Doing the IBX transcoder B workaround causes underruns on
pipe/transcoder A. Just hide them by disabling underrun reporting for
pipe A around the workaround.

It might be possible to avoid the underruns by moving the workaround
to be applied only when enabling pipe A. But I was too lazy to try it
right now, and the current method has been proven to work, so didn't
want to change it too hastily.

Note that this can re-enable underrun reporting on pipe A if was
already disabled due to a previous actual underrun. But that's OK, we
may just get a second underrun report if another real underron occurrs
on pipe A.

v2: Note that pipe A underruns can get re-enabled due to this (Jani)
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v1)
Link: http://patchwork.freedesktop.org/patch/msgid/1446225802-11180-1-git-send-email-ville.syrjala@linux.intel.com

ironlake_enaable_pch_transcoder() checks for CPT to see if it should
enable the timing override chicken bit, but
ironlake_disable_pch_transcoder() checks for !IBX to see if it should
clear the same bit. Change ironlake_disable_pch_transcoder() to check
for CPT as well to keep the two sides consistent.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446146763-31821-8-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>

Due to the shared error interrupt on IVB/HSW and CPT/PPT we may not
always get an interrupt on a FIFO underrun. But we can always do an
explicit check (like we do on GMCH platforms that have no underrun
interrupt).

v2: Drop stale kerneldoc for i9xx_check_fifo_underruns() (Daniel)
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1446225741-11070-1-git-send-email-ville.syrjala@linux.intel.com

Some hardware (IVB/HSW and CPT/PPT) have a shared error interrupt for
all the relevant underrun bits, so in order to keep the error interrupt
enabled, we need to have underrun reporting enabled on all PCH
transocders. Currently we leave the underrun reporting disabled when
the pipe is off, which means we won't get any underrun interrupts
when only a subset of the pipes are active.

Fix the problem by re-enabling the underrun reporting after the pipe has
been disabled. And to avoid the spurious underruns during pipe enable,
disable the underrun reporting before embarking on the pipe enable
sequence. So this way we have the error reporting disabled while
running through the modeset sequence.

v2: Re-enable PCH FIFO underrun reporting unconditionally on pre-HSW
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v1)
Link: http://patchwork.freedesktop.org/patch/msgid/1446225691-10928-1-git-send-email-ville.syrjala@linux.intel.com

As we did for ILK/SNB/IVB, move the PCH FIFO underrun enable to happen
after the encoder enable on HSW+. And again, for symmetry, move the
the disable to happen before encoder disable.

I've left out the vblank wait before the enable here because I don't
know if it's needed or not. Actually I don't know if this entire
change is needed as I don't have a HSW/BDW with VGA output.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446146763-31821-5-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>

We get spurious PCH FIFO underruns if we enable the reporting too soon
after enabling the crtc. Move it to be the last step, after the encoder
enable. Additionally we need an extra vblank wait, otherwise we still
get the underruns. Presumably the pipe/fdi isn't yet fully up and running
otherwise.

For symmetry, disable the PCH underrun reporting as the first thing,
just before encoder disable, when shutting down the crtc.

v2: Do the PCH underrun enable unconditionally (Jani, Daniel)
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org> (v1)
Link: http://patchwork.freedesktop.org/patch/msgid/1446225627-10809-1-git-send-email-ville.syrjala@linux.intel.com

Rather than looking at crtc->mode (which is the user mode) dig up the
sync polarity settings from the adjusted_mode when programming
TRANS_DP_CTL on CPT/PPT.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446146763-31821-3-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>

No point in doing the crtc->pipe->crtc->config->cpu_transcoder dance
when we can just do crtc->config->cpu_transcoder.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446146763-31821-2-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>

From our maintainer Daniel Vetter a few days ago:
  "Oh dear this is dead code. kdbg uses the fbcon, which always uses
  untiled, which means fbc will never be enabled. Also we have 0 users
  and 0 test coverage for kdbg on top of i915 (Jesse implemented it
  for fun years back). Imo just remove all this code."

