Add 3D support to the virtio-gpu.

* 'virtio-gpu-for-drm-next' of git://git.kraxel.org/linux:
  virtio-gpu: add page flip support
  virtio-gpu: mark as a render gpu
  virtio-gpu: add basic prime support
  virtio-gpu: add 3d/virgl support
  virtio-gpu: don't free things on ttm_bo_init failure
  virtio-gpu: wait for cursor updates finish
  virtio-gpu: add & use virtio_gpu_queue_fenced_ctrl_buffer
  virtio-gpu: add virtio_gpu_queue_ctrl_buffer_locked

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Also add DRM_RENDER_ALLOW to the ioctls.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

This is enough to enable DRI3.
Signed-off-by: NDave Airlie <airlied@redhat.com>
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Add the bits needed for opengl rendering support: query
capabilities, new virtio commands, drm ioctls.
Signed-off-by: NDave Airlie <airlied@redhat.com>
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

ttm_bo_init will call the destroy callback which will do all
this, this was causing a double free.
Signed-off-by: NDave Airlie <airlied@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Add helper function to handle the submission of fenced control requests.
Make sure we initialize the fence while holding the virtqueue lock, so
requests can't be reordered.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Add virtio_gpu_queue_ctrl_buffer_locked function, which does the same as
virtio_gpu_queue_ctrl_buffer but does not take the virtqueue lock.  The
caller must hold the lock instead.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Fixes userspace compile error since list_head is not exported to userspace
headers.

Suggested by Emil Velikov <emil.l.velikov@gmail.com> at
https://lkml.org/lkml/2015/6/3/792Signed-off-by: NMikko Rapeli <mikko.rapeli@iki.fi>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Fixes userspace compile error:

drm/sis_drm.h:68:19: error: field ‘obj_list’ has incomplete type
  struct list_head obj_list;

Suggested by Emil Velikov <emil.l.velikov@gmail.com> at
https://lkml.org/lkml/2015/6/3/792Signed-off-by: NMikko Rapeli <mikko.rapeli@iki.fi>
Signed-off-by: NDave Airlie <airlied@redhat.com>

This series:
* adds support for interlaced video modes to the ipu-v3 driver
  and dw_hdmi bridge.
* reworks the dw_hdmi connector enable/disable support, to ensure that
  when DRM disables the output, it stays disabled irrespective of the
  hotplug state.
* adds support for connector forcing, so we can force the hotplug state
  for this connector.
* adds the ALSA AHB audio driver to the bridge: Iwai has acked the
  audio driver.
* a few fixes to the ACR calculations to allow more modes to work with
  audio on iMX6.

Fabio has independently tested this series, so all patches here carry
his tested-by tag.

* 'drm-dwhdmi-devel' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  drm: bridge/dw_hdmi: replace CTS calculation for the ACR
  drm: bridge/dw_hdmi: remove ratio support from ACR code
  drm: bridge/dw_hdmi: adjust pixel clock values in N calculation
  drm: bridge/dw_hdmi: avoid being recursive in N calculation
  drm: bridge/dw_hdmi-ahb-audio: allow larger buffer sizes
  drm: bridge/dw_hdmi-ahb-audio: basic support for multi-channel PCM audio
  drm: bridge/dw_hdmi-ahb-audio: parse ELD from HDMI driver
  drm: bridge/dw_hdmi-ahb-audio: add audio driver
  drm: bridge/dw_hdmi: improve HDMI enable/disable handling
  drm: bridge/dw_hdmi: add connector mode forcing
  drm: bridge/dw_hdmi: add support for interlaced video modes
  gpu: imx: fix support for interlaced modes
  gpu: imx: simplify sync polarity setting

Merge commit '06d1ee32' of git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux into drm-next

Backmerge the drm-fixes pull from Linus's tree into drm-next.

This is to fix some conflicts and make future pulls cleaner

Another round of drm-misc. Unfortunately the DRM_UNLOCKED removal for
DRIVER_MODESET isn't complete yet for lack of review on 1-2 patches.
Otherwise just various stuff all over.

