gma500: fix cdv_intel_wait_for_vblank() prototype

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: fix two -Wmissing-field-initializers warnings

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: mark psb_driver_device_is_agp() and psb_driver_preclose() as static

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: psb_driver_load(): drop unused variables

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: fix ioctl confict

Move DRM_GMA_GET_PIPE_FROM_CRTC_ID to 0x08 to avoid confict with
DRM_GMA_GEM_MMAP.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: use DRM_IOCTL_DEF_DRV instead of custom PSB_IOCTL_DEF

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: mmu: mark psb_mmu_free_pt() and psb_mmu_pt_alloc_map_lock() as static

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: mmu: drop unused psb_get_default_pd_addr()

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: gtt: mark psb_gtt_entry() and psb_gtt_alloc() as static

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: mark psb_fbdev_destroy() and psb_fbdev_fini() as static

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: drop unused psbfb_suspend()/psbfb_resume()

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: gem_glue: add missing include

It fixes W=1 warnings:

drivers/gpu/drm/gma500/gem_glue.c:23:6: warning: no previous prototype for ‘drm_gem_object_release_wrap’ [-Wmissing-prototypes]
drivers/gpu/drm/gma500/gem_glue.c:44:5: warning: no previous prototype for ‘gem_create_mmap_offset’ [-Wmissing-prototypes]

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: medfield: drop a bit of dead code

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: mdfld_dsi_output_init() drop unused parameter

Nobody uses 'config' parameter.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: psbfb_create(): move depth initialization out of loop

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: initial medfield merge

We need to merge this ahead of some of the cleanup because a lot of needed
cleanup spans both new and old chips. If we try and clean up and the merge
we end up fighting ourselves.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
[With a load of the cleanup stuff folded in, register stuff reworked sanely]
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: rework register stuff sanely

Rework registers handling to prepare for Medfield.

Signed-off-by: Alan Cox <alan@linux.intel.com>
[split out from a single big patch]
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: make init_pm callback in struct psb_op optional

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: re-order calling on the fix setup so we set up after the DRM layer

Noted by Kirill A Shutemov

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

intel, gma500, lvds: Fix use after free and mem leak in psb_intel_lvds_init()

In psb_intel_lvds_init(), if we fail to allocate memory for
'psb_intel_connector' we free the memory we previously allocated for
'psb_intel_encoder', but we then proceed to use that free'd pointer
when we do 'psb_intel_encoder->dev_priv = lvds_priv;'.

We may also leak the memory we allocated for 'psb_intel_encoder' if we
'goto failed_connector;' and the variable goes out of scope.

While I was there anyway, I also removed the pointless 'if
(psb_intel_connector)' before freeing it at the 'failed_connector:'
label - kfree() deals gracefully with NULL pointers, so it is not
needed.

Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: Kconfig documentation tweak

Update this to better reflect the status

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: now move the Oaktrail save state into its own structure

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: clean up some of the struct fields we no longer use

Some this is Medfield stuff that may reappear in some form later, other
bits are just dead stuff

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

gma500: plug in more of the gamma functionality

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: cope with platformdev->id == -1

If there are not multiple instances of a platform device, the id
should apparently be set to -1.  Which results in a odd looking
bus-id like "platform:foodrm:-1".  Probably we should just treat
this case as id 0.

Signed-off-by: Rob Clark <rob@ti.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: drop setting vm_file to filp

Talking to Al Viro on irc, we can see no possible reason for doing
this, the upper mmap code does it. The code has been there since
first import into drm tree I can find.

Al tracked down this as a requirement pre 2.3.51 hasn't been needed since.

Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon: fix deferred page-flip detection logic on Avivo-based ASICs

This fixes page-flip-related flickering observed on Iconia Tab W500.

The update_pending status returned by radeon_page_flip is very accurate on
Avivo-based ASICs when vpos is negative.

Experiments were conducted on several ASIC generations ranging from RS690
to Cayman where the page flip was artificially timed to occur at a specific
vpos. With negative vpos, overriding update_pending always lead to
flickering.

