If the interrupt handler were to process a previous vblank interrupt and
the following flip pending interrupt at the same time, the page flip
would be completed too soon.

To eliminate this race, check the live pending flip status from the ISR
register before finishing the page flip.

v2: Added a comment explaining the logic (by Chris Wilson)
v3: Fix a typo in the comment
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Tested-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We still rely on a few LVDS bits, but restoring the enable bit can cause
trouble at this point, so don't.

v2: use the right mask to prevent restore (Daniel)
    conditionalize on KMS support (Denial)
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Otherwise, if the BIOS did anything wrong, our first I915_{WRITE,READ}
will give us "unclaimed register"  messages.

V2: Even earlier.
V3: Move it to intel_early_sanitize_regs.

Bugzilla: http://bugs.freedesktop.org/show_bug.cgi?id=58897Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We plan to treat GEN7_ERR_INT as an interrupt, so use this register
for the checks inside I915_WRITE. This way we can have the best of
both worlds: the error message with a register address and the

V2: Split in 2 patches: one for the macro, one for changing the
register, as requested by Ben.
V3: Rebase.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This avoids polluting i915_write##x and also allows us to reuse code
on i915_read##x.

v2: Rebase
v3: Convert the macros to static functions
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Some (but not all) of the HDMI registers can be used to control sDVO,
so those registers have two names. IMHO, when we're talking about
HDMI, we really should call the HDMI control register "hdmi_reg"
instead of "sdvox_reg", otherwise we'll just confuse people reading
our code (we now have platforms with HDMI but without SDVO). So now
"struct intel_hdmi" has a member called "hdmi_reg" instead of
"sdvox_reg".

Also, don't worry: "struct intel_sdvo" still has a member called
"sdvo_reg".

v2: Rebase (v1 was sent in May 2012).
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This way we can remove some duplicated code and avoid more mistakes
and regressions with these registers in the future.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

So use msecs_to_jiffies(10) to make the timeout the same as in the
"!has_aux_irq" case.

This patch was initially written by Daniel Vetter and posted on
pastebin a few weeks ago. I'm just bringing it to the mailing list.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Since basically every code called on these places comes from
intel_ddi.c
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

It is customary to end sysfs attributes with a newline.
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

GPU reset will drop all flips that are still in the ring. So after the
reset, call update_plane() for all CRTCs to make sure the primary
planes are scanning out from the correct buffer.

Also finish all pending flips. That means user space will get its
page flip events and won't get stuck waiting for them.

v2: Explicitly finish page flips instead of relying on FLIP_DONE
    interrupt being generated by the base address update.
v3: Make two loops over crtcs to avoid deadlocks with the crtc mutex
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
[danvet: Fixup long line complaint from checkpatch.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Since obj->pending_flips was never set, intel_pipe_set_base() never
actually waited for pending page flips to complete.

We really do want to wait for the pending flips, because otherwise the
mmio surface base address update could overtake the flip, and you
could end up with an old frame on the screen once the flip really
completes.

Just call intel_crtc_wait_pending_flips() prior to calling
intel_pipe_set_base() instead of calling just intel_finish_fb()
from intel_pipe_set_base(). Moving the call outside of
intel_pipe_set_base() avoids calling it twice from the full
modeset path.

v2: Wait for pending flips w/o holding struct_mutex
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We already managed to get it out of sync (Haswell has been promoted out
of this option), so let's remove all mentions to platforms.
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Use the new PM routines to indicate whether we need to VT switch at suspend
and resume time.  When a new driver is bound, set its flag accordingly,
and when unbound, remove it from the PM's console tracking list.
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Acked-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We already have the quirk entry for the mobile platform, but also
reports on some desktop versions. So be paranoid and set it
everywhere.

References: http://www.mail-archive.com/dri-devel@lists.freedesktop.org/msg33138.html
Cc: stable@vger.kernel.org
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: "Sankaran, Rajesh" <rajesh.sankaran@intel.com>
Reported-and-tested-by: NMihai Moldovan <ionic@ionic.de>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Whilst IOMMU is enabled for the Intel GPU on Ironlake, it appears that
using WC writes to update the PTE on the GPU fails miserably. The
result looks like the majority of the writes do not land leading to
lots of screen corruption and a hard system hang.

v2: s/</<=/ to preserve the current exclusion of Sandybridge
Reported-by: NNathan Myers <ncm@cantrip.org>
Bugzilla: https://bugzilla.freedesktop.org/show_bug.cgi?id=60391Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Tested-by: NNathan Myers <ncm@cantrip.org>
[danvet: Remove cc: stable and add tested-by.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

HSW no longer has the PIPECONF bit for limited range RGB output.
Instead the pipe CSC unit must be used to perform that task.

