ums-gem code correctly cancels the retire work (at lastclose time),
kms does not do so. Fix this by canceling the work right after ideling
the gpu.

While staring at the code I noticed that the work function is not
static. Fix this, too.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

When the module unloads, all users should be gone, hence all bo references
held by userspace, too. This should already result in an idle ringbuffer.
Still, be paranoid and idle gem before starting the unload dance.

Also kill the call to i915_gem_lastclose under an if (kms), it's a noop
for kms.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Kill any outstanding unpin_work when destroying the corresponding
crtc. Then flush the workqueue before the gem teardown, in case
any unpin work is still outstanding.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

idle_work wasn't cleaned up at all. It takes &dev->struct_mutex, but
accesss the mode_config crtc list (without any other locking!). Hence
this work needs to be canceled before calling drm_mode_config_cleanup.

As evidenced by the kernel's object debuggin code, the current code
also cleans up the timer to early (it gets rearmed). So move it right
before the final cleanup (it seems to work).

Also unconditionally set up the idle_timer in intel_increase_pllclock.
If we're unlucky the timer might fire right away, rendering the call
in the modesetting teardown pointless.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

With kms, interrupts now get disabled in the modesetting cleanup. So
free the error state afterwards, it currently gets allocated in
the interrupt handler.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

hotplug_work is queued by the hotplug interrupt and only either emits
a hotplug uevent or queues a crt poll slow-work. No other locking.  So
it's safe to cancel this work _after_ irq's have been turned off.  But
before the modesetting objects are destroyed because the hotplug
function accesses them (without locking).

The current code (for kms) only switches irqs off after modesetting
teardown, hence move the irq teardown into the modeset cleanup right
before the crtc cleanup.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

This is the first patch to clean up module unload races due to
outstanding timers/work. Preparatory step: Thou shalt not destroy
the workqueue when new work might still get enqued.

Now error_work gets queued by the hangcheck timer and only (atomically)
reads the chip wedged status. So cancel it right after the hangcheck
timer is killed. But the hangcheck is armed by interrupts, so move
everything after irqs are disabled.

Also change a del_timer to a del_timer_sync in the ums gem code, the
hangcheck timer is self-rearming.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

struct intel_dp contains both struct intel_encoder at the beginning (as
it's base-class) and an i2c adapater. When initializing, the i2c adapter
gets assigned

	intel_encoder->ddc_adaptor = &intel_dp->adapter

and the generic intel_encode_destroy happily calls kfree on this pointer.
Ouch. Fix this by using a dp specific cleanup function.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

This reverts commit b9421ae8.

This warning was so prelevant, even for apparently working machines,
that it was just causing fear, anxiety and panic.

The root cause still remains, so we will add some better debugging when
we focus on fixing it.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=17021Reported-by: NMaciej Rutecki <maciej.rutecki@gmail.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

This reverts commit 0f3ee801.

Enabling LVDS on pipe A was causing excessive wakeups on otherwise idle
systems due to i915 interrupts. So restrict the LVDS to pipe B once more,
whilst the issue is properly diagnosed.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=16307Reported-and-tested-by: NEnrico Bandiello <enban@postal.uv.es>
Poked-by: NFlorian Mickler <florian@mickler.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Adam Jackson <ajax@redhat.com>
Cc: stable@kernel.org

This reverts commit ce171780.

This commit has been independently bisected a few times as being the cause
of a s2ram failure.
Reported-and-tested-by: NKyle McMartin <kyle@mcmartin.ca>
Reported-and-tested-by: NAndy Isaacson <adi@hexapodia.org>
Cc: Zou Nan hai <nanhai.zou@intel.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

New pci ids for GT2 and GT2+ on desktop and mobile sandybridge,
and graphics device ids for server sandybridge. Also rename original
ids string to reflect GT1 version.
Signed-off-by: NZhenyu Wang <zhenyuw@linux.intel.com>
Cc: stable@kernel.org
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

MI_FLUSH is being deprecated, but still available on Sandybridge.
Make sure it's enabled as userspace still uses MI_FLUSH.
Signed-off-by: NZhenyu Wang <zhenyuw@linux.intel.com>
Cc: stable@kernel.org
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Sandybridge GTT has new cache control bits in PTE, which controls
graphics page cache in LLC or LLC/MLC, so we need to extend the mask
function to respect the new bits.

And set cache control to always LLC only by default on Gen6.
Signed-off-by: NZhenyu Wang <zhenyuw@linux.intel.com>
Cc: stable@kernel.org
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Signed-off-by: NZhenyu Wang <zhenyuw@linux.intel.com>
Cc: stable@kernel.org
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

It should shift bit 39-32 into pte's bit 11-4.
Reported-by: NTakashi Iwai <tiwai@suse.de>
Signed-off-by: NZhenyu Wang <zhenyuw@linux.intel.com>
Cc: stable@kernel.org
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Arguably this is a bug in drm-core in that we should not be called twice
in succession with DPMS_ON, however this is still occuring and we see
FDI link training failures on the second call leading to the occassional
blank display. For the time being ignore the repeated call.

Original patch by Dave Airlie <airlied@redhat.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

We were passing garbage values into the panel-fitter control register
when disabling it on Ironlake - those values (filter modes and reserved
MBZ bits) would have then be re-used the next time panel-fitting was
enabled.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

When we miss the flip prepare interrupt, we never get into the
software state needed to restart userspace, resulting in a freeze of a
full-screen OpenGL application (such as a compositor).

Work around this by checking DSPxSURF/DSPxBASE to see if the page flip
has actually happened. If it has, do the work we would have done when
the flip prepare interrupt comes in.