The same experiment on RV380 and RV410 showed that update_pending is not
accurate with negative vpos. In most cases update_pending == 1 is returned
although the flip would complete before the start of the next frame.
Therefore I left the behaviour unchanged for pre-AVIVO ASICs for
performance reasons, although this may result in flickering in rare cases.

This change also makes the logic a little easier to understand.

Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Mario Kleiner <mario.kleiner@tuebingen.mpg.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: remove unused code

remove declared but unused functions from drmP.h, fix the comments
where necessary. Also, remove drm_mem_info which is unused.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: clean up radeon_asic struct (v2)

v2: fix typo.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: reorganize surface callbacks

tidy up the radeon_asic struct.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: move clock/pcie setting callbacks into pm struct

tidy up radeon_asic struct.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: reorganize display callbacks

tidy up the radeon_asic struct.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: reorganize gart callbacks

tidy up the radeon_asic struct.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: make ring_start, ring_test, and ib_test per ring

Each ring type may need a different variant.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: remove unused cp callbacks from radeon_asic

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: reorganize irq callbacks

tidy up the radeon_asic struct.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: reorganize copy callbacks

tidy up the radeon_asic struct, handle multiple
rings better.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König<christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon: fix IB debugfs files for multiple cards

Storing pointers to the IBs in a static var just
leads to giving the same content back for all
cards in the system.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon: also make the cs_parse function per ring

Not all rings use PM4, so the cs_parser also needs to be per ring.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: no need to align IB like this

So don't confuse devs by doing so.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alex.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon: move ring syncing after bo validation

The function radeon_bo_list_validate can cause a
bo to move, resulting in a different sync_obj
and a dependency to wait for this move to finish.

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: reorganize pm callbacks

tidy up the radeon_asic struct.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: reorganize page flip callbacks

tidy up the radeon_asic struct.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: reorganize hpd callbacks

tidy up the radeon_asic struct.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: add a radeon asic callback for mc idle

Required for future functionality.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/radeon/kms: add wait_for_vblank asic callback

Required for future functionality.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

Merge tag 'drm-intel-next-2012-02-16-merge-resolved' of git://people.freedesktop.org/~danvet/drm-intel into drm-core-next

* tag 'drm-intel-next-2012-02-16-merge-resolved' of git://people.freedesktop.org/~danvet/drm-intel: (45 commits)
  Revert "drivers/gpu/drm/i915/intel_overlay.c needs seq_file.h"
  drm/i915/lvds: Always use the presence pin for LVDS on PCH
  drm/i915: Record the position of the request upon error
  drm/i915: Record the in-flight requests at the time of a hang
  drm/i915: Record the tail at each request and use it to estimate the head
  drm/i915: add missing SDVO bits for interlaced modes on ILK
  drm/i915: Fix race condition in accessing GMBUS
  drm/i915: add a "force-dvi" HDMI audio mode
  drm/i915: Don't lock panel registers when downclocking
  drm/i915: fix up locking inconsistency around gem_do_init
  drm/i915: enable forcewake voodoo also for gen6
  drm/i915: fixup seqno allocation logic for lazy_request
  drm/i915: outstanding_lazy_request is a u32
  drm/i915: check gtfifodbg after possibly failed writes
  drm/i915: catch gtfifo errors on forcewake_put
  drm/i915: use gtfifodbg
  drm/i915: set interlaced bits for TRANSCONF
  drm/i915: fixup overlay checks for interlaced modes
  drm/i915: allow interlaced mode output on the HDMI connector
  drm/i915: allow interlaced mode output on the SDVO connector
  ...

Merge remote-tracking branch 'airlied/drm-next' into for-airlied

Manually resolve the conflict between the new enum drm property
helpers in drm-next and the new "force-dvi" option that the "audio" output
property gained in drm-intel-next.

While resolving this conflict, switch the new drm_prop_enum_list to
use the newly introduced enum defines instead of magic values.

