the struct.

This appears to make VT switch more robust (still not perfect)
with an MST enabled Dell monitor.

Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/ast: Fix HW cursor image

The translation from the X driver to the KMS one typo'ed a couple
of array indices, causing the HW cursor to look weird (blocky with
leaking edge colors). This fixes it.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: stable@vger.kernel.org
Signed-off-by: Dave Airlie <airlied@redhat.com>

Merge tag 'ipu-fixes-3.18' of git://git.pengutronix.de/git/pza/linux into drm-next

IPUv3 fixes for v3.18

* tag 'ipu-fixes-3.18' of git://git.pengutronix.de/git/pza/linux:
  gpu: ipu-v3: Kconfig: Remove SOC_IMX6SL from IMX_IPUV3_CORE Kconfig
  gpu: ipu-v3: ipu-smfc: Do not leave DEBUG defined
  gpu: ipu-v3: Return proper error on ipu_add_client_devices error path
  gpu: ipu-v3: Select GENERIC_IRQ_CHIP to fix build error

Merge tag 'topic/core-stuff-2014-10-03' of git://anongit.freedesktop.org/drm-intel into drm-next

two core fixes

* tag 'topic/core-stuff-2014-10-03' of git://anongit.freedesktop.org/drm-intel:
  drm/core: use helper to check driver features
  drm/edid: Add missing interlaced flag to 576i@100 modes.

Merge tag 'drm-intel-next-fixes-2014-10-03' of git://anongit.freedesktop.org/drm-intel into drm-next

Bunch of fixes for 3.18. Major parts:
- ppgtt fixes (but full ppgtt is for 3.19) from Chris, Michel, ...
- hdmi pixel replication fixes (Clint Taylor)
- leftover i830M patches from Ville
- small things all over

* tag 'drm-intel-next-fixes-2014-10-03' of git://anongit.freedesktop.org/drm-intel: (21 commits)
  drm/i915: Enable pixel replicated modes on BDW and HSW.
  drm/i915: Don't spam dmesg with rps messages on vlv/chv
  drm/i915: Do not leak pages when freeing userptr objects
  drm/i915: Do not store the error pointer for a failed userptr registration
  Revert "drm/i915/bdw: BDW Software Turbo"
  drm/i915/bdw: Cleanup pre prod workarounds
  drm/i915: Use EIO instead of EAGAIN for sink CRC error.
  drm/i915: Extend BIOS stolen mem handling to all platform
  drm/i915: Match GTT space sanity checker with implementation
  drm/i915: HSW always use GGTT selector for secure batches
  drm/i915: add cherryview specfic forcewake in execlists_elsp_write
  drm/i915: fix another use-after-free in i915_gem_evict_everything
  drm/i915: Don't reinit hpd interrupts after gpu reset
  drm/i915: Wrap -EIO send-vblank event for failed pageflip in spinlock
  drm/i915: Drop any active reference before unbinding
  drm/i915: Objects on the unbound list may still have an active reference
  drm/i915/edp: use lane count and link rate from DPCD for eDP
  drm/i915/dp: add missing \n in the TPS3 debug message
  drm/i915/hdmi, dp: Do not dereference the encoder in the connector destroy
  drm/i915: Limit the watermark to at least 8 entries on gen2/3
  ...

Merge branch 'drm-next-3.18' of git://people.freedesktop.org/~agd5f/linux into drm-next

It looks like you missed my last 3.18 pull from 9/24.  This one
includes those patches and a few more on top.  The additional patches are:
- Maarten's radeon fence updates
- Some additional debugging output
From the previous pull request:
- Re-enable some dpm features that were previously disabled due
  to a bug that was fixed in 3.16
- Make some arrays static
- re-arrange some audio code to properly reflect connected status
  in the audio driver