Adding to what Daniel said: for kgdboc's KMS support,
intel_pipe_set_base_atomic() already manually disables FBC, so we
won't do the in_dbg_master() check there. This is essentially a revert
of:

commit c924b934
Author: Jason Wessel <jason.wessel@windriver.com>
Date:   Thu Aug 5 09:22:32 2010 -0500
    i915: when kgdb is active display compression should be off

Besides, it is not clear what is the exact problem caused by FBC, and
why other features such as PSR, DRRS, IPS and RPM are not also
checking for in_dbg_master(). IMHO we should either remove the code as
suggested by Daniel or we add some nice comments explaining why is FBC
so special.

v2: Rebase due to new patch order.

Cc: Jason Wessel <jason.wessel@windriver.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-13-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

Daniel was looking at this code and asked about whether fb->pitches[0]
is correct, then he suggested we should a comment to make sure it is
actually intentional.

For more information on the CFB size calculation, please see the
commit message of:

commit c4ffd409
Author: Paulo Zanoni <paulo.r.zanoni@intel.com>
Date:   Thu Oct 1 19:55:57 2015 -0300
    drm/i915: fix CFB size calculation
Requested-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-12-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

If we run igt/kms_frontbuffer_tracking, this message will appear
thousands of times, eating a significant part of our dmesg buffer.
It's part of the expected FBC behavior, so let's just silence it.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-10-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

Make sure we deactivate FBC at intel_fbc_init(), so we can remove the
call from intel_display.c. Currently we only have the "enabled"
software state, but later we'll have both "enabled" and "active", and
we'll add assertions to them, so just calling intel_fbc_disable() from
intel_modeset_init() won't work. It's better to make sure
intel_fbc_init() already puts the hardware in the expected state, so
we can put nice assertions in the other functions.

v2: Keep/improve the comment (Chris).
v3: Improve the commit message a little bit.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-9-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

If FBC is disabled we will still call intel_fbc_invalidate(), and as a
result we may call intel_fbc_deactivate(), which will try to touch
registers.

I'm pretty sure I saw this happen on a runtime suspended device, and
I'm almost sure I was running igt/pm_rpm. It produced the "you touched
registers while the device is suspended" WARNs. But this was some time
ago and I can't remember exactly which conditions were necessary to
reproduce the problem.

v2: Rebase to new series order.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-8-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

Don't try to list in comments the cases where we should enable or
disable FBC: it varies a lot with the hardware generations and the
code should be the documentation. Also notice that there's already a
huge gap between the comments and what's in the code.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>

Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-7-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

This change was part of the commit that makes intel_fbc_update()
receive an intel_crtc as argument instead of dev_priv, but since it
was polluting the diff with too many chunks I decided to move it to
its own commit.

It seems that our developers are favoring having this instead of the
old combination drm_crtc *crtc + intel_crtc *intel_crtc, and on the
mentioned commit we'll get rid of the drm_crtc variable, so let's do
an intermediate commit with the rename, so on the next commit we'll
have just struct intel_crtc *crtc.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-6-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

We're going to kill intel_fbc_find_crtc(), that's why a big part of
the logic moved from intel_fbc_find_crtc() to crtc_is_valid().

v2:
  - Rebase due to pipe_a_only change.
  - Split the multiline conditional (Chris).
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-5-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

We already check if the CRTC is visible, and it shouldn't be possible
to have a visible CRTC without an FB.

This was noticed by both Chris and Ville on different ocasions.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-4-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

Make the code easier to read.
Suggested-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-3-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

Although the term "nuke" is part of the FBC spec, it's not very
intuitive, so let's rename it to make it easier for people that are
not familiar with the spec.
Requested-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-2-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

Newlines are not needed and they're not used by the other messages. I
added the newline by mistake.
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446664257-32012-14-git-send-email-paulo.r.zanoni@intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

The bspec indicates that DDI A using four lanes is the only valid
configuration for Broxton (Broxton doesn't have a DDI E to split these
lanes with); the DDI_A_4_LANES bit of port A's DDI_BUF_CTL should always
be set by the BIOS.  However some BIOS versions seem to only be setting
this bit if eDP is actually lit up at boot time; if the BIOS doesn't
turn on the eDP panel because an external display is plugged in, then
this bit is never properly initialized.  The end result of this is that
we wind up calculating a lower max data rate than we should and may wind
up rejecting the native mode for panels that we should be able to drive.