* tag 'topic/drm-misc-2015-10-08' of git://anongit.freedesktop.org/drm-intel:
  drm: Stop using drm_vblank_count() as the hw frame counter
  drm/irq: Use unsigned int pipe in public API
  drm: Use DRM_ROTATE_MASK and DRM_REFLECT_MASK
  drm: Add DRM_ROTATE_MASK and DRM_REFLECT_MASK
  vga_switcheroo: Add missing locking
  vgaarb: use kzalloc in vga_arbiter_add_pci_device()
  drm: Don't zero vblank timestamps from the irq handler
  drm: Hack around CONFIG_AGP=m build failures
  drm/i915: Remove setparam ioctl
  drm: Remove dummy agp ioctl wrappers
  drm/vmwgfx: Stop checking for DRM_UNLOCKED
  drm/drm_ioctl.c: kerneldoc
  drm: Define a drm_invalid_op ioctl implementation
  drm: Remove __OS_HAS_AGP
  drm/doc: Update docs about device instance setup

Merge tag 'drm-intel-next-2015-09-28-merged' of git://anongit.freedesktop.org/drm-intel into drm-next

drm-intel-next-2015-09-28:
- fastboot by default for some systems (Maarten Lankhorts)
- piles of workarounds for bxt and skl
- more fbc work from Paulo
- fix hdmi hotplug detection (Sonika)
- first few patches from Ville to parametrize register macros, prep work for
  typesafe mmio functions
- prep work for nv12 rotation (Tvrtko Ursulin)
- various other bugfixes and improvements all over

I have another backmerge here since things became messy and I didn't
realize you resolved some of them already (usually you complain when
there's a conflict ...).

For 4.4 I plan one more feature round after this and then that's it.
* tag 'drm-intel-next-2015-09-28-merged' of git://anongit.freedesktop.org/drm-intel: (80 commits)
  drm/i915: Update DRIVER_DATE to 20150928
  drm/i915: fix task reference leak in i915_debugfs.c
  drm/i915: Defer adding preallocated stolen objects to the VM list
  drm/i915: Remove extraneous request cancel.
  drm/i915: Enable querying offset of UV plane with intel_plane_obj_offset
  drm/i915: Support NV12 in rotated GGTT mapping
  drm/i915: Support appending to the rotated pages mapping
  drm/i915: Support planar formats in tile height calculations
  drm/i915/bxt: Update revision id for BXT C0
  drm/i915: Parametrize CSR_PROGRAM registers
  drm/i915: Parametrize DDI_BUF_TRANS registers
  drm/i915: Parametrize TV luma/chroma filter registers
  drm/i915: Replace raw numbers with the approproate register name in ILK turbo code
  drm/i915: Parametrize ILK turbo registers
  drm/i915: Parametrize FBC_TAG registers
  drm/i915: Parametrize GEN7_GT_SCRATCH and GEN7_LRA_LIMITS
  drm/i915: Parametrize LRC registers
  drm/i915: Don't pass sdvo_reg to intel_sdvo_select_{ddc, i2c}_bus()
  drm/i915: Ignore "digital output" and "not HDMI output" bits for eDP detection
  drm/i915: Make sure we don't detect eDP on g4x
  ...

This is the first radeon and amdgpu pull for drm-next. Highlights include:
- Efficiency improvements to the CS checker for pre-SI asics
- Cursor fixes ported from radeon to amdgpu
- Enable GPU scheduler by default
- Add a bunch of GPUVM debugging options
- Add support for some new atombios opcodes
- Misc cleanups and fixes

