The idle work handler is self-arming - if it detects that it needs to
run again it will queue itself from its work handler. Take greater care
when trying to drain the idle work, and double check that it is flushed.

The free worker has a similar issue where it is armed by an RCU task
which may be running concurrently with us.

This should hopefully help with the sporadic WARN_ON(dev_priv->gt.awake)
from i915_gem_suspend.

v2: Reuse drain_freed_objects.
v3: Don't try to flush the freed objects from the shrinker, as it may be
underneath the struct_mutex already.
v4: do while and comment upon the excess rcu_barrier in drain_freed_objects
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161223145804.6605-2-chris@chris-wilson.co.uk

Since commit db6c2b41 ("drm/i915: Store the vma in an rbtree under
the object") the vma are once again sorted into GGTT first, then ppGTT
so that the typical case of walking the GGTT vma can stop as soon as we
find a non-ppGTT. Apply that optimisation.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161223145804.6605-1-chris@chris-wilson.co.uk

Now that we're disabling L2 clock gating MI_OVERLAY_OFF actually works
on 830, so let's use it.

v2: Nuke the unused dev_priv variable
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161222195222.9618-1-ville.syrjala@linux.intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

On enable intel_dsi_enable() directly calls intel_enable_dsi_pll(),
make intel_dsi_disable() also directly call intel_disable_dsi_pll(),
rather then hiding the call in intel_dsi_clear_device_ready(),
no functional changes.
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161201202925.12220-7-hdegoede@redhat.com

The upper bits of the vsync width, vsync offset and hsync width
were not parsed from the VBT. Parse these fields in this patch.

V2: Renamed lvds dvo timing structure members and code identation
fix (Jani's review comments)
V3: Corrected commit message, used "from the VBT"
Signed-off-by: NVincente Tsou <vincente.tsou@intel.com>
Signed-off-by: NMadhav Chauhan <madhav.chauhan@intel.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1482430993-3265-1-git-send-email-madhav.chauhan@intel.com

BSpec says:
"Overlay Clock Gating Must be Disabled: Overlay & L2 Cache clock gating
must be disabled in order to prevent device hangs when turning off overlay.SW
must turn off Ovrunit clock gating (6200h) and L2 Cache clock gating (C8h)."

We only turned off the overlay clock gating (due to lack of docs I
presume). After a bit of experimentation it looks like the the magic
C8h register lives in the PCI config space of device 0, and the magic
bit appears to be bit 2. Or at the very least this eliminates the GPU
death after MI_OVERLAY_OFF.

L2 clock gating seems to save ~80mW, so let's keep it on unless we need
to actually use the overlay.

Also let's move the OVRUNIT clock gating to the same place since we can,
and 845 supposedly doesn't need it.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481131693-27993-11-git-send-email-ville.syrjala@linux.intel.comAcked-by: NChris Wilson <chris@chris-wilson.co.uk>

Extract the primary plane pixel format via plane state when setting up
the overlay colorkey.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481131693-27993-10-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

Use two-dimensional arrays and named initializers to make the
overlay filter coefficient tables easier to parse for humans.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481131693-27993-9-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

The formula in Bspec for computing the overlay SWIDTHSW is overly
obfuscated. Simplify the formula to something that's easily parsed by
humans.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481131693-27993-8-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

Check pipe config gmch_pfit.control instead of using intel_panel_fitter_pipe()
to figure out if the pipe for the overlay is using the panel fitter.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481131693-27993-7-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

Replace the use of crtc->mode.h/vdisplay with the more appropriate
config->pipe_src_w/h.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481131693-27993-6-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

Do the overlay frontbuffer tracking properly so that it matches
the state of the overlay on/off/continue requests.

One slight problem is that intel_frontbuffer_flip_complete()
may get delayed by an arbitrarily liong time due to the fact that
the overlay code likes to bail out when a signal occurs. So the
flip may not get completed until the ioctl is restarted. But fixing
that would require bigger surgery, so I decided to ignore it for now.
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481131693-27993-5-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

The vma will be NULL if the overlay was previously off, so
dereferencing it will oops. Check for NULL before doing that.

Cc: stable@vger.kernel.org
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Fixes: 9b3b7841 ("drm/i915/overlay: Use VMA as the primary tracker for images")
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481131693-27993-2-git-send-email-ville.syrjala@linux.intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>

The i915_gem_active stuff doesn't like a NULL ->retire hook, but
the overlay code can set it to NULL. That obviously ends up oopsing.
Fix it by introducing a new helper to assign the retirement callback
that will switch out the NULL function pointer with
i915_gem_retire_noop.

