Merge branch 'drm-next-4.9' of git://people.freedesktop.org/~agd5f/linux into drm-next

First drm-next pull for radeon and amdgpu for 4.9.  Highlights:
- powerplay support for iceland asics
- improved GPU reset (both full asic and per block)
- UVD and VCE powergating for CZ and ST
- VCE clockgating for CZ and ST
- Support for pre-initialized (e.g., zeroed) vram buffers
- ttm cleanups
- virtual display support
- core and radeon/amdgpu support for page_flip_target
- lots of bug fixes and clean ups

* 'drm-next-4.9' of git://people.freedesktop.org/~agd5f/linux: (171 commits)
  drm/amdgpu: use memcpy_toio for VCE firmware upload
  drm/amdgpu: use memcpy_to/fromio for UVD fw upload
  drm/amd/powerplay: delete useless code in iceland_hwmgr.c.
  drm/radeon: switch UVD code to use UVD_NO_OP for padding
  drm/amdgpu: switch UVD code to use UVD_NO_OP for padding
  drm/radeon: add support for UVD_NO_OP register
  drm/amdgpu: add support for UVD_NO_OP register
  drm/amdgpu: fix VCE ib alignment value
  drm/amdgpu: fix IB alignment for UVD
  drm/amd/amdgpu: Print ring name in amdgpu_ib_schedule()
  drm/radeon: remove dead code, si_mc_load_microcode (v2)
  drm/radeon/cik: remove dead code (v2)
  drm/amd/powerplay: avoid NULL dereference, cz_hwmgr.c
  drm/amd/powerplay: avoid NULL pointer dereference
  drm/amdgpu/gmc8: remove dead code (v2)
  drm/amdgpu/gmc7: remove dead code (v2)
  drm/amdgpu: Fix indentation in dce_v8_0_audio_write_sad_regs()
  drm/amdgpu: Use correct mask in dce_v8_0_afmt_setmode() and fix comment typos.
  drm/amdgpu: cleanup amdgpu_vm_bo_update params
  drm/amdgpu: stop adding dummy entry in amdgpu_ttm_placement_init
  ...

Merge branch 'drm-intel-next' of git://anongit.freedesktop.org/drm-intel into drm-next

drm-intel-next-2016-08-22:
- bugfixes and cleanups for rcu-protected requests (Chris)
- atomic modeset fixes for gpu reset on pre-g4x (Maarten&Ville)
- guc submission improvements (Dave Gordon)
- panel power sequence cleanup (Imre)
- better use of stolen and unmappable ggtt (Chris), plus prep work to make that
  happen
- rework of framebuffer offsets, prep for multi-plane framebuffers (Ville)
- fully partial ggtt vmaps, including fenced ones (Chris)
- move lots more of the gem tracking from the object to the vma (Chris)
- tune the command parser (Chris)
- allow fbc without fences on recent platforms (Chris)
- fbc frontbuffer tracking fixes (Chris)
- fast prefaulting using io-mappping.h pgprot caching (Chris)

* 'drm-intel-next' of git://anongit.freedesktop.org/drm-intel: (141 commits)
  io-mapping: Fixup for different names of writecombine
  io-mapping.h: s/PAGE_KERNEL_IO/PAGE_KERNEL/
  drm/i915: Update DRIVER_DATE to 20160822
  drm/i915: Use remap_io_mapping() to prefault all PTE in a single pass
  drm/i915: Embed the io-mapping struct inside drm_i915_private
  io-mapping: Always create a struct to hold metadata about the io-mapping
  drm/i915/fbc: Allow on unfenced surfaces, for recent gen
  drm/i915/fbc: Don't set an illegal fence if unfenced
  drm/i915: Flush delayed fence releases after reset
  drm/i915: Reattach comment, complete type specification
  drm/i915/cmdparser: Accelerate copies from WC memory
  drm/i915/cmdparser: Use binary search for faster register lookup
  drm/i915/cmdparser: Check for SKIP descriptors first
  drm/i915/cmdparser: Compare against the previous command descriptor
  drm/i915/cmdparser: Improve hash function
  drm/i915/cmdparser: Only cache the dst vmap
  drm/i915/cmdparser: Use cached vmappings
  drm/i915/cmdparser: Add the TIMESTAMP register for the other engines
  drm/i915/cmdparser: Make initialisation failure non-fatal
  drm/i915: Stop discarding GTT cache-domain on unbind vma
  ...

Merge branch 'for-next' of git://people.freedesktop.org/~seanpaul/dogwood into drm-next

This pull request contains the following rockchip drm changes:

  - Introduce support for rk3399 vop/crtc
  - Add PSR framework to the rockchip driver
  - Implement PSR in the rockchip analogix edp driver
  - Fix panel on/off in analogix to avoid damaging panels
  - Some miscellaneous fixes to clean up logs and code readability