Let's workaround this BIOS bug by just turning the DDI_A_4_LANES bit on
in our driver's internal state if we recognize that we're running on BXT
where it should have been on anyway.

Cc: Imre Deak <imre.deak@intel.com>
Cc: Bob Paauwe <bob.j.paauwe@intel.com>
Signed-off-by: NMatt Roper <matthew.d.roper@intel.com>
Reviewed-by: NBob Paauwe <bob.j.paauwe@intel.com>
Tested-by: NBob Paauwe <bob.j.paauwe@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446764012-27251-1-git-send-email-matthew.d.roper@intel.com

Currently there's no trace in dmesg when the gen2/3 dotclock checks
reject the modeset. Add some to avoid further head scratching.

While at it refactor the code a bit to look nicer.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446241178-432-1-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NAnder Conselvan de Oliveira <conselvan2@gmail.com>

VMA offsets are 64 bits. Plane surface offsets are in ggtt and
the hardware register to set this is thus 32 bits. Be explicit
about these and convert carefully to from vma to final size.

This will make sparse happy by not creating 32bit pointers out
of 64bit vma offsets.

Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446204375-29831-1-git-send-email-mika.kuoppala@intel.comReviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

We have had one case where buggy csr/dmc firmware version influenced
gt side and caused a hang. Add dmc firmware loading state and
version to error state.

v2: - Rebased on top of Damien's patches
    - included fw load state
v3: include dmc info only if platform supports it (Chris)
v4: move *csr to branch scope (Chris)
v5: remove dependency to csr_state

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> (v4)
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446124879-22240-1-git-send-email-mika.kuoppala@intel.com
Tested-by: Daniel Stone <daniels@collabora.com> # SKL
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

We check these to determine firmware loading status. Include
them to help to debug causes of firmware loading fails.

v2: Move all CSR specific registers to i915_reg.h (Ville)
v3: Rebase
v4: Rebase (RPM ref)
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446220487-32691-1-git-send-email-mika.kuoppala@intel.com
Tested-by: Daniel Stone <daniels@collabora.com> # SKL
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

For bxt CSR firmware exposes a count of dc5 entries. Expose
it through debugs
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Tested-by: Daniel Stone <daniels@collabora.com> # SKL
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

The CSR firmware expose two counters, handy to check if we are indeed
entering DC5/DC6.

v2: Rebase
v3: Take RPM ref before reading (Imre)
Signed-off-by: NDamien Lespiau <damien.lespiau@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com> (v1)
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446220412-32574-1-git-send-email-mika.kuoppala@intel.com
Tested-by: Daniel Stone <daniels@collabora.com> # SKL
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

Create a new debufs file for it, we'll have a few more things to add
there.

v2: Fix checkpatch warning about static const array
v3: use named initializers (Ville)
v4: strip out csr_state as it will be removed in future (Ville, Imre)

Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> (v1)
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445950025-5793-3-git-send-email-mika.kuoppala@intel.comReviewed-by: NImre Deak <imre.deak@intel.com>
Tested-by: Daniel Stone <daniels@collabora.com> # SKL
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

There is known issue on GT interrupt delivery with DC6 and
firmwares <1.21. There is a suspicion that this causes
spurious gpu hangs on driver init and with some workloads,
as upgrading the firmware to 1.21 makes these problems
disappear.

As of now the current version included in distribution
firmware packages is very like to be 1.19. Play it safe and
refuse to load a firmware version that may affect gpu
side stability.