* 'drm-next-4.4' of git://people.freedesktop.org/~agd5f/linux: (42 commits)
  drm/amdgpu: fix lockup when clean pending fences
  drm/amdgpu: add timer to fence to detect scheduler lockup
  drm/amdgpu: add VM CS mapping trace point
  drm/amdgpu: add option to clear VM page tables after every submit
  drm/amdgpu: add option to stop on VM fault
  drm/amdgpu: only print meaningful VM faults
  drm/amdgpu: also trace already allocated VMIDs
  drm/amdgpu: Drop unnecessary #include <linux/vga_switcheroo.h>
  drm/radeon: Drop unnecessary #include <linux/vga_switcheroo.h>
  drm/amdgpu: clean up pageflip interrupt handling
  drm/amdgpu: rework sdma structures
  drm/amdgpu: unpin cursor BOs on suspend and pin them again on resume
  drm/amdgpu/dce8: Fold set_cursor() into show_cursor()
  drm/amdgpu/dce8: Clean up reference counting and pinning of the cursor BOs
  drm/amdgpu/dce8: Move hotspot handling out of set_cursor
  drm/amdgpu/dce8: Re-show the cursor after a modeset (v2)
  drm/amdgpu/dce8: Use cursor_set2 hook for enabling / disabling the HW cursor
  drm/amdgpu/dce11: Fold set_cursor() into show_cursor()
  drm/amdgpu/dce11: Clean up reference counting and pinning of the cursor BOs
  drm/amdgpu/dce11: Move hotspot handling out of set_cursor
  ...

* removes the now unused DRM slave encoder support, which all users have
  migrated away from, allowing us to simplify the code.
* ensure all pending interrupts are processed together, rather than
  needing the handler to be re-entered each time.
* use more HDMI helpers to setup the info frames.
* fix EDID read handling by ensuring that we always wait the specified time
  before attempting to read the EDID, no matter where the EDID read request
  came from.

* 'drm-tda998x-devel' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  drm/i2c: tda998x: clean up after struct tda998x_priv2 removal
  drm/i2c: tda998x: kill struct tda998x_priv2
  drm/i2c: tda998x: move connector into struct tda998x_priv
  drm/i2c: tda998x: remove encoder pointer
  drm/i2c: tda998x: remove DRM slave encoder support
  drm/i2c: tda998x: use more HDMI helpers
  drm/i2c: tda998x: handle all outstanding interrupts
  drm/i2c: tda998x: convert to u8/u16/u32 types
  drm/i2c: tda998x: re-implement "Fix EDID read timeout on HDMI connect"
  drm/i2c: tda998x: report whether we actually handled the IRQ
  drm/i2c: tda998x: remove useless NULL checks

* remove support for the non-component support from the Armada DRM driver,
  switching it to component-only mode.
* create a "armada plane" to allow the primary and overlay planes to share
  some code.
* increase efficiency by using inherently atomic operations, rather than
  spinlocking to achieve atomicity.  Eg, if we want to exchange a value,
  using xchg().
* increase PM savings by stopping the external pixel clock when we're in
  DPMS mode.

* 'drm-armada-devel' of git://ftp.arm.linux.org.uk/~rmk/linux-arm:
  drm/armada: move frame wait wakeup into plane work
  drm/armada: convert overlay plane vbl worker to a armada plane worker
  drm/armada: move CRTC flip work to primary plane work
  drm/armada: move frame wait into armada_frame
  drm/armada: move the locking for armada_drm_vbl_event_remove()
  drm/armada: move the update of dplane->ctrl0 out of spinlock
  drm/armada: move write to dma_ctrl0 to armada_drm_crtc_plane_disable()
  drm/armada: provide a common helper to disable a plane
  drm/armada: allocate primary plane ourselves
  drm/armada: add primary plane creation
  drm/armada: introduce generic armada_plane struct
  drm/armada: update armada overlay to use drm_universal_plane_init()
  drm/armada: use xchg() to atomically update dplane->old_fb
  drm/armada: factor out retirement of old fb
  drm/armada: rename overlay identifiers
  drm/armada: redo locking and atomics for armada_drm_crtc_complete_frame_work()
  drm/armada: disable CRTC clock during DPMS
  drm/armada: use drm_plane_force_disable() to disable the overlay plane
  drm/armada: move vbl code into armada_crtc
  drm/armada: remove non-component support

The first lockup fence will lock the fence list of scheduler.
Then cancel the delayed workqueues for all clean pending fences
without waiting the workqueues to finish.