* 'drm-next-3.18' of git://people.freedesktop.org/~agd5f/linux:
  drm/radeon/kv: add uvd/vce info to dpm debugfs output
  drm/radeon/ci: add uvd/vce info to dpm debugfs output
  drm/radeon: export reservation_object from dmabuf to ttm
  drm/radeon: cope with foreign fences inside the reservation object
  drm/radeon: cope with foreign fences inside display
  drm/radeon/cik: write gfx ucode version to ucode addr reg
  drm/radeon/si: print full CS when we hit a packet 0
  drm/radeon: remove unecessary includes
  drm/radeon/combios: declare legacy_connector_convert as static
  drm/radeon/atombios: declare connector convert tables as static
  drm/radeon: drop btc_get_max_clock_from_voltage_dependency_table
  drm/radeon/dpm: drop clk/voltage dependency filters for BTC
  drm/radeon/dpm: drop clk/voltage dependency filters for CI
  drm/radeon/dpm: drop clk/voltage dependency filters for SI
  drm/radeon/dpm: drop clk/voltage dependency filters for NI
  drm/radeon: disable audio when we disable hdmi (v2)
  drm/radeon: split audio enable between eg and r600 (v2)
  drm/radeon: consolidate r600_audio.c into r600_hdmi.c
  drm/radeon: consolidate duplicate encode is digital function
  drm/radeon: fix register name to match internal name

drm/radeon/kv: add uvd/vce info to dpm debugfs output

Track whether UVD or VCE are enabled in debugfs.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/ci: add uvd/vce info to dpm debugfs output

Track whether UVD or VCE are enabled in debugfs.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: export reservation_object from dmabuf to ttm

Adds an extra argument to radeon_bo_create, which is only used in radeon_prime.c.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: cope with foreign fences inside the reservation object

Not the whole world is a radeon! :-)

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: cope with foreign fences inside display

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/core: use helper to check driver features

The patch replaces direct access to driver_features field
by calls to helper function.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/radeon/cik: write gfx ucode version to ucode addr reg

Helpful for debugging as the version shows up in a
register dump.

Cc: Jay Cornwall <jay.cornwall@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/si: print full CS when we hit a packet 0

We should never have a packet 0 in the command stream.
Dump the full command stream to help debug.

bug:
https://bugs.freedesktop.org/show_bug.cgi?id=84500

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: remove unecessary includes

The requirement for these was removed in:
0a168933976eb483da91161316bbbbcb74d00486

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/combios: declare legacy_connector_convert as static

legacy_connector_convert is used in radeon_combios.c only, so declare it as static.

Signed-off-by: Michele Curti <michele.curti@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/atombios: declare connector convert tables as static

The tables:
 * supported_devices_connector_convert
 * supported_devices_connector_object_id_convert
 * object_connector_convert
are used in redeon_atombios.c only, so declare them as static.

Signed-off-by: Michele Curti <michele.curti@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: drop btc_get_max_clock_from_voltage_dependency_table

It's no longer used now that the underlying bugs are fixed.

Reviewed-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Tested-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/dpm: drop clk/voltage dependency filters for BTC

No longer needed now that the underlying bug was fixed in
e07929810f0a19ddd756558290c7d72827cbfcd9
(drm/radeon/dpm: fix typo in vddci setup for eg/btc).

bug:
https://bugs.freedesktop.org/show_bug.cgi?id=69721

Reviewed-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/dpm: drop clk/voltage dependency filters for CI

Not sure this was ever necessary for CI, was just done
to be on the safe side.

bug:
https://bugs.freedesktop.org/show_bug.cgi?id=69721

Reviewed-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/dpm: drop clk/voltage dependency filters for SI

Not sure this was ever necessary for SI, was just done
to be on the safe side.

bug:
https://bugs.freedesktop.org/show_bug.cgi?id=69721

Reviewed-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/dpm: drop clk/voltage dependency filters for NI

No longer needed now that the underlying bug was fixed in
b0880e87c1fd038b84498944f52e52c3e86ebe59
(drm/radeon/dpm: fix vddci setup typo on cayman).

bug:
https://bugs.freedesktop.org/show_bug.cgi?id=69721

Reviewed-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Tested-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: disable audio when we disable hdmi (v2)

This should allow the audio driver to get a better
idea of whether the sink is connected or not.

v2: fix copy/paste typo noticed by David Henningsson

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: split audio enable between eg and r600 (v2)

Clean up the enable sequence as well.