Conflicts:
drivers/gpu/drm/i915/intel_modes.c

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

drm: add some caps for userspace to discover more info for dumb KMS driver (v2)

For the simple KMS driver case we need some more info about what the preferred
depth and if a shadow framebuffer is preferred.

I've only added this for intel/radeon which support the dumb ioctls so far.

If you need something really fancy you should be writing a real X.org driver.

v2: drop cursor information, just return an error from the cursor ioctls
and we can make userspace fallback to sw cursor in that case, cursor
info was getting too messy, best to start smaller.

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

drm: move pci bus master enable into driver.

The current enabling of bus mastering in the drm midlayer allows a large
race condition under kexec. When a kexec'ed kernel re-enables bus mastering
for the GPU, previously setup dma blocks may cause writes to random pieces
of memory. On radeon the writeback mechanism can cause these sorts of issues.

This patch doesn't fix the problem, but it moves the bus master enable under
the individual drivers control so they can move enabling it until later in
their load cycle and close the race.

Fix for radeon kms driver will be in a follow-up patch.

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

Revert "drivers/gpu/drm/i915/intel_overlay.c needs seq_file.h"

This reverts commit e167976ee7f5fe4b80f7e8f55e087f6c67cf9562,
Since this was already fixed in commit
3bd3c9329973a93fa3ef5e9840f2fd6fa2889e3f some days before this
commit cause seq_file.h to be included twice.

Signed-off-by: Danny Kukawka <danny.kukawka@bisect.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/radeon/evergreen: make texdw[] array larger

We store stuff in texdw[7] so this array needs to have 8 elements.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/i915/lvds: Always use the presence pin for LVDS on PCH

With the introduction of the PCH, we gained an LVDS presence pin but we
continued to use the existing logic that asserted that LVDS was only
supported on certain mobile chipsets. However, there are desktop
IronLake systems with LVDS attached which we fail to detect. So for PCH,
trust the LVDS presence pin and quirk all the lying manufacturers.

Tested-by: Daniel Woff <wolff.daniel@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43171
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: Record the position of the request upon error

So that we can tally the request against the command sequence in the
ringbuffer, or merely jump to the interesting locations.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: Record the in-flight requests at the time of a hang

Being able to tally the list of outstanding requests with the sequence
of commands in the ringbuffer is often useful evidence with respect to
driver corruption.

Note that since this is the umpteenth per-ring data structure to be added
to the error state, I've coallesced the nearby loops (the ringbuffer and
batchbuffer) into a single structure along with the list of requests.  A
later task would be to refactor the ring register state into the same
structure.

v2: Fix pretty printing of requests so that they are parsed correctly by
intel_error_decode and use the 0x%08x format for seqno for consistency

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: Record the tail at each request and use it to estimate the head

By recording the location of every request in the ringbuffer, we know
that in order to retire the request the GPU must have finished reading
it and so the GPU head is now beyond the tail of the request. We can
therefore provide a conservative estimate of where the GPU is reading
from in order to avoid having to read back the ring buffer registers
when polling for space upon starting a new write into the ringbuffer.

A secondary effect is that this allows us to convert
intel_ring_buffer_wait() to use i915_wait_request() and so consolidate
upon the single function to handle the complicated task of waiting upon
the GPU. A necessary precaution is that we need to make that wait
uninterruptible to match the existing conditions as all the callers of
intel_ring_begin() have not been audited to handle ERESTARTSYS
correctly.

By using a conservative estimate for the head, and always processing all
outstanding requests first, we prevent a race condition between using
the estimate and direct reads of I915_RING_HEAD which could result in
the value of the head going backwards, and the tail overflowing once
again. We are also careful to mark any request that we skip over in
order to free space in ring as consumed which provides a
self-consistency check.