With < 1.23 there is a palette and dmc ram corruption issue
so blacklist anything below that.

v2: Refuse to load fw instead of notifying the user
v3: Rebase on header version changes
v4: Refuse to load anything less than 1.23
v5: Give enough information for user for finding correct fw (Chris)
v6: better url and formatting (Chris)
v7: move error log for each fail path (Mika)
    bail out earlier in load path (Imre)
v8: Fix the version check (Imre)

Cc: Animesh Manna <animesh.manna@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Dave Gordon <david.s.gordon@intel.com>
Cc: Arun Siluvery <arun.siluvery@linux.intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Patrik Jakobsson <patrik.jakobsson@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@gmail.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
References: https://01.org/linuxgraphics/downloads/skldmcver121
References: https://01.org/linuxgraphics/downloads/skylake-dmc-1.23
Testcase: igt/gem_exec_nop
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1446220336-32392-1-git-send-email-mika.kuoppala@intel.com
Tested-by: Daniel Stone <daniels@collabora.com> # SKL
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

That can be handy later on to tell which DMC firmware version the user
has, by just looking at the dmesg.

v2: use DRM_DEBUG_DRIVER (Chris)
v3: use DRM_INFO (Marc Herbert)

Cc: Marc Herbert <marc.herbert@intel.com>
Signed-off-by: Damien Lespiau <damien.lespiau@intel.com> (v1)
Signed-off-by: NMika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445950025-5793-1-git-send-email-mika.kuoppala@intel.comReviewed-by: NImre Deak <imre.deak@intel.com>
Tested-by: Daniel Stone <daniels@collabora.com> # SKL
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

We had two failure modes here:

1.
Deadlock in intelfb_alloc failure path where it calls
drm_framebuffer_remove, which grabs the struct mutex and intelfb_create
(caller of intelfb_alloc) was already holding it.

2.
Deadlock in intelfb_create failure path where it calls
drm_framebuffer_unreference, which grabs the struct mutex and
intelfb_create was already holding it.

[Daniel Vetter on why struct_mutex needs to be locked in the second half
of intelfb_create: "The vma [for the fbdev] is pinned, the problem is
that we re-lookup it a few times, which is racy. We should instead track
the vma directly, but oh well we don't."]

v2:
   * Reformat commit msg to 72 chars. (Lukas Wunner)
   * Add third failure mode. (Lukas Wunner)

v5:
   * Rebase on drm-intel-nightly 2015y-09m-01d-09h-06m-08s UTC,
     rephrase commit message. (Jani Nicula)

v6:
   * In intelfb_alloc, if __intel_framebuffer_create failed,
     fb will be an ERR_PTR, thus not null. So in the failure
     path we need to check for IS_ERR_OR_NULL to avoid calling
     drm_framebuffer_remove on the ERR_PTR. (Lukas Wunner)
   * Since this is init code a drm_framebuffer_unreference should
     be all we need. drm_framebuffer_remove is for framebuffers
     that userspace has created - and is getting somewhat
     defeatured. (Daniel Vetter)

v7:
   * Clarify why struct_mutex needs to be locked in the second half
     of intelfb_create. (Daniel Vetter)

Fixes: 60a5ca01 ("drm/i915: Add locking around
    framebuffer_references--")
Reported-by: NLukas Wunner <lukas@wunner.de>
Signed-off-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
[Lukas: Create v3 + v4 + v5 + v6 + v7 based on Tvrtko's v2]
Signed-off-by: NLukas Wunner <lukas@wunner.de>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/47d4e88c91b3bf0f7a280cabec54c8c8cf0cf6f2.1446892879.git.lukas@wunner.deSigned-off-by: NJani Nikula <jani.nikula@intel.com>

In intelfb_alloc(), if the call to intel_pin_and_fence_fb_obj() fails,
the bo is unrefed twice: By drm_framebuffer_remove() and once more by
drm_gem_object_unreference(). Fix it.
Reported-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NLukas Wunner <lukas@wunner.de>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/cd7b33330621a350b0159ec5e098297b139cfaf7.1446892879.git.lukas@wunner.deSigned-off-by: NJani Nikula <jani.nikula@intel.com>

Currently when allocating a framebuffer fails, the gem object gets
unrefed at the bottom of the call stack in __intel_framebuffer_create,
not where it gets refed, which is in intel_framebuffer_create_for_mode
(via i915_gem_alloc_object) and in intel_user_framebuffer_create
(via drm_gem_object_lookup).