Change-Id: I9bec826de1aa49d587b0662f3fb4a95333979429
Signed-off-by: NJunwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>

Change-Id: I67e987db0efdca28faa80b332b75571192130d33
Signed-off-by: NJunwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: NDavid Zhou <david1.zhou@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>

Output all VM mappings a command submission uses.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>

This makes it much easier to find when userspace misses to send some buffers.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>

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

Port of radeon commit 9b7d786b.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>

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

This was added to two radeon files even though they don't use any
vga_switcheroo symbols, the amdgpu fork inherited them:

Added to amdgpu_acpi.c by commit d7a2952f ("drm/radeon: Add
support for the ATIF ACPI method to the radeon driver").

Added to amdgpu_bios.c by commit 6a9ee8af ("vga_switcheroo:
initial implementation (v15)").
Signed-off-by: NLukas Wunner <lukas@wunner.de>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This was added to three files even though they don't use any
vga_switcheroo symbols:

Added to radeon_acpi.c by commit d7a2952f ("drm/radeon: Add
support for the ATIF ACPI method to the radeon driver").

Added to radeon_asic.c by commit 0a10c851 ("drm/radeon: create
radeon_asic.c").

Added to radeon_bios.c by commit 6a9ee8af ("vga_switcheroo:
initial implementation (v15)").
Signed-off-by: NLukas Wunner <lukas@wunner.de>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Check to make sure we aren't touching a non-existent
display controller and simplify the code.
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Rework the sdma structures in the driver to
consolidate all of the sdma info into a single
structure and allow for asics that may have
different numbers of sdma instances.
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Everything is evicted from VRAM before suspend, so we need to make
sure all BOs are unpinned and re-pinned after resume. Fixes broken
mouse cursor after resume introduced by commit b9729b17.

Port of radeon commit:
f3cbb17bReviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Port of radeon commit:
8991668aReviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Take a GEM reference for and pin the new cursor BO, unpin and drop the
GEM reference for the old cursor BO in dce8 crtc_cursor_set2, and use
amdgpu_crtc->cursor_addr in dce8 set_cursor.

This fixes dce8 cursor_reset accidentally incrementing the cursor BO
pin count, and cleans up the code a little.

Port of radeon commit:
cd404af0Reviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

It's only needed in dce8 crtc_cursor_set2.

Port of radeon commit:
2e007e61Reviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Setting a mode seems to clear the cursor registers, so we need to
re-program them to make sure the cursor is visible.

Port of radeon commit:
6d3759fa

v2: change radeon reference in error message
Reviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The cursor_set2 hook provides the cursor hotspot position within the
cursor image. When the hotspot position changes, we can adjust the cursor
position such that the hotspot doesn't move on the screen. This prevents
the cursor from appearing to intermittently jump around on the screen
when the position of the hotspot within the cursor image changes.

Port of radeon commits:
78b1a601
3feba08dReviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Port of radeon commit:
8991668aReviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Take a GEM reference for and pin the new cursor BO, unpin and drop the
GEM reference for the old cursor BO in dce11 crtc_cursor_set2, and use
amdgpu_crtc->cursor_addr in dce11 set_cursor.

This fixes dce11 cursor_reset accidentally incrementing the cursor BO
pin count, and cleans up the code a little.

Port of radeon commit:
cd404af0Reviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

It's only needed in dce11 crtc_cursor_set2.

Port of radeon commit:
2e007e61Reviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Setting a mode seems to clear the cursor registers, so we need to
re-program them to make sure the cursor is visible.

Port of radeon commit:
6d3759fa

v2: change radeon reference in error output
Reviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The cursor_set2 hook provides the cursor hotspot position within the
cursor image. When the hotspot position changes, we can adjust the cursor
position such that the hotspot doesn't move on the screen. This prevents
the cursor from appearing to intermittently jump around on the screen
when the position of the hotspot within the cursor image changes.

Port of radeon commits:
78b1a601
3feba08dReviewed-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>