The CSC pre offset are set to 0, since the incoming data is full
[0:255] range RGB, the coefficients are programmed to compress the
data into [0:219] range, and then we use either the CSC_MODE black
screen offset bit, or the CSC post offsets to shift the data to
the correct [16:235] range.

Also have to change the confiuration of all planes so that the
data is sent through the pipe CSC unit. For simplicity send the
plane data through the pipe CSC unit always, and in case full
range output is requested, the pipe CSC unit is set up with an
identity transform to pass the plane data through unchanged.

I've been told by some hardware people that the use of the pipe
CSC unit shouldn't result in any measurable increase in power
consumption numbers.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=53881
Cc: stable@vger.kernel.org
Cc: Jani Nikula <jani.nikula@intel.com>
Tested-by: NJani Monoses <jani@ubuntu.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Print out the HW context object information per ring. Even though the
existing code only utilizes the render ring, it's simple enough to
support future expansion.

I had this in a patch somewhere in a rev of the original implementation,
but I must have lost it.
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
[danvet: s/context/default context/ bikeshed applied.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Yet another remnant ... this might explain why l3 remapping didn't
really work on HSW.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=57441Spotted-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Cc: stable@vger.kernel.org
Reviewed-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The bit controlling whether PIPE_CONTROL DW/QW write targets
the global GTT or PPGTT moved moved from DW 2 bit 2 to
DW 1 bit 24 on IVB.

I verified on IVB that the fix is in fact effective. Without the fix
none of the scratch writes actually landed in the pipe control page.
With the fix the writes show up correctly.

v2: move PIPE_CONTROL_GLOBAL_GTT_IVB setup to where other flags are set
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The Intel PRM says the M1 and M2 divisors must be in the range of 10-20 and 5-9.
Since we do all calculations based on them being register values (which are
subtracted by 2) we need to specify them accordingly.
Signed-off-by: NPatrik Jakobsson <patrik.r.jakobsson@gmail.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=56359
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The Intel PRM says the M1 and M2 divisors must be in the range of 10-20 and 5-9.
Since we do all calculations based on them being register values (which are
subtracted by 2) we need to specify them accordingly.
Signed-off-by: NPatrik Jakobsson <patrik.r.jakobsson@gmail.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Similarly to:

  commit 6a0d1df3d3a0d2370541164eb0595fe35dcd6de3
  Author: Damien Lespiau <damien.lespiau@intel.com>
  Date:   Tue Dec 11 15:18:28 2012 +0000

      drm/i915: Preserve the FDI line reversal override bit on CPT

DDI port support lane reversal to easy the PCB layouting work. Let's
preserve the bit configured by the BIOS (until we find how to correctly
retrieve the information from the VBT, but this does sound more fragile
then just relying on the BIOS that has, hopefully, been validated
already.
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The FDI link has supported link reversal to make the PCB layout
engineer's life easier for quite a while and we have always presered
this bit as we programmed FDI_RX_CTL with a read/modify/write sequence.

We're trying to take a bit more control over what the BIOS leaves in
various register and with the introduction of DDI, started to program
FDI_RX_CTL fully.

There's a fused bit to indicate DMI link reversal and FDI defaults to
mirroring that configuration. We have a bit to override that behaviour
that we need to preserve from the BIOS.
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Reviewed-by: NRodrigo Vivi <rodrigo.vivi@gmail.com>
Reviewed-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Amending
commit 4518f611
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Wed Jan 23 16:16:35 2013 +0100

    drm/i915: dump UTS_RELEASE into the error_state

CC: stable@vger.kernel.org
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Support for real RGB332 is a rarity, most hardware only really support
C8. So use C8 instead of RGB332 when determining the format based on
depth/bpp.

This fixes 8bpp fbcon on i915, since i915 will only accept C8 and not
RGB332.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=59572Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Tested-by: mlsemon35@gmail.com
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Set depth/bits_per_pixel to 8 for C8 format.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: NDave Airlie <airlied@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

As explained by Chris Wilson gem objects in stolen memory are always
coherent with the GPU so we don't need to ever flush the CPU caches for
these.