* 'for-next' of git://people.freedesktop.org/~seanpaul/dogwood:
  drm/rockchip: analogix_dp: drop unnecessary probe deferral "error" print
  drm/rockchip: Enable vblank without event
  drm/rockchip: Improve analogix-dp psr handling
  drm/rockchip: A couple small fixes to psr
  drm/rockchip: Use a spinlock to protect psr state
  drm/rockchip: Don't use a delayed worker for psr state changes
  drm/rockchip: Convert psr_list_mutex to spinlock and use it
  drm/rockchip: analogix_dp: implement PSR function
  drm/bridge: analogix_dp: add the PSR function support
  drm/rockchip: add an common abstracted PSR driver
  drm/rockchip: vop: export line flag function
  drm/bridge: analogix_dp: Ensure the panel is properly prepared/unprepared
  dt-bindings: add compatible strings for big/little rockchip vops
  dt-bindings: sort Rockchip vop compatible by chip's number
  drm/rockchip: vop: add rk3399 vop support
  drm/rockchip: vop: introduce VOP_REG_MASK
  drm/rockchip: sort registers define by chip's number

drm/amdgpu: use memcpy_toio for VCE firmware upload

Try to be clean here, even when it's a noop on x86.

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

drm/amdgpu: use memcpy_to/fromio for UVD fw upload

Also use the firmware size from the header instead of calculating it.

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

drm/amd/powerplay: delete useless code in iceland_hwmgr.c.

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

drm/radeon: switch UVD code to use UVD_NO_OP for padding

Replace packet2's with packet0 writes to UVD_NO_OP.  The
value written to UVD_NO_OP does not matter.

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

drm/amdgpu: switch UVD code to use UVD_NO_OP for padding

Replace packet2's with packet0 writes to UVD_NO_OP.  The
value written to UVD_NO_OP does not matter.

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

drm/radeon: add support for UVD_NO_OP register

Writes to this register are the preferred way to do NOPs.

Bump the driver version as well.

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

drm/amdgpu: add support for UVD_NO_OP register

Writes to this register are the preferred way to do NOPs.

Bump the driver version as well.

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

drm/amdgpu: fix VCE ib alignment value

The VCE rings only require single dword alignment.

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

drm/amdgpu: fix IB alignment for UVD

According to the hw team, it should be 16, not 8.

Cc: Peter Fang <peter.fang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

io-mapping: Fixup for different names of writecombine

Somehow architectures can't agree on this. And for good measure make
sure we have a fallback which should work everywhere (fingers
crossed).

v2: Make it compile properly, needs a defined() for the #elif.

Fixes: ac96b5566926 ("io-mapping.h: s/PAGE_KERNEL_IO/PAGE_KERNEL/")
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: linux-mm@kvack.org
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160823202233.4681-1-daniel.vetter@ffwll.ch
(cherry picked from commit 80c33624e4723c4e22d9917cd676067ebf652dc2)

io-mapping.h: s/PAGE_KERNEL_IO/PAGE_KERNEL/

PAGE_KERNEL_IO is an x86-ism. Though it is used to define the pgprot_t
used for the iomapped region, it itself is just PAGE_KERNEL. On all
other arches, PAGE_KERNEL_IO is undefined so in a general header we must
refrain from using it.

v2: include pgtable for pgprot_combine()

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Fixes: cafaf14a5d8f ("io-mapping: Always create a struct to hold metadata about the io-mapping")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: linux-mm@kvack.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20160823155024.22379-1-chris@chris-wilson.co.uk
(cherry picked from commit ac96b5566926af83463ddcf4655856033c092f26)

drm/doc: Document uapi requirements in DRM

Everyone knows them, except all the new folks joining from the ARM
side haven't lived through all the pain of the past years and are
entirely surprised when I raise this. Definitely time to document
this.

Last time this was a big discussion was about 6 years ago, when qcom
tried to land a kernel driver without userspace. Dave Airlie made the
rules really clear:

http://airlied.livejournal.com/73115.html

This write-up here is essentially what I've put into a presentation a
while ago, which was also reviewed by Dave:

http://blog.ffwll.ch/2015/05/gfx-kernel-upstreaming-requirements.html

v2: Fix typos Eric&Rob spotted.

v3: Nitpick from Jani.

Cc: Dave Airlie <airlied@gmail.com>
Cc: Oded Gabbay <oded.gabbay@gmail.com>
Cc: Russell King <rmk+kernel@armlinux.org.uk>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Cc: Eric Anholt <eric@anholt.net>
Cc: Thomas Hellstrom <thellstrom@vmware.com>
Cc: Sinclair Yeh <syeh@vmware.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Cc: Benjamin Gaignard <benjamin.gaignard@linaro.org>
Cc: Mark Yao <mark.yao@rock-chips.com>
Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: Rob Clark <robdclark@gmail.com>
Cc: CK Hu <ck.hu@mediatek.com>
Cc: Xinliang Liu <z.liuxinliang@hisilicon.com>
Cc: Philipp Zabel <p.zabel@pengutronix.de>
Cc: Stefan Agner <stefan@agner.ch>
Cc: Inki Dae <inki.dae@samsung.com>
Cc: Maxime Ripard <maxime.ripard@free-electrons.com>
Cc: Boris Brezillon <boris.brezillon@free-electrons.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Brian Starkey <brian.starkey@arm.com>
Cc: Liviu Dudau <liviu.dudau@arm.com>
Cc: Alexey Brodkin <abrodkin@synopsys.com>
Acked-by: Dave Airlie <airlied@gmail.com>
Reviewed-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