V2: clean up duplicate defines

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: consolidate r600_audio.c into r600_hdmi.c

Most of that functionality is only used by r600_hdmi.c
and I'm planning to change that further.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: consolidate duplicate encode is digital function

Only need one copy.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: fix register name to match internal name

no functional change.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/edid: Add missing interlaced flag to 576i@100 modes.

CEA VICs 44 and 45 were missing DRM_MODE_FLAG_INTERLACE.

Signed-off-by: Clint Taylor <clinton.a.taylor@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Merge branch 'for-airlied-next' of git://people.freedesktop.org/~mlankhorst/linux into drm-next

fixups for nouveau and fencing

* 'for-airlied-next' of git://people.freedesktop.org/~mlankhorst/linux:
  drm/nouveau: export reservation_object from dmabuf to ttm
  drm/ttm: add reservation_object as argument to ttm_bo_init
  drm: Pass dma-buf as argument to gem_prime_import_sg_table
  drm/nouveau: assign fence_chan->name correctly
  drm/nouveau: specify if interruptible wait is desired in nouveau_fence_sync
  drm/nouveau: bump driver patchlevel to 1.2.1

Merge tag 'topic/core-stuff-2014-09-29' of git://anongit.freedesktop.org/drm-intel into drm-next

Ok, here's the update core-stuff pull request with the locking fixup patch
fixed up with another patch.

* tag 'topic/core-stuff-2014-09-29' of git://anongit.freedesktop.org/drm-intel:
  drm: Drop grab fpriv->fbs_lock in drm_fb_release
  drm/udl: use container_of to resolve udl_fbdev from drm_fb_helper
  drm/ast: use container_of to resolve ast_fbdev from drm_fb_helper
  drm/gma500: use container_of to resolve psb_fbdev from drm_fb_helper
  drm/qxl: use container_of to resolve qxl_fbdev from drm_fb_helper
  drm/nouveau: use container_of to resolve nouveau_plane from drm_plane
  drm/nouveau: use container_of to resolve nouveau_fbdev from drm_fb_helper
  drm/radeon: use container_of to resolve radeon_fbdev from drm_fb_helper
  drm/mgag200: use container_of to resolve mga_fbdev from drm_fb_helper
  drm/cirrus: use container_of to resolve cirrus_fbdev from drm_fb_helper
  drm: Improve debug output for drm_wait_one_vblank
  drm: Fixup locking for universal cursor planes
  drm: Don't update vblank timestamp when the counter didn't change

drm/nouveau: export reservation_object from dmabuf to ttm

Adds an extra argument to nouveau_bo_new, which is only used in nouveau_prime.c.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>

drm/i915: Enable pixel replicated modes on BDW and HSW.

Haswell and later silicon has added a new pixel replication register
to the pipe timings for each transcoder. Now in addition to the
DPLL_A_MD register for the pixel clock double, we also need to write
to the TRANS_MULT_n (0x6002c) register to double the pixel data. Writing
to the DPLL only double the pixel clock.

ver2: Macro name change from MULTIPLY to PIPE_MULTI. (Daniel)
ver3: Do not set pixel multiplier if transcoder is eDP (Ville)
ver4: Macro name change to PIPE_MULT and default else pixel_multiplier

Cc: Ville =?iso-8859-1?Q?Syrj=E4l=E4?= <ville.syrjala@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Clint Taylor <clinton.a.taylor@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
[danvet: Appease checkpatch and move one hunk back into the right
place that git am misplace!?]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/ttm: add reservation_object as argument to ttm_bo_init

This allows importing reservation objects from dma-bufs.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>

drm: Pass dma-buf as argument to gem_prime_import_sg_table

Allows importing dma_reservation_objects from a dma-buf.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>

drm/nouveau: assign fence_chan->name correctly

Make nouveau_fence_chan refcounted, to make trace_fence_destroy
always return the correct name without a race condition.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>

drm/nouveau: specify if interruptible wait is desired in nouveau_fence_sync

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>

drm/nouveau: bump driver patchlevel to 1.2.1

Allows userspace to detect shared fences are supported.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>

drm/nouveau: nv84+: fix fence context seqno's

This fixes a regression introduced by "drm/nouveau: rework to new fence interface"
(commit 29ba89b2371d466).