Cc: stable@vger.kernel.org
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Fixes: 0d9bdd88 ("drm/i915: Convert intel_overlay to request tracking")
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20161207175647.10018-1-chris@chris-wilson.co.uk

Initialize overlay->last_flip properly instead of leaving it zeroed.

Cc: stable@vger.kernel.org
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Fixes: 0d9bdd88 ("drm/i915: Convert intel_overlay to request tracking")
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161221144547.27319-1-ville.syrjala@linux.intel.com

Trying to determine the pixel rate of the pipe can't be done until we
know the clock, which means it can't be done until the encoder
.get_config() hooks have been called. So let's move the min_pixclk[]
stuff to the end of intel_modeset_readout_hw_state() when we actually
have gathered all the required infromation.

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Mika Kahola <mika.kahola@intel.com>
Cc: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Fixes: 565602d7 ("drm/i915: Do not acquire crtc state to check clock during modeset, v4.")
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161220153902.15621-1-ville.syrjala@linux.intel.comReviewed-by: NAnder Conselvan de Oliveira <conselvan2@gmail.com>
Reviewed-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>

Apparently some VLV BIOSen like to leave the VDD force bit enabled
even for power seqeuncers that aren't properly hooked up to any
port. That will result in a imbalance in the AUX power domain
refcount when we stat to use said power sequencer as edp_panel_vdd_on()
will not grab the power domain reference if it sees that the VDD is
already on.

To fix this let's make sure we turn off the VDD force bit when we
initialize the power sequencer registers. That is, unless it's
being done from the init path since there we are actually
initializing the registers for the current power sequencer and
we don't want to turn VDD off needlessly as that would require
waiting for the power cycle delay before we turn it back on.

This fixes the following kind of warnings:
WARNING: CPU: 0 PID: 123 at ../drivers/gpu/drm/i915/intel_runtime_pm.c:1455 intel_display_power_put+0x13a/0x170 [i915]()
WARN_ON(!power_domains->domain_use_count[domain])
...

v2: Fix typos in comment (David)

Cc: stable@vger.kernel.org
Cc: Matwey V. Kornilov <matwey.kornilov@gmail.com>
Tested-by: NMatwey V. Kornilov <matwey.kornilov@gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98695Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161220165117.24801-1-ville.syrjala@linux.intel.comReviewed-by: NDavid Weinehall <david.weinehall@linux.intel.com>

There is at least one APL based system using port A in DP mode
(connecting to an on-board DP->VGA adaptor). Atm we'll configure port A
unconditionally as eDP which is incorrect in this case. Fix this by
relying on the VBT DDI port 'internal port' flag instead on all ports on
DDI platforms. For now chicken out from using VBT for port A before
GEN9.

v2:
- Move the DDI port info lookup to intel_bios_is_port_edp() (David, Jani)
- Use the DDI port info on all DDI platforms starting from port B.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: David Weinehall <david.weinehall@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> (v1)
Reviewed-by: NDavid Weinehall <david.weinehall@linux.intel.com>
Signed-off-by: NImre Deak <imre.deak@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1482315444-24750-1-git-send-email-imre.deak@intel.com

when num_pipes is zero, it indicates there is no display and HDMI
audio doesn't exist.

v2: Move the check from caller to callee for consistency.

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Signed-off-by: NElaine Wang <elaine.wang@intel.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1482142746-21663-1-git-send-email-elaine.wang@intel.com

v2: Addressed Jani's Review comments (renamed bit field macros)
Signed-off-by: NDeepak M <m.deepak@intel.com>
Signed-off-by: NMadhav Chauhan <madhav.chauhan@intel.com>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Signed-off-by: NJani Nikula <jani.nikula@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481792500-30863-2-git-send-email-madhav.chauhan@intel.com

This macro got useless after commit 8d9c20e1
"drm/i915: Remove .is_mobile field from platform struct"

that removed is_mobile split from VLV definition.
Also this was never reused on any following platform.

So let's clean up a bit here.

Cc: Carlos Santa <carlos.santa@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: NRodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1482184508-18346-1-git-send-email-rodrigo.vivi@intel.com

Don't even tell the mm allocator to handle the first page of stolen on
the affected platforms. This means that we won't inherit the FB in
case the BIOS decides to put it at the start of stolen. But the BIOS
should not be putting it at the start of stolen since it's going to
get corrupted. I suppose the bug here is that some pixels at the very
top of the screen will be corrupted, so it's not exactly easy to
notice.