Merge tag 'topic/drm-misc-2016-08-23' of git://anongit.freedesktop.org/drm-intel into drm-next

A few bigger things:
- start of splitting drm_crtc.c into more manageable and better documneted
  chunks
- DRM_DEV_* logging (Sean)

* tag 'topic/drm-misc-2016-08-23' of git://anongit.freedesktop.org/drm-intel: (46 commits)
  drm/fb-helper: Make docs for fb_set_suspend wrappers consistent
  drm/rockchip: Delete unnecessary assignment for the field "owner"
  drm/bridge: dw-hdmi: Delete unnecessary assignment for the field "owner"
  drm/rockchip: Don't continue trying to enable crtc on failure
  drm/fb-helper: Add drm_fb_helper_set_suspend_unlocked()
  drm/fb-helper: Fix the dummy remove_conflicting_framebuffers
  drm/udl: Ensure channel is selected before using the device.
  drm: Avoid calling dev_printk(.dev = NULL)
  drm: avoid exposing kernel stack in compat_drm_getstats
  reservation: fix small comment typo
  drm: Don't implement empty prepare_fb()/cleanup_fb()
  virtio-gpu: Use memdup_user() rather than duplicating its implementation
  GPU-DRM-Savage: Use memdup_user() rather than duplicating
  drm: Allow drivers to modify plane_state in prepare_fb/cleanup_fb
  drm/rockchip: Use DRM_DEV_ERROR in vop
  drm: Introduce DRM_DEV_* log messages
  drm/edid: CEA mode 64 1080p100 vsync pulse width incorrect
  Revert "drm/hisilicon: Don't set drm_device->platformdev"
  dma-buf: fix kernel-doc warning and typos
  drm: Fix kerneldoc in drm_plane_helper.c
  ...

drm/fb-helper: Make docs for fb_set_suspend wrappers consistent

I figured I might as well go ocd and make them booleans and rename the
locked version too.

v2: Review from Noralf.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Cc: Noralf Trønnes <noralf@tronnes.org>
Fixes: cfe63423d9be ("drm/fb-helper: Add drm_fb_helper_set_suspend_unlocked()")
Acked-by: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160823152727.31788-1-daniel.vetter@ffwll.ch

drm/amd/amdgpu: Print ring name in amdgpu_ib_schedule()

If the ring isn't ready lets print out which ring name
to help debugging.

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: remove dead code, si_mc_load_microcode (v2)

In an if block for (running == 0) running cannot be non-zero.

v2: agd: remove unused variable

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon/cik: remove dead code (v2)

In an if block for (running == 0) running cannot be non-zero.

v2: agd: remove unused variable

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: avoid NULL dereference, cz_hwmgr.c

if (a == NULL || a->b == NULL)
leads to a NULL pointer dereference if a == NULL.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: avoid NULL pointer dereference

if (a == NULL || a->b == NULL)
leads to a NULL pointer dereference if a == NULL.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gmc8: remove dead code (v2)

In an if block for (running == 0) running cannot be non-zero.

v2: agd: remove unused variable

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gmc7: remove dead code (v2)

In an if block for (running == 0) running cannot be non-zero.

v2: agd: remove unused variable

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix indentation in dce_v8_0_audio_write_sad_regs()

Fixed indentation for readability.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/rockchip: analogix_dp: drop unnecessary probe deferral "error" print

This driver is the only user of of_drm_find_panel() which prints an
error before doing probe deferral, yielding messages like this on boot,
before eventually succeeding:

[    2.234271] [drm:rockchip_dp_probe] *ERROR* failed to find panel
...
[    4.797539] [drm:rockchip_dp_probe] *ERROR* failed to find panel
...

Let's just drop the message.

Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: Enable vblank without event

vblank should be enabled regardless of whether an event
is expected back. This is especially important for a cursor
plane.

Reviewed-by: Yakir Yang <ykk@rock-chips.com>
Tested-by: Yakir Yang <ykk@rock-chip.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: Improve analogix-dp psr handling

Remove the delayed worker, opting instead for the non-delayed
variety. Also introduce a lock to ensure we don't have races
with the worker and psr_state. Finally, cancel and wait for
the worker to finish when disabling the bridge.

Reviewed-by: Yakir Yang <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: A couple small fixes to psr

A few things that need tidying up, no functional changes.

Reviewed-by: Yakir Yang <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: Use a spinlock to protect psr state

The handling of psr state is racey, shore that up with
a per-psr driver lock.

Reviewed-by: Yakir Yang <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: Don't use a delayed worker for psr state changes

The delayed worker isn't needed and is racey. Remove it and do
the state change in line.

Reviewed-by: Yakir Yang <ykk@rock-chips.com>
Tested-by: Yakir Yang <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: Convert psr_list_mutex to spinlock and use it

This patch converts the psr_list_mutex to a spinlock and locks
all access to psr_list to avoid races (however unlikely they
were).

Reviewed-by: Yakir Yang <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: analogix_dp: implement PSR function

Alway enable the PSR function for Rockchip analogix_dp driver. If panel
don't support PSR, then the core analogix_dp would ignore this setting.