Given sufficient abuse, such as a set of unthrottled GPU bound
cairo-traces, avoiding the use of I915_RING_HEAD gives a 10-20% boost on
Sandy Bridge (i5-2520m):
  firefox-paintball  18927ms -> 15646ms: 1.21x speedup
  firefox-fishtank   12563ms -> 11278ms: 1.11x speedup
which is a mild consolation for the performance those traces achieved from
exploiting the buggy autoreported head.

v2: Add a few more comments and make request->tail a conservative
estimate as suggested by Daniel Vetter.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
[danvet: resolve conflicts with retirement defering and the lack of
the autoreport head removal (that will go in through -fixes).]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: add missing SDVO bits for interlaced modes on ILK

This was pointed by Jesse Barnes. The code now seems to follow the
specification but I don't have an SDVO device to really test this.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Merge tag 'drm-intel-next-2012-02-07' of git://people.freedesktop.org/~danvet/drm-intel into drm-core-next

* tag 'drm-intel-next-2012-02-07' of git://people.freedesktop.org/~danvet/drm-intel: (29 commits)
  drm/i915: Handle unmappable buffers during error state capture
  drm/i915: rewrite shmem_pread_slow to use copy_to_user
  drm/i915: rewrite shmem_pwrite_slow to use copy_from_user
  drm/i915: fall through pwrite_gtt_slow to the shmem slow path
  drm/i915: add debugfs file for swizzling information
  drm/i915: fix swizzle detection for gen3
  drm/i915: Remove the upper limit on the bo size for mapping into the CPU domain
  drm/i915: add per-ring fault reg to error_state
  drm/i915: reject GTT domain in relocations
  drm/i915: remove the i915_batchbuffer_info debugfs file
  drm/i915: capture error_state also for stuck rings
  drm/i915: refactor debugfs create functions
  drm/i915: refactor debugfs open function
  drm/i915: don't trash the gtt when running out of fences
  drm/i915: Separate fence pin counting from normal bind pin counting
  drm/i915/ringbuffer: kill snb blt workaround
  drm/i915: collect more per ring error state
  drm/i915: refactor ring error state capture to use arrays
  drm/i915: switch ring->id to be a real id
  drm/i915: set AUD_CONFIG N_value_index for DisplayPort
  ...

drm/i915: Fix race condition in accessing GMBUS

GMBUS has several ports and each has it's own corresponding
I2C adpater. When multiple I2C adapters call gmbus_xfer() at
the same time there is a race condition in using the underlying
GMBUS controller. Fixing this by adding a mutex lock when calling
gmbus_xfer().

v2: Moved gmbus_mutex below intel_gmbus and added comments.
Rebased to drm-intel-next-queued.

Signed-off-by: Yufeng Shen <miletus@chromium.org>
[danvet: Shortened the gmbus_mutex comment a bit and add the patch
revision comment to the commit message.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: add a "force-dvi" HDMI audio mode

When HDMI-DVI converter is used, it's not only necessary to turn off
audio, but also to disable HDMI_MODE_SELECT and video infoframe. Since
the DVI mode is mainly tied to audio functionality from end user POV,
add a new "force-dvi" audio mode:

xrandr --output HDMI1 --set audio force-dvi

Note that most users won't need to set this and happily rely on the EDID
based DVI auto detection.

Reported-by: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: Don't lock panel registers when downclocking

This patch replaces the locking from the downclock routines with an assert
to ensure the registers are indeed unlocked. Without this patch, pre-SNB
devices would lock the registers when downclocking which would cause a
WARNING on suspend/resume with downclocking enabled.

Note: To hit this bug, you need to have lvds downclocking enabled.

Signed-off-by: Sean Paul <seanpaul@chromium.org>
Acked-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/radeon: add support for evergreen/ni tiling informations v11

evergreen and northern island gpu needs more informations for 2D tiling
than previous r6xx/r7xx. Add field to tiling ioctl to allow userspace
to provide those.