This invites mistakes: __intel_framebuffer_create is also called from
intelfb_alloc, and as discovered by Tvrtko Ursulin, a double unref
was introduced there with a8bb6818 ("drm/i915: Fix error path leak
in fbdev fb allocation").

As suggested by Ville Syrjälä, fix the double unref and improve code
clarity by moving the unref away from __intel_framebuffer_create to
where the gem object gets refed.

Based on Tvrtko Ursulin's original v2.

v3: On fb alloc failure, unref gem object where it gets refed,
    fix double unref in separate commit (Ville Syrjälä)

v4: Lock struct_mutex on unref (Chris Wilson)

v5: Rebase on drm-intel-nightly 2015y-09m-01d-09h-06m-08s UTC,
    rephrase commit message (Jani Nicula)
Tested-by: NPierre Moreau <pierre.morrow@free.fr>
    [MBP  5,3 2009  nvidia MCP79 + G96        pre-retina]
Tested-by: NPaul Hordiienko <pvt.gord@gmail.com>
    [MBP  6,2 2010  intel ILK + nvidia GT216  pre-retina]
Tested-by: NWilliam Brown <william@blackhats.net.au>
    [MBP  8,2 2011  intel SNB + amd turks     pre-retina]
Tested-by: NLukas Wunner <lukas@wunner.de>
    [MBP  9,1 2012  intel IVB + nvidia GK107  pre-retina]
Tested-by: NBruno Bierbaumer <bruno@bierbaumer.net>
    [MBP 11,3 2013  intel HSW + nvidia GK107  retina]

Fixes: a8bb6818 ("drm/i915: Fix error path leak in fbdev fb
    allocation")
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NLukas Wunner <lukas@wunner.de>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/2161c5062ef5d6458f8ae14d924a26d4d1dba317.1446892879.git.lukas@wunner.deSigned-off-by: NJani Nikula <jani.nikula@intel.com>

The function name implies it should get intel_dp, and it mostly used
where there is an intel_dp in the context.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NSivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445594525-7174-8-git-send-email-ander.conselvan.de.oliveira@intel.com

In order to prepare for a link training with DDI, the state machine
would call intel_ddi_prepare_link_retrain(). To remove the dependency to
the hardware information, replace that direct call with a callback.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NSivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445594525-7174-7-git-send-email-ander.conselvan.de.oliveira@intel.com

No functional changes, just moving code around.

v2: Rebase
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NSivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445594525-7174-6-git-send-email-ander.conselvan.de.oliveira@intel.com

Move register write from intel_dp_update_link_train() into
intel_dp_set_signal_levels(). This creates a better split between the
i915 specific code and the generic link training part. Note that this
causes an extra register write in intel_dp_reset_link_train(), since
both intel_dp_set_signal_levels() and intel_dp_set_link_train() write
to the DP register.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NSivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445594525-7174-5-git-send-email-ander.conselvan.de.oliveira@intel.com

Move the call to intel_dp_get_adjust_train() out of
intel_dp_update_link_train() and call it instead from the clock recovery
and channel equalization features. A follow up patch will remove the DP
register write from that function, so that it handles only the DPCD
write.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NSivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445594525-7174-4-git-send-email-ander.conselvan.de.oliveira@intel.com

Split the register write with the new link training pattern out of
intel_dp_set_link_train(), so that the i915 specific code is in a
separate function.
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NSivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445594525-7174-3-git-send-email-ander.conselvan.de.oliveira@intel.com

It just makes the code more confusing, so just reference intel_dp_>DP
directly.

Note that this also fix a bug where the value of intel_dp->DP could be
different than the last value written to the hw, due to an early return
that would skip the 'intel_dp->DP = DP' line.

v2: Don't preserve old DP value on failure. (Sivakumar)
  - Don't call drm_dp_clock_recovery_ok() twice. (Sivakumar)
  - Keep return type of clock recovery and channel equalization
    functions as void. (Ander)

v3: Remove DP parameter from intel_dp_set_signal_levels(). (Sivakumar)
Signed-off-by: NAnder Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Reviewed-by: NSivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1445594525-7174-2-git-send-email-ander.conselvan.de.oliveira@intel.com