Signed-off-by: Yakir Yang <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/bridge: analogix_dp: add the PSR function support

The full name of PSR is Panel Self Refresh, panel device could refresh
itself with the hardware framebuffer in panel, this would make lots of
sense to save the power consumption.

This patch have exported two symbols for platform driver to implement
the PSR function in hardware side:
- analogix_dp_active_psr()
- analogix_dp_inactive_psr()

Reviewed-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Yakir Yang <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: add an common abstracted PSR driver

The PSR driver have exported four symbols for specific device driver, and
it's safe to call them in interrupt context:
- rockchip_drm_psr_register()
- rockchip_drm_psr_unregister()
- rockchip_drm_psr_enable()
- rockchip_drm_psr_disable()
- rockchip_drm_psr_flush()

Encoder driver should call the register/unregister interfaces to hook
itself into common PSR driver, encoder have implement the 'psr_set'
callback which use the set PSR state in hardware side.

Crtc driver would call the enable/disable interfaces when vblank is
enable/disable, after that the common PSR driver would call the encoder
registered callback to set the PSR state.

Fb driver would call the flush interface in 'fb->dirty' callback, this
helper function would force all PSR enabled encoders to exit from PSR
for 3 seconds.

Signed-off-by: Yakir Yang <ykk@rock-chips.com>
[seanpaul removed leftover psr_enabled/psr_work kruft from drm_vop.c]
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: vop: export line flag function

VOP have integrated a hardware counter which indicate the exact display
line that vop is scanning. And if we're interested in a specific line,
we can set the line number to vop line_flag register, and then vop would
generate a line_flag interrupt for it.

For example eDP PSR function is interested in the vertical blanking
period, then driver could set the line number to zero.

This patch have exported a symbol that allow other driver to listen the
line flag event with given timeout limit:
-  rockchip_drm_wait_line_flag()

Signed-off-by: Yakir Yang <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/bridge: analogix_dp: Ensure the panel is properly prepared/unprepared

Instead of just preparing the panel on bind, actually prepare/unprepare
during modeset/disable. The panel must be prepared in order to read hpd
status and edid, so we need to keep state around the prepares in order
to ensure we don't accidentally turn the panel off at the wrong time.

Reviewed-by: Yakir Yang <ykk@rock-chips.com>
Tested-by: Yakir Yang <ykk@rock-chips.com>
Reviewed-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

dt-bindings: add compatible strings for big/little rockchip vops

This patch documents the compatible strings for the big and little vop
in rockchip's drm driver.

Cc: Rob Herring <robh+dt@kernel.org>
Cc: Pawel Moll <pawel.moll@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Ian Campbell <ijc+devicetree@hellion.org.uk>
Cc: Kumar Gala <galak@codeaurora.org>
Reviewed-by: Tomasz Figa <tfiga@chromium.org>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
[seanpaul removed superfluous description per tfiga's review]
Signed-off-by: Sean Paul <seanpaul@chromium.org>

dt-bindings: sort Rockchip vop compatible by chip's number

Reorder the compatible vop devices to be sorted by chip number
in ascending order.

Reviewed-by: Tomasz Figa <tfiga@chromium.org>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
[seanpaul added commit description per tfiga's review]
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: vop: add rk3399 vop support

There are two VOP in rk3399 chip, respectively VOP_BIG and VOP_LIT.
most registers layout of this two vop is same, their framework are both
VOP_FULL, the Major differences of this two is that:

VOP_BIG max output resolution is 4096x2160.
VOP_LIT max output resolution is 2560x1600

VOP_BIG support four windows.
VOP_LIT only support two windows.

RK3399 vop register layout is similar with rk3288, so some feature
can reuse with rk3288.

Reviewed-by: Tomasz Figa <tfiga@chromium.org>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: vop: introduce VOP_REG_MASK

Some new vop register support mask, bit[16-31] is mask,
bit[0-15] is value, the mask is correspond to the value.

Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
[seanpaul masked 'v' per tfiga's review comments]
Reviewed-by: Mark Yao <mark.yao@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: sort registers define by chip's number

No functional changes, sort the vop registers to make
code more readable.

Signed-off-by: Mark Yao <mark.yao@rock-chips.com>
[seanpaul resolved conflict with name change from _3066 to _3036]
Reviewed-by: Mark Yao <mark.yao@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>

drm/rockchip: Delete unnecessary assignment for the field "owner"

The field "owner" is set by the core.
Thus delete an unneeded initialisation.

Generated by: scripts/coccinelle/api/platform_no_drv_owner.cocci
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: http://patchwork.freedesktop.org/patch/msgid/41e48fd3-f713-7225-1df2-3b1f4758f949@users.sourceforge.net

drm/bridge: dw-hdmi: Delete unnecessary assignment for the field "owner"

The field "owner" is set by the core.
Thus delete an unneeded initialisation.

Generated by: scripts/coccinelle/api/platform_no_drv_owner.cocci
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: http://patchwork.freedesktop.org/patch/msgid/7c97565d-eee3-45e9-b494-9e802977502c@users.sourceforge.net

drm/rockchip: Don't continue trying to enable crtc on failure

If vop_enable fails, don't continue on, it causes system hangs.