The v8 cs checking change to track color view on r6xx/r7xx doesn't
affect old userspace as old userspace always emited 0 for this register.

v2 fix r6xx/r7xx 2D tiling computation
v3 fix r6xx/r7xx height align for untiled surface & add support for
   tile split on evergreen and newer
v4 improve tiling debugging output
v5 fix tile split code for evergreen and newer
v6 set proper tile split for crtc register
v7 fix tile split limit value
v8 add COLOR_VIEW checking to r6xx/r7xx checker, add evergreen cs
   checking, update safe reg for r600, evergreen and cayman.
   Evergreen checking need some work around for stencil alignment
   issues
v9 fix tile split value range, fix compressed texture handling and
   mipmap calculation, allow evergreen check to be silencious in
   front of current broken userspace (depth/stencil alignment issue)
v10 fix eg 3d texture and compressed texture, fix r600 depth array,
    fix r600 color view computation, add support for evergreen stencil
    split
v11 more verbose debugging in some case

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

drm/radeon/kms: add support for streamout v7

v2: agd5f: add strmout CS checking, copy_dw register checking

v3: agd5f: don't use cs_check_reg() for copy_dw checking as it
will incorrectly patch the command stream for certain regs.

v4: agd5f: add warning if safe reg check fails for copy_dw

v5: agd5f: add stricter checking for 6xx/7xx

v6: agd5f: add range checking for copy_dw on eg+,
add sx_surface_sync to safe reg list for 7xx.

v7: agd5f: add stricter checking for eg+

Signed-off-by: Marek Olšák <maraeo@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm: Fix kcalloc parameters swapped

The first parameter should be "number of elements" and the second parameter
should be "element size".

Signed-off-by: Axel Lin <axel.lin@gmail.com>
Acked-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

Merge branch 'drm-vmware-next' into drm-core-next

* drm-vmware-next:
  drm/vmwgfx: Bump driver minor
  vmwgfx: Move function declaration to correct header
  drm/vmwgfx: Treat out-of-range initial width and height as host errors
  vmwgfx: Pick up the initial size from the width and height regs
  vmwgfx: Add page flip support
  vmwgfx: Pipe fence out of screen object dirty functions
  vmwgfx: Make it possible to get fence from execbuf
  vmwgfx: Clean up pending event references to struct drm_file objects on close
  vmwgfx: Rework fence event action

drm/vmwgfx: Bump driver minor

Bump driver minor to signal availability of the page-flip ioctl.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

vmwgfx: Move function declaration to correct header

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/vmwgfx: Treat out-of-range initial width and height as host errors

And assign the initial width and height to the minimum in that case.
Strange values (-1) from these registers have been reported by users.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

vmwgfx: Pick up the initial size from the width and height regs

Signed-off-by: Jakob Bornecrantz <jakob@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

vmwgfx: Add page flip support

Signed-off-by: Jakob Bornecrantz <jakob@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

vmwgfx: Pipe fence out of screen object dirty functions

Signed-off-by: Jakob Bornecrantz <jakob@vmware.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

vmwgfx: Make it possible to get fence from execbuf

Signed-off-by: Jakob Bornecrantz <jakob@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

vmwgfx: Clean up pending event references to struct drm_file objects on close

Pending events may have stale pointer references to struct drm_file objects
after a file has been closed, but before the event is supposed to be
attached to the drm file. Remove such events on file close.

Tested with "modetest".

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jakob Bornecrantz <jakob@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

vmwgfx: Rework fence event action

Signed-off-by: Jakob Bornecrantz <jakob@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/modes: do not enforce an odd vtotal for interlaced modes

CEA actually specifies an interlaced mode with even vtotal and
supplies a diagram showing how this is supposed to work.

Note that interlaced modes with an even vtotal seem to be a fairly
recent invention. All modelines lore I could dig up with googling says
that vtotal for interlaced modes _needs_ to be odd. But the even
modelines in CEA are not a spec-bug, there's a figure in CEA-861-E
called "Figure 5 Special Interlaced Video Format Timing (Even Vtotal)"
that explains how it's supposed to work. Furthermore intel Bspec
explicitly mentions that both odd and even interlaced vtotal are
supported (VTOTAL register in the south display engine of PCH split
chips).