The fence sequence should not be reset after creation, the old value is used instead.
On destruction the final value is written, to prevent another source of accidental
wraparound in case of a channel being destroyed after a hang, and unblocking any other
channel that may wait on the about-to-be-deleted channel to signal.

I'm nothing if not optimistic about any hope of recovery from that. ;-)

Reported-by: Ted Percival <ted@tedp.id.au>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Tested-by: Ted Percival <ted@tedp.id.au>
Acked-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/i915: Don't spam dmesg with rps messages on vlv/chv

If the GPU frequency isn't going to change don't spam dmesg with
debug messages about it.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

drm/i915: Do not leak pages when freeing userptr objects

sg_alloc_table_from_pages() can build us a table with coalesced ranges which
means we need to iterate over pages and not sg table entries when releasing
page references.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: "Barbalho, Rafael" <rafael.barbalho@intel.com>
Tested-by: Rafael Barbalho <rafael.barbalho@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@vger.kernel.org
[danvet: Remove unused local variable sg.]
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

drm/i915: Do not store the error pointer for a failed userptr registration

If we fail to create our mmu notification, we report the error back and
currently store the error inside the i915_mm_struct. This not only causes
subsequent registerations of the same mm to fail (an issue if the first
was interrupted by a signal and needed to be restarted) but also causes
us to eventually try and free the error pointer.

[   73.419599] BUG: unable to handle kernel NULL pointer dereference at 000000000000004c
[   73.419831] IP: [<ffffffff8114af33>] mmu_notifier_unregister+0x23/0x130
[   73.420065] PGD 8650c067 PUD 870bb067 PMD 0
[   73.420319] Oops: 0000 [#1] SMP DEBUG_PAGEALLOC
[   73.420580] CPU: 0 PID: 42 Comm: kworker/0:1 Tainted: G        W      3.17.0-rc6+ #1561
[   73.420837] Hardware name: Intel Corporation SandyBridge Platform/LosLunas CRB, BIOS ASNBCPT1.86C.0075.P00.1106281639 06/28/2011
[   73.421405] Workqueue: events __i915_mm_struct_free__worker
[   73.421724] task: ffff880088a81220 ti: ffff880088168000 task.ti: ffff880088168000
[   73.422051] RIP: 0010:[<ffffffff8114af33>]  [<ffffffff8114af33>] mmu_notifier_unregister+0x23/0x130
[   73.422410] RSP: 0018:ffff88008816bd50  EFLAGS: 00010286
[   73.422765] RAX: 0000000000000003 RBX: ffff880086485400 RCX: 0000000000000000
[   73.423137] RDX: ffff88016d80ee90 RSI: ffff880086485400 RDI: 0000000000000044
[   73.423513] RBP: ffff88008816bd70 R08: 0000000000000001 R09: 0000000000000000
[   73.423895] R10: 0000000000000320 R11: 0000000000000001 R12: 0000000000000044
[   73.424282] R13: ffff880166e5f008 R14: ffff88016d815200 R15: ffff880166e5f040
[   73.424682] FS:  0000000000000000(0000) GS:ffff88016d800000(0000) knlGS:0000000000000000
[   73.425099] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   73.425537] CR2: 000000000000004c CR3: 0000000087f5f000 CR4: 00000000000407f0
[   73.426157] Stack:
[   73.426597]  ffff880088a81248 ffff880166e5f038 fffffffffffffffc ffff880166e5f008
[   73.427096]  ffff88008816bd98 ffffffff814a75f2 ffff880166e5f038 ffff8800880f8a28
[   73.427603]  ffff88016d812ac0 ffff88008816be00 ffffffff8106321a ffffffff810631af
[   73.428119] Call Trace:
[   73.428606]  [<ffffffff814a75f2>] __i915_mm_struct_free__worker+0x42/0x80
[   73.429116]  [<ffffffff8106321a>] process_one_work+0x1ba/0x610
[   73.429632]  [<ffffffff810631af>] ? process_one_work+0x14f/0x610
[   73.430153]  [<ffffffff810636db>] worker_thread+0x6b/0x4a0
[   73.430671]  [<ffffffff8108d67d>] ? trace_hardirqs_on+0xd/0x10
[   73.431501]  [<ffffffff81063670>] ? process_one_work+0x610/0x610
[   73.432030]  [<ffffffff8106a206>] kthread+0xf6/0x110
[   73.432561]  [<ffffffff8106a110>] ? __kthread_parkme+0x80/0x80
[   73.433100]  [<ffffffff8169c22c>] ret_from_fork+0x7c/0xb0
[   73.433644]  [<ffffffff8106a110>] ? __kthread_parkme+0x80/0x80
[   73.434194] Code: 0f 1f 84 00 00 00 00 00 66 66 66 66 90 8b 46 4c 85 c0 0f 8e 10 01 00 00 55 48 89 e5 41 55 41 54 53 48 89 f3 49 89 fc 48 83 ec 08 <48> 83 7f 08 00 0f 84 b1 00 00 00 48 c7 c7 40 e6 ac 82 e8 26 65
[   73.435942] RIP  [<ffffffff8114af33>] mmu_notifier_unregister+0x23/0x130
[   73.437017]  RSP <ffff88008816bd50>
[   73.437704] CR2: 000000000000004c