Reviewed-by: Yakir Yank <ykk@rock-chips.com>
Tested-by: Yakir Yank <ykk@rock-chips.com>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: http://patchwork.freedesktop.org/patch/msgid/1471302749-2811-1-git-send-email-seanpaul@chromium.org

drm/fb-helper: Add drm_fb_helper_set_suspend_unlocked()

This adds a function that also takes the console lock before calling
fb_set_suspend() in contrast to drm_fb_helper_set_suspend() which is
a plain wrapper around fb_set_suspend().
Resume is run asynchronously using a worker if the console lock is
already taken. This is modelled after the i915 driver.

Signed-off-by: Noralf Trønnes <noralf@tronnes.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1471953246-29602-1-git-send-email-noralf@tronnes.org

drm/fb-helper: Fix the dummy remove_conflicting_framebuffers

We always need to remove conflicting framebuffers if any other fb driver
is enabled, and not just if we are setting up an fbdev ourselves.

Unfortunately remove_conflicting_framebuffers() was incorrectly stubbed
out if !fbdev rather than !fb leading to major memory corruption (and
corrupt filesystems) upon boot.

Fixes: 44adece57e26 ("drm/fb-helper: Add a dummy remove_conflicting_framebuffers")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Tobias Jakobi <tjakobi@math.uni-bielefeld.de>
Cc: Noralf Trønnes <noralf@tronnes.org>
Cc: tomi.valkeinen@ti.com
Cc: dh.herrmann@gmail.com
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20160822204452.2509-1-chris@chris-wilson.co.uk

drm/udl: Ensure channel is selected before using the device.

Lift configuration command from udlfb. This appears to be essential for
at least a Rextron VCUD-60, without which no URB communication occurs.

Signed-off-by: Jamie Lentin <jm@lentin.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1471904254-26545-1-git-send-email-jm@lentin.co.uk

drm/amdgpu: Use correct mask in dce_v8_0_afmt_setmode() and fix comment typos.

We were using the same mask twice. Looking at radeon, it seems
we should be using HDMI_AVI_INFO_CONT instead as the second mask.

Being there, fix typos in comments and improved readability.

I haven't looked at other DCEs, the mask may also be wrong for them.

Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: cleanup amdgpu_vm_bo_update params

Make it more obvious what we are doing here.

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

drm/amdgpu: stop adding dummy entry in amdgpu_ttm_placement_init

AMDGPU_GEM_CREATE_NO_CPU_ACCESS and AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED are
obviously mutual exclusive. So stop adding a dummy entry without effect when
both are specified.

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

drm/amdgpu: cleanup amdgpu_ttm_placement_init

Make it more clear what this function does. No intendet functional change.

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

drm/amdgpu: fix lru size grouping v2

Adding a BO can make it the insertion point for larger sizes as well.

v2: add a comment about the guard structure.

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

drm/amd/powerplay:  Tidy up cz_hwmgr.c

Clean up whitespace and formatting.

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add AMDGPU_INFO_NUM_EVICTIONS

For profiling.

v2: really bump the minor version

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: recover vram bo from shadow after gpu reset V2

V2:
1. don't directly submit to many jobs at the same time.
2. delete unrelated printk.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: link all shadow bo V2

V2:
1. use mutex instead of spinlock for shadow list, since its process could
sleep.
2. move list_del to bo destroy phase.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update pt shadow while updating pt V2

V2:
move shadow parameter to amdgpu_pte_update_params.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: minutemaidpark@hotmail.com
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update pd shadow while updating pd V2

V2:
Checking if shadow is valid.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: sync bo and shadow V3

Use shadow flag to judge which direction to sync.
V2:
Don't need bo pin, so remove it.

V3:
1. Split to two functions, one is backup_to_shadow, another is
restore_from_shadow.
2. Clean up previous shadow direction difinitions.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add direct submision option for copy_buffer

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm: Avoid calling dev_printk(.dev = NULL)

Since dev_printk likes to print "(NULL device *):" when passed in a NULL
pointer, we have to manually call printk() ourselves.

Fixes: c4e68a583202 ("drm: Introduce DRM_DEV_* log messages")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Eric Engestrom <eric.engestrom@imgtec.com>
Cc: Sean Paul <seanpaul@chromium.org>
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Link: http://patchwork.freedesktop.org/patch/msgid/20160819073750.16610-1-chris@chris-wilson.co.uk

drm: avoid exposing kernel stack in compat_drm_getstats

The C standard does not specify the size of the integer used
to store an enum. Hence in structure drm_stats32_t alignment
bytes may exist.

To avoid exposing bytes from the kernel stack it is
necessary to initialize variable s32 completely.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1471802179-2886-1-git-send-email-xypron.glpk@gmx.de

reservation: fix small comment typo

Signed-off-by: Rob Clark <robdclark@gmail.com>
[danvet: Mark up as function for proper cross-linking.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1471640134-30888-1-git-send-email-robdclark@gmail.com

drm/i915: Update DRIVER_DATE to 20160822

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

drm/amdgpu: add need backup function V2

V2:
add checking if need backup in amdgpu_bo_create.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/amdgpu: S3 resumed failed after 4-5 times loop

Phenomenon: software hang when device resume back, read UVD fence is 0xffffffff
and read pcie pid is 0xffff.
The issue is caused by VCE reset when update cg setting. according to HW programming
guide, adjust update VCE cg sequence.
The patch apply to VCE2.0.