Acked-by: Adam Jackson <ajax@redhat.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/i915: fix up locking inconsistency around gem_do_init

The locking in our setup and teardown paths is rather arbitrary, but
generally we try to protect gem stuff with dev->struct_mutex. Further,
the ums/gem ioctl to setup gem _does_ take the look. So fix up this
benign inconsistency.

Notice while reading through code.

v2: Rebased on top of the ppgtt code.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: enable forcewake voodoo also for gen6

We still have reports of missed irqs even on Sandybridge with the
HWSTAM workaround in place. Testing by the bug reporter gets rid of
them with the forcewake voodoo and no HWSTAM writes.

Because I've slightly botched the rebasing I've left out the ACTHD
readback which is also required to get IVB working. Seems to still
work on the tester's machine, so I think we should go with the more
minmal approach on SNB. Especially since I've only found weak evidence
for holding forcewake while waiting for an interrupt to arrive, but
none for the ACTHD readback.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45181
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45332
Tested-by: Nicolas Kalkhof nkalkhof()at()web.de
Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: fixup seqno allocation logic for lazy_request

Currently we reserve seqnos only when we emit the request to the ring
(by bumping dev_priv->next_seqno), but start using it much earlier for
ring->oustanding_lazy_request. When 2 threads compete for the gpu and
run on two different rings (e.g. ddx on blitter vs. compositor)
hilarity ensued, especially when we get constantly interrupted while
reserving buffers.

Breakage seems to have been introduced in

commit 6f392d548658a17600da7faaf8a5df25ee5f01f6
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Sat Aug 7 11:01:22 2010 +0100

    drm/i915: Use a common seqno for all rings.

This patch fixes up the seqno reservation logic by moving it into
i915_gem_next_request_seqno. The ring->add_request functions now
superflously still return the new seqno through a pointer, that will
be refactored in the next patch.

Note that with this change we now unconditionally allocate a seqno,
even when ->add_request might fail because the rings are full and the
gpu died. But this does not open up a new can of worms because we can
already leave behind an outstanding_request_seqno if e.g. the caller
gets interrupted with a signal while stalling for the gpu in the
eviciton paths. And with the bugfix we only ever have one seqno
allocated per ring (and only that ring), so there are no ordering
issues with multiple outstanding seqnos on the same ring.

v2: Keep i915_gem_get_seqno (but move it to i915_gem.c) to make it
clear that we only have one seqno counter for all rings. Suggested by
Chris Wilson.

v3: As suggested by Chris Wilson use i915_gem_next_request_seqno
instead of ring->oustanding_lazy_request to make the follow-up
refactoring more clearly correct. Also improve the commit message
with issues discussed on irc.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=45181
Tested-by: Nicolas Kalkhof nkalkhof()at()web.de
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: outstanding_lazy_request is a u32

So don't assign it false, that's just confusing ... No functional
change here.

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

drm/i915: check gtfifodbg after possibly failed writes

If we don't have a sufficient number of free entries in the FIFO, we
proceed to do a write anyway. With this check we should have a clue if
that write actually failed or not.

After some discussion with Daniel Vetter regarding his original
complaint, we agreed upon this.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: catch gtfifo errors on forcewake_put

This is similar to a patch I wrote several months ago. It's been updated
for the new FORCEWAKE_MT. As recommended by Chris Wilson, use WARN()
instead of DRM_ERROR, so we can get a backtrace.

This shouldn't impact performance too much as the extra register read
can replace the POSTING_READ we had previously.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: use gtfifodbg

Add register definitions for GTFIFODBG, and clear it during init time to
make sure state is correct.

This register tells us if either a read, or a write occurred while the
fifo was full. It seems like bit 2 is an OR of bit 0 and bit 1, so we
check that as well, but the documents are not quite clear.

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by (v1): Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: set interlaced bits for TRANSCONF

I'm not sure why they are needed (I didn't notice any difference in my
tests), but these bits are in our documentation and they are also set by
the Windows driver.

Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: fixup overlay checks for interlaced modes

The drm core _really_ likes to frob around with the crtc timings and
put halfed vertical timings (in fields) in there. Which confuses the
overlay code, resulting in it's refusal to display anything at the
lower half of an interlaced pipe.