We have confirmation that the first page of stolen does actually get
corrupted, so I really think we should do this in order to avoid any
possible future headaches, even if that means losing BIOS framebuffer
inheritance. Let's not use the HW in a way it's not supposed to be
used.

Notice that now ggtt->stolen_usable_size won't reflect the ending
address of the stolen usable range anymore, so we have to fix the
places that rely on this. To simplify, we'll just use U64_MAX.

v2: don't even put the first page on the mm (Chris)
v3: drm_mm_init() takes size instead of end as argument (Ville)
v4: add a comment explaining the reserved ranges (Chris)
    use 0 for start and U64_MAX for end when possible (Chris)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94605
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/1481808235-27607-1-git-send-email-paulo.r.zanoni@intel.com

BSpec got updated and this workaround is now listed as standard
required programming for all subsequent projects. This is confirmed to
fix Skylake screen flickering issues (probably caused by the fact that
we initialized a ring in the first page of stolen, but I didn't 100%
confirm this theory).

v2: this is the patch that fixes the screen flickering, document it.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94605
Cc: stable@vger.kernel.org
Tested-by: NDominik Klementowski <dominik232@gmail.com>
Signed-off-by: NPaulo Zanoni <paulo.r.zanoni@intel.com>
Acked-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1481727338-9901-1-git-send-email-paulo.r.zanoni@intel.com

In commit 0c40ce13 ("drm/i915: Trim the object sg table"), we expect
to copy exactly orig_st->nents across and allocate the table thusly.
The copy loop should therefore end with the new_sg being NULL.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161219124346.550-2-chris@chris-wilson.co.ukReviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>

If we at first do not succeed with attempting to remap our physical
pages using a coalesced scattergather list, try again with one
scattergather entry per page. This should help with swiotlb as it uses a
limited buffer size and only searches for contiguous chunks within its
buffer aligned up to the next boundary - i.e. we may prematurely cause a
failure as we are unable to utilize the unused space between large
chunks and trigger an error such as:

	 i915 0000:00:02.0: swiotlb buffer is full (sz: 1630208 bytes)
Reported-by: NJuergen Gross <jgross@suse.com>
Tested-by: NJuergen Gross <jgross@suse.com>
Fixes: 871dfbd6 ("drm/i915: Allow compaction upto SWIOTLB max segment size")
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Cc: <drm-intel-fixes@lists.freedesktop.org>
Link: http://patchwork.freedesktop.org/patch/msgid/20161219124346.550-1-chris@chris-wilson.co.ukReviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

In keeping with commit f802cf7e ("drm/i915/debugfs: use
rb_entry()"), convert the primary user of the rbtrees over to using
rb_entry rather than the equivalent container_of.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20161220104003.8044-1-chris@chris-wilson.co.ukReviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

To make the code clearer, use rb_entry() instead of container_of() to
deal with rbtree.
Signed-off-by: NGeliang Tang <geliangtang@gmail.com>
Signed-off-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/62ce937ae9a341421942b4418515610d055fa653.1482158544.git.geliangtang@gmail.com

Let's take usage of IS_LP to simplify the gem stolen
initialization as suggest by Tvrtko.

Also assume that all new LP platforms follows the chv+
and others bdw+.

v2: Remove the wrong commit message about bxt and glk. (Ander)

Cc: Ander Conselvan de Oliveira <ander.conselvan.de.oliveira@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: NRodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: NAnder Conselvan de Oliveira <conselvan2@gmail.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1482174347-24911-1-git-send-email-rodrigo.vivi@intel.com

gen8 is used for both Broadwell and Cherryview but this
function here is only Cherryview and all next atom LP platforms.
So let's rename it to avoid confusion as suggested by Ville.

Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NRodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: NAnder Conselvan de Oliveira <conselvan2@gmail.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1482096988-400-2-git-send-email-rodrigo.vivi@intel.com

Valleyview/Baytrail (gen7_lp) and Cherryview/Braswell (gen8_lp)
are both Atom platforms like Broxton/Apollolake and Geminilake.

So let's expand this is_lp back to these platforms and
create the IS_LP(dev_priv) so we can start simplifying a bit
our if/else for platform lists.
Signed-off-by: NRodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: NAnder Conselvan de Oliveira <conselvan2@gmail.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1482096988-400-1-git-send-email-rodrigo.vivi@intel.com

VLV apparently gets upset if the PPS for a pipe currently driving an
external DP port gets used for VDD stuff on another eDP port. The DP
port falls over and fails to retrain when this happens, leaving the
user staring at a black screen.