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

drm/amdgpu: use domain's gpu_offset for start addr

Signed-off-by: Flora Cui <Flora.Cui@amd.com>
Reviewed-by: Ken Wang <Qingqing.Wang@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update gart_pin_size only if the bo is pined to GTT

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

drm/amdgpu: check domain sanity in amdgpu_bo_pin_restricted()

abort if the bo is pined to other domain already

Signed-off-by: Flora Cui <Flora.Cui@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdgpu: move ttm stuff to amdgpu_ttm.h

Signed-off-by: Flora Cui <Flora.Cui@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: enable power containment features for tonga.

v2: fix build error introduced when fix code style problems.

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

drm/i915: Use remap_io_mapping() to prefault all PTE in a single pass

Very old numbers indicate this is a 66% improvement when remapping the
entire object for fence contention - due to the elimination of
track_pfn_insert and its strcmp.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Testcase: igt/gem_fence_upload/performance
Testcase: igt/gem_mmap_gtt
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160819155428.1670-6-chris@chris-wilson.co.uk

drm/i915: Embed the io-mapping struct inside drm_i915_private

As io_mapping.h now always allocates the struct, we can avoid that
allocation and extra pointer dance by embedding the struct inside
drm_i915_private

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160819155428.1670-5-chris@chris-wilson.co.uk

io-mapping: Always create a struct to hold metadata about the io-mapping

Currently, we only allocate a structure to hold metadata if we need to
allocate an ioremap for every access, such as on x86-32. However, it
would be useful to store basic information about the io-mapping, such as
its page protection, on all platforms.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: linux-mm@kvack.org
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160819155428.1670-4-chris@chris-wilson.co.uk

drm/i915/fbc: Allow on unfenced surfaces, for recent gen

Only fbc1 is tied to using a fence. Later iterations of fbc are more
flexible and allow operation on unfenced frontbuffers.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: "Zanoni, Paulo R" <paulo.r.zanoni@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160819155428.1670-3-chris@chris-wilson.co.uk

drm/i915/fbc: Don't set an illegal fence if unfenced

If the frontbuffer doesn't have an associated fence, it will have a
fence reg of -1. If we attempt to OR in this register into the FBC
control register we end up setting all control bits, oops!

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: "Zanoni, Paulo R" <paulo.r.zanoni@intel.com>
Reviwed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160819155428.1670-2-chris@chris-wilson.co.uk

drm/i915: Flush delayed fence releases after reset

What I never hit in testing, but Mika immediately did, was a GPU hang
with a pending fence release (where a tiled object has been changed by
the user to be untiled, and the update has not yet been committed to the
fence register). As the stride/tiling is 0, this causes a divide-by-zero
error when trying to write the new fence parameters:

[   28.784518] drm/i915: Resetting chip after gpu hang
[   28.784551] divide error: 0000 [#1] PREEMPT SMP
[   28.784565] Modules linked in: nls_iso8859_1 nls_cp437 vfat fat mxm_wmi x86_pkg_temp_thermal snd_hda_codec_hdmi kvm irqbypass snd_hda_codec_realtek snd_hda_codec_generic snd_hda_intel snd_hda_codec serio_raw snd_hwdep snd_hda_core snd_pcm snd_seq_midi snd_seq_midi_event snd_rawmidi snd_seq snd_timer snd_seq_device snd soundcore mac_hid wmi efivarfs autofs4 raid10 raid456 libcrc32c async_raid6_recov async_memcpy async_pq raid6_pq async_xor xor async_tx raid0 multipath linear psmouse e1000e ptp pps_core nvme nvme_core i915 i2c_algo_bit drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops drm video
[   28.784738] CPU: 0 PID: 1692 Comm: kworker/0:2 Not tainted 4.8.0-rc2+ #895
[   28.784752] Hardware name: System manufacturer System Product Name/Z170M-PLUS, BIOS 1803 05/09/2016
[   28.784786] Workqueue: events_long i915_hangcheck_elapsed [i915]
[   28.784814] task: ffff923c18f59d40 task.stack: ffff923c1b7e4000
[   28.784827] RIP: 0010:[<ffffffffc0475b5f>]  [<ffffffffc0475b5f>] fence_write+0x9f/0x3b0 [i915]
[   28.784854] RSP: 0018:ffff923c1b7e7b30  EFLAGS: 00010246
[   28.784866] RAX: 00000000008ca000 RBX: ffff923c18540000 RCX: 0000000000000020
[   28.784880] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 000000000596d000
[   28.784894] RBP: ffff923c1b7e7b68 R08: 0000000000000000 R09: 0000000000000000
[   28.784908] R10: 0000000000000000 R11: 00000000008ca000 R12: ffff923c1ef9d600
[   28.784921] R13: 0000000000100040 R14: 0000000000100044 R15: ffff923c18549908
[   28.784935] FS:  0000000000000000(0000) GS:ffff923c36c00000(0000) knlGS:0000000000000000
[   28.784951] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   28.784962] CR2: 00007f193373c893 CR3: 0000000419c78000 CR4: 00000000003406f0
[   28.784976] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[   28.784990] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[   28.785004] Stack:
[   28.785009]  000000000596c03b ffff923c1b7e7b68 ffff923c18549938 0000000000000009
[   28.785026]  ffff923c18540000 ffff923c18549280 ffff923c18547ce8 ffff923c1b7e7b90
[   28.785044]  ffffffffc04761f9 ffff923c18540000 ffff923c18547d00 ffff923c18548ff8
[   28.785062] Call Trace:
[   28.785078]  [<ffffffffc04761f9>] i915_gem_restore_fences+0x39/0x50 [i915]
[   28.785102]  [<ffffffffc047fe89>] i915_gem_reset+0x179/0x300 [i915]