Tested-by: Christopher Egert <cme3000@gmail.com>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: allow interlaced mode output on the HDMI connector

Signed-off-by: Peter Ross <pross@xvid.org>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Christopher Egert <cme3000@gmail.com>
Tested-by: Alfonso Fiore <alfonso.fiore@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: allow interlaced mode output on the SDVO connector

Signed-off-by: Peter Ross <pross@xvid.org>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Christopher Egert <cme3000@gmail.com>
Tested-by: Alfonso Fiore <alfonso.fiore@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: correctly program the VSYNCSHIFT register

The hw seems to use this to correctly insert the required delay
before/after an even/odd interlaced field. This might also explain
why we need to substract 1 half-line from vtotal - if the hw just
adds the delay programmend in VSYNCSHIFT the total frame time would be
about that too long.

These registers seems to only exist on gen4 and later. For paranoia
also program it to 0 for progressive modes, but according to
documentation the hw should just ignore it in this case.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Alfonso Fiore <alfonso.fiore@gmail.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: don't allow interlaced pipeconf on gen2

gen2 doesn't support it, so be a bit more paranoid and add a check to
ensure that we never ever set an unsupported interlaced bit.

Ensure that userspace can't set an interlaced mode by resetting
interlace_allowed for the crt on gen2. dvo and lvds are the only other
encoders that gen2 supports and these already disallow interlaced
modes.

Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: fixup interlaced support on ilk+

According to Paulo Zanoni, this is what windows does.

Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Alfonso Fiore <alfonso.fiore@gmail.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: fixup interlaced vertical timings confusion, part 2

According to bspec, we need to subtract an additional line from vtotal
for interlaced modes and vblank_end needs to equal vtotal. All other
timing fields do not need this special treatment, so kill it.

Bspec says that this is irrespective of whether the interlaced mode
has an odd or even vtotal, both modes are supported.

Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Christopher Egert <cme3000@gmail.com>
Tested-by: Alfonso Fiore <alfonso.fiore@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: fixup interlaced vertical timings confusion, part 1

We have a pretty decent confusion about vertical timings of interlaced
modes. Peter Ross has written a patch that makes interlace modes work
on a lot more platforms/output combinations by doubling the vertical
timings.

The issue with that patch is that core drm _does_ support specifying
whether we want these vertical timings in fields or frames, we just
haven't managed to consistently use this facility. The relavant
function is drm_mode_set_crtcinfo, which fills in the crtc timing
information.

The first thing to note is that the drm core keeps interlaced modes in
frames, but displays modelines in fields. So when the crtc modeset
helper copies over the mode into adjusted_mode it will already contain
vertical timings in half-frames. The result is that the fixup code in
intel_crtc_mode_fixup doesn't actually do anything (in most cases at
least).

Now gen3+ natively supports interlaced modes and wants the vertical
timings in frames. Which is what sdvo already fixes up, at least under
some conditions.

There are a few other place that demand vertical timings in fields
but never actually deal with interlaced modes, so use frame timings
for consistency, too. These are:
- lvds panel,
- dvo encoders - dvo is the only way gen2 could support interlaced
  mode, but currently we don't support any encoders that do.
- tv out - despite that the tv dac sends out an interlaced signal it
  expects a progressive mode pipe configuration.
All these encoders enforce progressive modes by resetting
interlace_allowed.

Hence we always want crtc vertical timings in frames. Enforce this in
our crtc mode_fixup function and rip out any redudant timing
computations from the encoders' mode_fixup function.

v2-4: Adjust the vertical timings a bit.

v5: Split out the 'subtract-one for interlaced' fixes.

v6: Clarify issues around tv-out and gen2.

Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Christopher Egert <cme3000@gmail.com>
Tested-by: Alfonso Fiore <alfonso.fiore@gmail.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: clean up interlaced pipeconf bit definitions

- Clarify which bits are for which chips.
- Note that gen2 can't do interlaced directly (only via dvo tv chips).
- Move the mask to the top to make it clearer how wide this field is.
- Add defintions for all possible values.