Let's fix it by also tracking which pipe is driving which DP/eDP port.
We'll track this under intel_dp so that we'll share the protection
of the pps_mutex alongside the pps_pipe tracking, since the two
things are intimately related.

I had plans to reduce the protection of pps_mutex to cover only eDP
ports, but with this we can't do that. Well, for for VLV/CHV at least.
For other platforms it should still be possible, which would allow
AUX communication to occur in parallel for multiple DP ports.

v2: Drop stray crap from a comment (Imre)
    Grab pps_mutex when clearing active_pipe
    Fix a typo in the commit message
v3: Make vlv_active_pipe() static

Cc: stable@vger.kernel.org
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1481738423-29738-1-git-send-email-ville.syrjala@linux.intel.com

The kref_put_mutex() returns with the mutex held after freeing the
object - so we must remember to drop it...

Fixes: 69df05e1 ("drm/i915: Simplify releasing context reference")
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161219101357.28140-1-chris@chris-wilson.co.ukReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>

A fairly trivial move of a matching pair of routines (for preparing a
request for construction) onto an engine vfunc. The ulterior motive is
to be able to create a mock request implementation.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-7-chris@chris-wilson.co.uk

PIN_HIGH is an expensive operation (in comparison to allocating from the
hole stack) unsuitable for frequent use (such as switching between
contexts). However, the kernel context should be pinned just once for
the lifetime of the driver, and here it is appropriate to keep it out of
the mappable range (in order to maximise mappable space for users).
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-6-chris@chris-wilson.co.uk

As the shadow gvt is not user accessible and does not have an associated
vm, we can mark it as closed during its construction. This saves leaking
the internal knowledge of i915_gem_context into gvt/.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-5-chris@chris-wilson.co.uk

A few users only take the struct_mutex in order to release a reference
to a context. We can expose a kref_put_mutex() wrapper in order to
simplify these users, and optimise taking of the mutex to the final
unref.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-4-chris@chris-wilson.co.uk

The requests conversion introduced a nasty bug where we could generate a
new request in the middle of constructing a request if we needed to idle
the system in order to evict space for a context. The request to idle
would be executed (and waited upon) before the current one, creating a
minor havoc in the seqno accounting, as we will consider the current
request to already be completed (prior to deferred seqno assignment) but
ring->last_retired_head would have been updated and still could allow
us to overwrite the current request before execution.

We also employed two different mechanisms to track the active context
until it was switched out. The legacy method allowed for waiting upon an
active context (it could forcibly evict any vma, including context's),
but the execlists method took a step backwards by pinning the vma for
the entire active lifespan of the context (the only way to evict was to
idle the entire GPU, not individual contexts). However, to circumvent
the tricky issue of locking (i.e. we cannot take struct_mutex at the
time of i915_gem_request_submit(), where we would want to move the
previous context onto the active tracker and unpin it), we take the
execlists approach and keep the contexts pinned until retirement.
The benefit of the execlists approach, more important for execlists than
legacy, was the reduction in work in pinning the context for each
request - as the context was kept pinned until idle, it could short
circuit the pinning for all active contexts.

We introduce new engine vfuncs to pin and unpin the context
respectively. The context is pinned at the start of the request, and
only unpinned when the following request is retired (this ensures that
the context is idle and coherent in main memory before we unpin it). We
move the engine->last_context tracking into the retirement itself
(rather than during request submission) in order to allow the submission
to be reordered or unwound without undue difficultly.

And finally an ulterior motive for unifying context handling was to
prepare for mock requests.

v2: Rename to last_retired_context, split out legacy_context tracking
for MI_SET_CONTEXT.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-3-chris@chris-wilson.co.uk

Just a simple move to avoid a forward declaration, though the diff likes
to present itself as a move of intel_logical_ring_alloc_request_extras()
in the opposite direction.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-2-chris@chris-wilson.co.uk

i915_vma_move_to_active() requires the struct_mutex for serialisation
with retirement, so mark it up with lockdep_assert_held().
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-1-chris@chris-wilson.co.uk

Convert some of the obvious hand-rolled ranged overflow sanity checks to
our shiny new range_overflows macro.

Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Suggested-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NMatthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161213203222.32564-4-matthew.auld@intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>