Reported-by: Mika Kuoppala <mika.kuoppala@intel.com>
Fixes: 49ef5294cda2 ("drm/i915: Move fence tracking from object to vma")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160819155428.1670-1-chris@chris-wilson.co.uk

drm/i915: Reattach comment, complete type specification

In the recent patch
bc3d674 drm/i915: Allow userspace to request no-error-capture upon ...
the final version moved the flags and the associated #defines around
so they were adjacent; unfortunately, they ended up between a comment
and the thing (hw_id) to which the comment applies :(

So this patch reshuffles the comment and subject back together.

Also, as we're touching 'hw_id', let's change it from just 'unsigned'
to a fully-specified 'unsigned int', because some code checking tools
(including checkpatch) object to plain 'unsigned'.

Fixes: bc3d674462e5 ("drm/i915: Allow userspace to request no-error-capture...")
Signed-off-by: Dave Gordon <david.s.gordon@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/1471616622-6919-1-git-send-email-david.s.gordon@intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

drm: Don't implement empty prepare_fb()/cleanup_fb()

The plane .prepare_fb() and .cleanup_fb() helpers are optional, there's
no need to implement empty stubs, and no need to explicitly set the
function pointers to NULL either.

Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
[danvet: Resolved conflicts with Chris' patch.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

virtio-gpu: Use memdup_user() rather than duplicating its implementation

Reuse existing functionality from memdup_user() instead of keeping
duplicate source code.

This issue was detected by using the Coccinelle software.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/401e68fc-5515-7a7a-be2e-503dee676b34@users.sourceforge.net

GPU-DRM-Savage: Use memdup_user() rather than duplicating

Reuse existing functionality from memdup_user() instead of keeping
duplicate source code.

This issue was detected by using the Coccinelle software.

Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/021efa32-c259-e799-a80e-630972291815@users.sourceforge.net

drm: Allow drivers to modify plane_state in prepare_fb/cleanup_fb

The drivers have to modify the atomic plane state during the prepare_fb
callback so they track allocations, reservations and dependencies for
this atomic operation involving this fb. In particular, how else do we
set the plane->fence from the framebuffer!

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818180017.20508-1-chris@chris-wilson.co.uk

drm/i915/cmdparser: Accelerate copies from WC memory

If we need to use clflush to prepare our batch for reads from memory, we
can bypass the cache instead by using non-temporal copies.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.william.auld@gmail.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-39-chris@chris-wilson.co.uk

drm/i915/cmdparser: Use binary search for faster register lookup

A significant proportion of the cmdparsing time for some batches is the
cost to find the register in the mmiotable. We ensure that those tables
are in ascending order such that we could do a binary search if it was
ever merited. It is.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-38-chris@chris-wilson.co.uk

drm/i915/cmdparser: Check for SKIP descriptors first

If the command descriptor says to skip it, ignore checking for anyother
other conflict.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-37-chris@chris-wilson.co.uk

drm/i915/cmdparser: Compare against the previous command descriptor

On the blitter (and in test code), we see long sequences of repeated
commands, e.g. XY_PIXEL_BLT, XY_SCANLINE_BLT, or XY_SRC_COPY. For these,
we can skip the hashtable lookup by remembering the previous command
descriptor and doing a straightforward compare of the command header.
The corollary is that we need to do one extra comparison before lookup
up new commands.

v2: Less magic mask (ok, it is still magic, but now you cannot see!)

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-36-chris@chris-wilson.co.uk

drm/i915/cmdparser: Improve hash function

The existing code's hashfunction is very suboptimal (most 3D commands
use the same bucket degrading the hash to a long list). The code even
acknowledge that the issue was known and the fix simple:

/*
 * If we attempt to generate a perfect hash, we should be able to look at bits
 * 31:29 of a command from a batch buffer and use the full mask for that
 * client. The existing INSTR_CLIENT_MASK/SHIFT defines can be used for this.
 */

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-35-chris@chris-wilson.co.uk

drm/i915/cmdparser: Only cache the dst vmap

For simplicity, we want to continue using a contiguous mapping of the
command buffer, but we can reduce the number of vmappings we hold by
switching over to a page-by-page copy from the user batch buffer to the
shadow. The cost for saving one linear mapping is about 5% in trivial
workloads - which is more or less the overhead in calling kmap_atomic().