Fixes regression from commit ad46cb533d586fdb256855437af876617c6cf609
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Thu Aug 7 14:20:40 2014 +0100

    drm/i915: Prevent recursive deadlock on releasing a busy userptr

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=84207
Testcase: igt/gem_render_copy_redux
Testcase: igt/gem_userptr_blits/create-destroy-sync
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jacek Danecki <jacek.danecki@intel.com>
Cc: "Gong, Zhipeng" <zhipeng.gong@intel.com>
Cc: Jacek Danecki <jacek.danecki@intel.com>
Cc: "Ursulin, Tvrtko" <tvrtko.ursulin@intel.com>
Cc: stable@vger.kernel.org
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

Revert "drm/i915/bdw: BDW Software Turbo"

This reverts commit c76bb61a71083b2d90504cc6d0dda2047c5d63ca.

It's apparently too broken so that Rodrigo submitted a patch to add a
config option for it. Given that the design is also ... suboptimal and
that I've only merged this to get lead engineers and managers off my
back for one second let's just revert this.

/me puts on combat gear again

It was worth a shot ...

References: http://mid.mail-archive.com/1411686380-1953-1-git-send-email-rodrigo.vivi@intel.com
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Daisy Sun <daisy.sun@intel.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

drm: Drop grab fpriv->fbs_lock in drm_fb_release

Paulo Zanoni reported a lockdep splat with a locking inversion between
fpriv->fbs_lock and the modeset locks. This issue was introduced in

commit f2b50c1161590c3bcdbf3455fe4c575f1c1bd293
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Fri Sep 12 17:07:32 2014 +0200

    drm: Fixup locking for universal cursor planes

This here is actually one of the rare cases where lockdep hits a false
positive: The deadlock only happens in drm_fb_release, which cleans up
the file private structure when all the references are gone. So the
locking is the very last one and no one else can deadlock. It also
doesn't protect anything at all, since all ioctls are guaranteed to
have returned at this point - otherwise they'd still hold a reference
on the file.

So let's just drop it and replace it with a big comment.

Cc: David Herrmann <dh.herrmann@gmail.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Paulo Zanoni <przanoni@gmail.com>
Reported-and-Tested-by: Paulo Zanoni <przanoni@gmail.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

gpu: ipu-v3: Kconfig: Remove SOC_IMX6SL from IMX_IPUV3_CORE Kconfig

SOC_IMX6SL does not have the IPU block, so remove it from the Kconfig entry.

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>

drm/udl: use container_of to resolve udl_fbdev from drm_fb_helper

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/ast: use container_of to resolve ast_fbdev from drm_fb_helper

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/gma500: use container_of to resolve psb_fbdev from drm_fb_helper

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/qxl: use container_of to resolve qxl_fbdev from drm_fb_helper

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/nouveau: use container_of to resolve nouveau_plane from drm_plane

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/nouveau: use container_of to resolve nouveau_fbdev from drm_fb_helper

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/radeon: use container_of to resolve radeon_fbdev from drm_fb_helper

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/mgag200: use container_of to resolve mga_fbdev from drm_fb_helper

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/cirrus: use container_of to resolve cirrus_fbdev from drm_fb_helper

Use container_of instead of casting first structure member.