Also, add debugfs information to tell us what's going on (based on the
patch from Chris Wilson attached to bugs.fdo bug #29798).
Signed-off-by: NSimon Farnsworth <simon.farnsworth@onelan.co.uk>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

We reset intel_encoder for every matching encoder whilst iterating over
the encoders attached to this crtc when changing mode. As such in a
cloned configuration intel_encoder may not correspond to the correct
is_edp encoder.

By scoping intel_encoder to the loop, not only is the compiler able to
spot this mistake, we also improve readiability for ourselves.
[It might not be a mistake, within this function it is unclear as to
whether it is permissable for eDP to be cloned...]
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

We were failing when trying to allocate the resource for MMIO of the
MCHBAR because we forgot to specify what type of resource we wanted.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: stable@kernel.org
Reviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>

Only stop trying if the aux channel sucessfully reports that the
transmission was completed, otherwise try again. On the 5th failure,
bail and report that something is amiss.

This fixes a sporadic failure in reading the EDID for my external panel
over DP.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

If the VBIOS tells us the mapping of the SDVO device onto the DDC bus,
use it. However, if there is no VBIOS available that mapping is
uninitialised and we should fallback to our earlier guess.

Fix regression introduced in b1083333 (which in turn is a fix for the
regression caused by the introduction of this guess, 14571b4c).

References:

  Bug 29499 - [945GM] Screen disconnected because of missing VBIOS
  https://bugs.freedesktop.org/show_bug.cgi?id=29499

  Bug 15109 - i945GM fails to detect EDID on DVI port
  https://bugzilla.kernel.org/show_bug.cgi?id=15109Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reported-and-tested-by: NPaul Neumann <paul104x@yahoo.de>
Cc: Adam Jackson <ajax@redhat.com>
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Cc: stable@kernel.org

Adam Hill reported that his Arrandale system required a much longer, up
to 200x500us, wait for the panel to initialise or else modesetting would
fail.

References:

  https://bugs.freedesktop.org/show_bug.cgi?id=29141Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reported-and-tested-by: NAdam Hill <sidepipeuk@yahoo.co.uk>

Lets try to avoid repeating old bugs.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>

i965 uses the Display Registers to compute the offset from the display
base so the new base does not need adjusting when flipping. The older
chipsets use a fence to access the display and so do perceive the
surface as linear and have a single base register which is reprogrammed
using the flip.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Reported-by: NMarty Jack <martyj19@comcast.net>
Reviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>

To simplify the IS_GEN[234] macros and to enable switching.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

copy_to_user() returns the number of bytes remaining to be copied and
I'm pretty sure we want to return a negative error code here.
Signed-off-by: NDan Carpenter <error27@gmail.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

copy_to_user returns the number of bytes remaining to be copied, but we
want to return a negative error code here.  These are returned to
userspace.
Signed-off-by: NDan Carpenter <error27@gmail.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

Make sure we always detect when we fail to correctly allocate the Isoch
Flush Page and print an error to warn the user about the likely memory
corruption that will result in invalid rendering or worse.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

So set the coherent dma mask accordingly. This dma mask is only used
for physical objects, so it won't really matter allocation-wise.

Now this never really surfaced because sane 32bit kernels only have 1G
of lowmem. But some eager testers (distros?) still carry around the patch
to adjust lowmem via a kconfig option. And the kernel seems to favour
high allocations on boot-up, hence the overlay blowing up reliably.

Because the patch is tiny and nicely shows how broken gen2 is it's imho
worth to merge despite the fact that mucking around with the lowmem/
highmem division is (no longer) supported.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=28318
Cc: stable@kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

The vblank status bit is a sticky bit that must be cleared with a write
of '1' prior to polling for the next vblank.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Tested-by: NSitsofe Wheeler <sitsofe@yahoo.com>

jbarnes: I'd still rather see a lock, but I think you're right that
we don't generally wait in code that needs not to miss an interrupt.
Reviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>

Partial revert of 9d0498a2.
Signed-off-by: NPekka Enberg <penberg@kernel.org>
Tested-by: NHugh Dickins <hughd@google.com>
Tested-by: NSven Joachim <svenjoac@gmx.de>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

This fixes blur-like screen corruption on the following card:

VGA compatible controller [0300]: Intel Corporation 82G33/G31 Express
Integrated Graphics Controller [8086:29c2] (rev 10)

intel_sdvo_mode_set() should not return prematurely just because some
features are not supported.

https://bugzilla.kernel.org/show_bug.cgi?id=17151Signed-off-by: NPavel Roskin <proski@gnu.org>
Reported-by: NJonathan Corbet <corbet@lwn.net>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
[ickle: Relax a couple more checks for failing LVDS modesetting]
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

This reverts commit 86f100b1.

The kref API requires the handlecount to be initialised to one on object
creation (so that kref_get() doesn't complain upon first use) so the
dalliance in the drivers is required in order to sink the initial
floating reference.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: stable@kernel.org

drivers/gpu/drm/i915/intel_overlay.c: In function 'intel_overlay_print_error_state':

drivers/gpu/drm/i915/intel_overlay.c:1467: error: implicit declaration of function 'seq_printf'

Addresses https://bugzilla.kernel.org/show_bug.cgi?id=16811Reported-by: NMartin Ziegler <ziegler@uni-freiburg.de>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: Dave Airlie <airlied@linux.ie>
Cc: Andre Muller <andremuellerster@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Every driver used the default implementation. Fold that one into
the only callsite and drop the callback.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NDave Airlie <airlied@redhat.com>

Totally unused.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NDave Airlie <airlied@redhat.com>

There's no point in jumping through two indirections. So kill one
and call the kernels agp functions directly.
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NDave Airlie <airlied@redhat.com>