This patch doesn't change any code.

v2: Paulo Zanoni pointed out that the pixel doubling modes do no
longer exist on ivb.

Cc: Peter Ross <pross@xvid.org>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Tested-by: Christopher Egert <cme3000@gmail.com>
Tested-by: Alfonso Fiore <alfonso.fiore@gmail.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Merge remote-tracking branch 'airlied/drm-fixes' into drm-intel-next-queued

Back-merge from drm-fixes into drm-intel-next to sort out two things:

- interlaced support: -fixes contains a bugfix to correctly clear
  interlaced configuration bits in case the bios sets up an interlaced
  mode and we want to set up the progressive mode (current kernels
  don't support interlaced). The actual feature work to support
  interlaced depends upon (and conflicts with) this bugfix.

- forcewake voodoo to workaround missed IRQ issues: -fixes only enabled
  this for ivybridge, but some recent bug reports indicate that we
  need this on Sandybridge, too. But in a slightly different flavour
  and with other fixes and reworks on top. Additionally there are some
  forcewake cleanup patches heading to -next that would conflict with
  currrent -fixes.

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

Merge branch 'drm-intel-fixes' of git://git./linux/kernel/git/keithp/linux into drm-fixes

* 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/keithp/linux:
  drm/i915: fixup interlaced bits clearing in PIPECONF on PCH_SPLIT (v2)
  drm/i915: no lvds quirk for AOpen MP45
  drm/i915: Force explicit bpp selection for intel_dp_link_required
  drm/i915: fixup interlaced bits clearing in PIPECONF on PCH_SPLIT
  drm/i915:: Disable FBC on SandyBridge

drm/i915: enable ppgtt

We want to unconditionally enable ppgtt for two reasons:
- Windows uses this on snb and later.
- We need the basic hw support to work before we can think about real
  per-process address spaces and other cool features we want.

But Chris Wilson was complaining all over irc and intel-gfx that this
will blow up if we don't have a module option to disable it. Hence add
one, to prevent this.

ppgtt support seems to slightly change the timings and make crashy
things slightly more or less crashy. Now in my testing and the testing
this got on troublesome snb machines, it seems to have improved things
only. But on ivb it makes quite a few crashes happen much more often,
see

https://bugs.freedesktop.org/show_bug.cgi?id=41353

Luckily Eugeni Dodonov seems to have a set of workarounds that fix
this issue.

v2: Don't try to enable ppgtt on pre-snb.

v3: Pimp commit message and make Chris Wilson less grumpy by adding a
module option.

v4: New try at making Chris Wilson happy.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: ppgtt debugfs info

This was pretty usefull for debugging, might be useful for diagnosing
issues.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: ppgtt register definitions

Split out for easier cross-checking of the boring pieces with bspec.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: ppgtt binding/unbinding support

This adds support to bind/unbind objects and wires it up. Objects are
only put into the ppgtt when necessary, i.e. at execbuf time.

Objects are still unconditionally put into the global gtt.

v2: Kill the quick hack and explicitly pass cache_level to ppgtt_bind
like for the global gtt function. Noticed by Chris Wilson.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drm/i915: initialization/teardown for the aliasing ppgtt

This just adds the setup and teardown code for the ppgtt PDE and the
last-level pagetables, which are fixed for the entire lifetime, at
least for the moment.

v2: Kill the stray debug printk noted by and improve the pte
definitions as suggested by Chris Wilson.

v3: Clean up the aperture stealing code as noted by Ben Widawsky.

v4: Paint the init code in a more pleasing colour as suggest by Chris
Wilson.

v5: Explain the magic numbers noticed by Ben Widawsky.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

agp/intel-gtt: export the gtt pagetable iomapping

We need this because ppgtt page directory entries need to be in the
global gtt pagetable.

Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com>
Signed-Off-by: Daniel Vetter <daniel.vetter@ffwll.ch>