Signed-off-by: Fabian Frederick <fabf@skynet.be>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm: Improve debug output for drm_wait_one_vblank

This replicates what we've done in i915 in

commit 31e4b89acbd7b19c9a8557e6e660a583a0b97daa
Author: Damien Lespiau <damien.lespiau@intel.com>
Date:   Mon Aug 18 13:51:00 2014 +0100

    drm/i915: Print the pipe on which the vblank wait times out

to make sure that when we switch i915 to drm_wait_one_vblank that the
debug output doesn't regress.

Cc: Damien Lespiau <damien.lespiau@intel.com>
Cc: Thomas Wood <thomas.wood@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

drm: Fixup locking for universal cursor planes

Bunch of things amiss:
- Updating crtc->cursor_x/y was done without any locking. Spotted by
  David Herrmann.
- Dereferencing crtc->cursor->fb was using the wrong lock, should take
  the crtc lock.
- Grabbing _all_ modeset locks torpedoes the reason why we added
  fine-grained locks originally: Cursor updates shouldn't stall on
  background stuff like probing outputs.

Best is to just grab the crtc lock around everything and drop all the
other locking. The only issue is that we can't switch planes between
crtcs with that, so make sure that never happens when someone uses
universal plane helpers. This shouldn't be a possible regression ever
since legacy ioctls also only grabbed the crtc lock, so switching
crtcs was never possible for the underlying plane object. And i915
(the only user of universal cursors thus far) has fixed cursor->crtc
links.

Cc: David Herrmann <dh.herrmann@gmail.com>
Cc: Pallavi G<pallavi.g@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Tested-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

drm: Don't update vblank timestamp when the counter didn't change

If we already have a timestamp for the current vblank counter, don't
update it with a new timestmap. Small errors can creep in between two
timestamp queries for the same vblank count, which could be confusing to
userspace when it queries the timestamp for the same vblank sequence
number twice.

This problem gets exposed when the vblank disable timer is not used
(or is set to expire quickly) and thus we can get multiple vblank
disable<->enable transition during the same frame which would all
attempt to update the timestamp with the latest estimate.

Testcase: igt/kms_flip/flip-vs-expired-vblank
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Mario Kleiner <mario.kleiner.de@gmail.com>
v2:Mario: Trivial rebase on top of current drm-next (13-Sep-2014)
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915/bdw: Cleanup pre prod workarounds

as these have been fixed in production hw and hurt performance
if applied.

v2: adjust requested ring space (Ville)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=83482
Tested-by: zhoujian <jianx.zhou@intel.com>
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/doc: Fixup drm_irq kerneldoc includes.

Only !P can be used together with a function list.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: Extract <drm/drm_gem.h>

v2: Don't forget git add, noticed by David.

Cc: David Herrmann <dh.herrmann@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Acked-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: Move internal debugfs functions to drm_internal.h

In my header cleanup I've missed the debugfs functions completely.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: Move leftover ioctl declarations to drm_internal.h

Somehow I've missed these three, fix this up asap. Plus move
drm_master_create since while at it.

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: Move drm_vm_open_locked into drm_internal.h

Leftover from my previous header cleanup.

This depends upon the patch to rework exynos mmap support, otherwise
it'll break exynos.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: move drm_mmap to <drm/drm_legacy.h>

Now that we've removed the copypasted users in gem/ttm we can
relegate the legacy buffer mapping support to where it belongs.
Also give it the proper drm_legacy_ prefix.

While at it statify drm_mmap_locked, somehow I've missed that in my
previous header rework.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/gem: Don't call drm_mmap from drm_gem_mmap

The only user I could dig out was i915 back when ums+gem was still a
thing. But we've just very much killed that, and even when someone
screams about that we should resurrect that with a special hack
(wrapping drm_gem_mmap) in i915, not in the core code.