This fixes a breakage - at least with the compact sg patches applied -
during the resume/restore gtt mappings path, when we tried to clflush an
FB object in stolen memory, but since stolen objects don't have backing
pages we passed an invalid page pointer to drm_clflush_page().
Signed-off-by: NImre Deak <imre.deak@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

If a GPU reset occurs while a page flip has been submitted to the ring,
the flip will never complete once the ring has been reset.

The GPU reset can be detected by sampling the reset_counter before the
flip is submitted, and then while waiting for the flip, the sampled
counter is compared with the current reset_counter value.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
[danvet: Move the reset_counter assignment to an earlier place in
common code as discussed on the mailing list.]
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=60140Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The pending flip mask no longer set anywhere, so trying to wait for
while it's non-zero is a no-op. Remove it completely.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

This has been lost in the locking rework for intel_alloc_context_page:

commit 2c34b850
Author: Ben Widawsky <ben@bwidawsk.net>
Date:   Sat Mar 19 18:14:26 2011 -0700

    drm/i915: fix ilk rc6 teardown locking

Cc: Ben Widawsky <ben@bwidawsk.net>
Reviewed-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

We already print the HWS addresses during init, so do the same for the
pipe control page. Reduces guesswork when looking at hex addresses
later.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

With the previous patch "drm/i915: disable shared panel fitter for
pipe" we now disable the panel fitter at the right spot in the modeset
sequence in the crtc functions on all platforms. Hence the disabling
in intel_disable_lvds is no longer required and potentially harmful
(since the plane is still enabled at this point).

Similarly on the enabling side we enable the panel fitter in the lvds
callback only once the plane is enabled. Which is too late. Hence move
this into a new intel_pre_enable_lvds callback.

Finally we can ditch lvds_encoder->pfit_dirty - this was required to
work around the crtc helper semantics, but with the new i915 modeset
infrastructure we should enable/disable the pfit only when enabling or
disabling the entire output pipeline. So separate state tracking for
the pfit is no longer required.
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
[danvet: Bikeshed the commit message a bit to stress that now we
enable/disable the pfit on i9xx platforms at the right point of time
compared to the old code.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

If encoder is switched off by BIOS, but the panel fitter is left on,
we never try to turn off the panel fitter and leave it still attached
to the pipe - which can cause blurry output elsewhere.

Based on work by Chris Wilson <chris@chris-wilson.co.uk>

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=58867
Cc: stable@vger.kernel.org
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Tested-by: NAndreas Sturmlechner <andreas.sturmlechner@gmail.com>
[danvet: Remove the redundant HAS_PCH_SPLIT check and add a tiny
comment.]
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Some early bios versions seem to ship with the wrong tuning values for
the MCH, possible resulting in pipe underruns under load. Especially
on DP outputs this can lead to black screen, since DP really doesn't
like an occasional whack from an underrun.

Unfortunately the registers seem to be locked after boot, so the only
thing we can do is politely point out issues and suggest a BIOS
upgrade.

Arthur Runyan pointed us at this issue while discussion DP bugs - thus
far no confirmation from a bug report yet that it helps. But at least
some of my machines here have wrong values, so this might be useful in
understanding bug reports.

v2: After a bit more discussion with Art and Ben we've decided to only
the check the watermark values, since the OREF ones could be be a
notch more aggressive on certain machines.

Cc: Ben Widawsky <ben@bwidawsk.net>
Cc: Runyan, Arthur J <arthur.j.runyan@intel.com>
Reviewed-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

It's not that the assertion is incorrect, but rather that we can call
do_destroy early in loading, and we will falsely BUG().

Since contexts have been in for a while now, and in the internal APIs
are pretty stable, it should be fairly safe to remove this.

v2: Remove unused dev_priv, and dev
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The ring initialization will differ a bit in upcoming generations, and
this split will prepare the code for what's needed.

This patch also fixes a bug introduced in:
commit 99433931
Author: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date:   Tue Jan 22 14:12:17 2013 +0200

    drm/i915: use gem_set_seqno() on hardware init

After doing the extraction, the bad error handling became obvious.  I
acknowledge that this should be two patches, but it's a pretty
small/trivial patch. If requested, I can certainly do the fix as a
distinct patch.

v2: Should be cleanup blt, not init blt on failure (Chris)

v3: Forgot to git add on v2

Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NBen Widawsky <ben@bwidawsk.net>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>