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-34-chris@chris-wilson.co.uk

drm/i915/cmdparser: Use cached vmappings

The single largest factor in the overhead of parsing the commands is the
setup of the virtual mapping to provide a continuous block for the batch
buffer. If we keep those vmappings around (against the better judgement
of mm/vmalloc.c, which we offset by handwaving and looking suggestively
at the shrinker) we can dramatically improve the performance of the
parser for small batches (such as media workloads). Furthermore, we can
use the prepare shmem read/write functions to determine  how best we
need to clflush the range (rather than every page of the object).

The impact of caching both src/dst vmaps is +80% on ivb and +140% on byt
for the throughput on small batches. (Caching just the dst vmap and
iterating over the src, doing a page by page copy is roughly 5% slower
on both platforms. That may be an acceptable trade-off to eliminate one
cached vmapping, and we may be able to reduce the per-page copying overhead
further.) For *this* simple test case, the cmdparser is now within a
factor of 2 of ideal performance.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-33-chris@chris-wilson.co.uk

drm/i915/cmdparser: Add the TIMESTAMP register for the other engines

Since I have been using the BCS_TIMESTAMP to measure latency of
execution upon the blitter ring, allow regular userspace to also read
from that register. They are already allowed RCS_TIMESTAMP!

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-32-chris@chris-wilson.co.uk

drm/i915/cmdparser: Make initialisation failure non-fatal

If the developer adds a register in the wrong order, we BUG during boot.
That makes development and testing very difficult. Let's be a bit more
friendly and disable the command parser with a big warning if the tables
are invalid.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-31-chris@chris-wilson.co.uk

drm/i915: Stop discarding GTT cache-domain on unbind vma

Since commit 43566dedde54 ("drm/i915: Broaden application of
set-domain(GTT)") we allowed objects to be in the GTT domain, but unbound.
Therefore removing the GTT cache domain when removing the GGTT vma is no
longer semantically correct.

An unfortunate side-effect is we lose the wondrously named
i915_gem_object_finish_gtt(), not to be confused with
i915_gem_gtt_finish_object()!

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Akash Goel <akash.goel@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-30-chris@chris-wilson.co.uk

drm/i915: Bump the inactive tracking for all VMA accessed

We track the LRU access for eviction and bump the last access for the
user GGTT on set-to-gtt. When we do so we need to not only bump the
primary GGTT VMA but all partials as well. Similarly we want to
bump the last access tracking for when unpinning an object from the
scanout so that they do not get promptly evicted and hopefully remain
available for reuse on the next frame.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-29-chris@chris-wilson.co.uk

drm/i915: Track display alignment on VMA

When using the aliasing ppgtt and pageflipping with the shrinker/eviction
active, we note that we often have to rebind the backbuffer before
flipping onto the scanout because it has an invalid alignment. If we
store the worst-case alignment required for a VMA, we can avoid having
to rebind at critical junctures.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-28-chris@chris-wilson.co.uk

drm/i915: Fallback to using unmappable memory for scanout

The existing ABI says that scanouts are pinned into the mappable region
so that legacy clients (e.g. old Xorg or plymouthd) can write directly
into the scanout through a GTT mapping. However if the surface does not
fit into the mappable region, we are better off just trying to fit it
anywhere and hoping for the best. (Any userspace that is capable of
using ginormous scanouts is also likely not to rely on pure GTT
updates.) With the partial vma fault support, we are no longer
restricted to only using scanouts that we can pin (though it is still
preferred for performance reasons and for powersaving features like
FBC).

v2: Skip fence pinning when not mappable.
v3: Add a comment to explain the possible ramifications of not being
    able to use fences for unmappable scanouts.
v4: Rebase to skip over some local patches
v5: Rebase to defer until after we have unmappable GTT fault support

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Deepak S <deepak.s@linux.intel.com>
Cc: Damien Lespiau <damien.lespiau@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-27-chris@chris-wilson.co.uk

drm/i915: Choose not to evict faultable objects from the GGTT

Often times we do not want to evict mapped objects from the GGTT as
these are quite expensive to teardown and frequently reused (causing an
equally, if not more so, expensive setup). In particular, when faulting
in a new object we want to avoid evicting an active object, or else we
may trigger a page-fault-of-doom as we ping-pong between evicting two
objects.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-26-chris@chris-wilson.co.uk

drm/i915: Drop ORIGIN_GTT for untracked GTT writes

If FBC is set on a framebuffer that is unmapped, all GTT faults will be
from a partial mapping. Writes by the user through the partial VMA are
then untracked by the FBC and so we must use the ORIGIN_CPU when flushing
the I915_GEM_DOMAIN_GTT.

v2: Keep ORIGIN_CPU for set-to-domain(.write=CPU)

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: "Zanoni, Paulo R" <paulo.r.zanoni@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-25-chris@chris-wilson.co.uk

drm/i915: Convert partial ggtt vma to full ggtt if it spans the entire object

If we want to create a partial vma from a chunk that is the same size as
the object, create a normal ggtt vma instead. The benefit is that it
will match future requests for the normal ggtt.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-24-chris@chris-wilson.co.uk

drm/i915: Fix partial GGTT faulting

We want to always use the partial VMA as a fallback for a failure to
bind the object into the GGTT. This extends the support partial objects
in the GGTT to cover everything, not just objects too large.

v2: Call the partial view, view not partial.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20160818161718.27187-23-chris@chris-wilson.co.uk