So good riddance to another entry point of the legacy buffer mapping
code.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/<ttm-based-drivers>: Don't call drm_mmap

Really, the legacy buffer api should be dead, especially for all these
newfangled drivers. I suspect this is copypasta from the transitioning
days, which probably originated in radeon.

Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Herrmann <dh.herrmann@gmail.com>
Cc: Rashika <rashika.kheria@gmail.com>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Fabian Frederick <fabf@skynet.be>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Alexandre Courbot <acourbot@nvidia.com>
Cc: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Cc: Christian Engelmayer <cengelma@gmx.at>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Reviewed-by: David Herrmann <dh.herrmann@gmail.com>
Acked-by: Ben Skeggs <bskeggs@redhat.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: change drm_err return type to void

The return value is not used by callers of this function
nor by uses of the DRM_ERROR macro so change the function
to return void.

Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/ttm: Clean usage of ttm_io_prot() with TTM_PL_FLAG_CACHED

Today, most callers of ttm_io_prot() check TTM_PL_FLAG_CACHED before
calling it since on some archs it will unconditionally create non-cached
mappings.

But not all callers do which is incorrect as far as I can tell.

Instead, move that check inside ttm_io_port() itself for all archs
and make powerpc use the same implementation as ia64 and arm

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: powerpc can use a simpler drm_io_prot()

What the code does is equivalent to the x86 code, so let's use
it as well

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>

Merge branch 'exynos-drm-next' of git://git./linux/kernel/git/daeinki/drm-exynos into drm-next

   Sorry for late. This pull request includes some enhancements
   for Exynos drm, new feature supports, cleanups and fixups
   like below,

   - Consider low power transmission for drm mipi dsi module,
     and also add non-continuous clock mode support for Exynos
     mipi dsi driver.
   - Add Exynos3250 SoC support.
   - Enhance and clean up ipp framework and fimc driver.
   - Update to use component match support and fix up
     de-initialization order.
   - Remove a direct mmap interface and relevant stuff specific to
     Exynos drm, use drm generic mmap interface instead.
     And we will remove the specific interface from userspace
     library, libdrm soon.
   - Use universal plane which allows to replace fake primary plane
     with the real one.
   - Some code cleanups and fixups.

* 'exynos-drm-next' of git://git.kernel.org/pub/scm/linux/kernel/git/daeinki/drm-exynos: (40 commits)
  drm/exynos: switch to universal plane API
  drm/exynos: use drm generic mmap interface
  drm/exynos: remove DRM_EXYNOS_GEM_MAP_OFFSET ioctl
  drm/exynos: factor out initial setting of each driver
  drm/exynos/hdmi: unregister connector on removal
  drm/exynos/dp: unregister connector on removal
  drm/exynos/dpi: unregister connector and panel on removal
  drm/exynos/dsi: unregister connector on removal
  drm/exynos/fb: free exynos framebuffer on error
  drm/exynos/fbdev: fix fbdev gem object cleanup
  drm/exynos: fix drm driver de-initialization order
  drm/exynos/ipp: traverse ipp drivers list safely
  drm/exynos: update to use component match support
  drm/exynos/ipp: add file checks for ioctls
  drm/exynos/ipp: remove file argument from node related functions
  drm/exynos/fimc: fix source buffer registers
  drm/exynos/fimc: simplify buffer queuing
  drm/exynos/fimc: do not enable fimc twice
  drm/exynos/fimc: avoid clearing overflow bits
  drm/exynos/ipp: remove events during command cleaning
  ...

drm/exynos: switch to universal plane API

The patch replaces legacy functions
drm_plane_init() / drm_crtc_init() with
drm_universal_plane_init() and drm_crtc_init_with_planes().
It allows to replace fake primary plane with the real one.
Additionally the patch leaves cleanup of crtcs to core,
this way planes and crtcs are cleaned in correct order.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos: use drm generic mmap interface

This patch removes DRM_EXYNOS_GEM_MMAP ictrl feature specific
to Exynos drm and instead uses drm generic mmap.

We had used the interface specific to Exynos drm to do mmap directly,
not to use demand paging which maps each page with physical memory
at page fault handler. We don't need the specific mmap interface
because the drm generic mmap which uses vm offset manager stuff can
also do mmap directly.

This patch makes a userspace region to be mapped with whole physical
memory region allocated by userspace request when mmap system call is
requested.

Changelog v2:
- do not set VM_IO, VM_DONTEXPEND and VM_DONTDUMP. These flags were already
  set by drm_gem_mmap
- do not include <linux/anon_inodes.h>, which isn't needed anymore.

Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos: remove DRM_EXYNOS_GEM_MAP_OFFSET ioctl

This interface and relevant codes aren't used anymore.

Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos: factor out initial setting of each driver

From fimd driver and vidi driver, dev->irq_enabled and
dev->vblank_disable_allowed are set and also mixer needs them even if
missed. It's duplicated so set them when loads drm driver.

Signed-off-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/hdmi: unregister connector on removal

During component removal driver should unregister connector.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/dp: unregister connector on removal

During component removal driver should unregister connector.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/dpi: unregister connector and panel on removal

During component removal it should unregister connector
and optionally detach the panel.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/dsi: unregister connector on removal

During component unbind connector should be unregistered.
Also DSI host should be unregistered after KMS cleanup.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/fb: free exynos framebuffer on error

In case drm_framebuffer_init fails exynos_fb should be freed
before returning an error.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/fbdev: fix fbdev gem object cleanup

exynos_gem_obj is used by exynos_drm_fbdev_destroy so it cannot be destroyed
before calling the latter. exynos_gem_obj will be destroyed anyway by
exynos_drm_fbdev_destroy->...->exynos_drm_fb_destroy.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos: fix drm driver de-initialization order

Since components have their own cleanup routines calling
drm_mode_config_cleanup before component_unbind_all causes errors
due to double free of KMS objects. The patch fixes it by changing
de-initialization order. Now it is exactly opposite to init order.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: traverse ipp drivers list safely

On ipp subsystem removal list of ipp drivers is traversed
and their members are deleted. To do it properly safe version
of list_for_each* should be used.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos: update to use component match support

Update Exynos's DRM driver to use component match support rater than
add_components.

Changelog v2:
- release devices and drivers if failed.
- change compare_of to compare_dev.

Signed-off-by: Inki Dae <inki.dae@samsung.com>
Tested-by: Andrzej Hajda <a.hajda@samsung.com>

drm/exynos/ipp: add file checks for ioctls

Process should not have access to ipp nodes created by another
process. The patch adds necessary checks.
It also simplifies lookup for command node.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: remove file argument from node related functions

Since file pointer is preserved in c_node passing it
as argument in node functions is redundant.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/fimc: fix source buffer registers

FIMC in default mode of operation uses only one input buffer,
but the driver used also second buffer, as a result only the
first frame was processed correctly. The patch fixes it.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/fimc: simplify buffer queuing

The patch removes redundant checks, redundant HW reads
and simplifies code.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/fimc: do not enable fimc twice

The patch removes redundant H/W activation.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/fimc: avoid clearing overflow bits

Overflow bits shall be cleared by H/W.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: remove events during command cleaning

Events were removed only during stop command, as a result
there were memory leaks if program prematurely exited.
This patch fixes it.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: stop hardware before freeing memory

Memory shouldn't be freed when hardware is still running.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: replace work_struct casting with better constructs

Type casting should be avoided if possible. In case of
work_struct it can be simply replaced by reference to member field.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: clean memory nodes on command node cleaning

The nodes should be removed before removing command node.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: move nodes cleaning to separate function

The patch introduces ipp_clean_mem_nodes function which replaces
redundant code. Additionally memory node function definitions
are moved up to increase its visibility.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: free partially allocated resources on error

In case of allocation errors some already allocated buffers
were not freed. The patch fixes it.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: remove unused field in command node

Since command node have file pointer dev field became useless.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: remove only related commands on file close

On file close driver should remove only command nodes created
via this file.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: move file reference from memory to command node

Command node should contain file reference to distinguish commands
created by different processes.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>

drm/exynos/ipp: cancel works before command node clean

All pending works should be canceled prior to its removal.

Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
Reviewed-by: Joonyoung Shim <jy0922.shim@samsung.com>
Signed-off-by: Inki Dae <inki.dae@samsung.com>