assert_rpm_wakelock_held is triggered from i915_pmic_bus_access_notifier
even though it gets unregistered on (runtime) suspend, this is caused
by a race happening under the following circumstances:

intel_runtime_pm_put does:

   atomic_dec(&dev_priv->pm.wakeref_count);

   pm_runtime_mark_last_busy(kdev);
   pm_runtime_put_autosuspend(kdev);

And pm_runtime_put_autosuspend calls intel_runtime_suspend from
a workqueue, so there is ample of time between the atomic_dec() and
intel_runtime_suspend() unregistering the notifier. If the notifier
gets called in this windowd assert_rpm_wakelock_held falsely triggers
(at this point we're not runtime-suspended yet).

This commit adds disable_rpm_wakeref_asserts and
enable_rpm_wakeref_asserts calls around the
intel_uncore_forcewake_get(FORCEWAKE_ALL) call in
i915_pmic_bus_access_notifier fixing the false-positive WARN_ON.

Changes in v2:
-Reword comment explaining why disabling the wakeref asserts is
 ok and necessary

Cc: stable@vger.kernel.org
Reported-by: NFKr <bugs-freedesktop@ubermail.me>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110150301.9601-2-hdegoede@redhat.com

smatch does not track initialised values as well as gcc, and this
triggers many warnings by smatch not presented by gcc. Silence smatch by
initialising the error values to -ENODEV, which we use to denote
internal errors. (If we see a selftest fail with a silent -ENODEV, we
know smatch was right!)

v2: smatch was right about igt_create_vma(), it may unlikely fail on the
first object allocation which we want to be loud about.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20171114223346.25958-1-chris@chris-wilson.co.ukReviewed-by: NDhinakaran Pandiyan <dhinakaran.pandiyan@intel.com>

Resuming GEM presumes it can talk to hw, in particular to ensure the
kernel context is loaded upon resume for powersaving. If the GuC is
still asleep at this point, we upset the HW. Rearrange the resume such
that we restore the original order of init-hw, resume-guc, use-gem.

Fixes: 37cd3300 ("drm/i915: Remove redundant intel_autoenable_gt_powersave()")
References: a1c41994 ("drm/i915/guc: Add host2guc notification for suspend and resume")
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Alex Dai <yu.dai@intel.com>
Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171114130300.25677-2-chris@chris-wilson.co.ukReviewed-by: NSagar Arun Kamble <sagar.a.kamble@intel.com>

Display is not sending a PMRsp when a PMReq is received
at the same time that all planes are turned off.
State machine in the dcprunit is stuck in the WAIT4DONE
state which means that there is no fill_done.

WA: disable arbiter clock gating, set bit [27] of 0x46530

v2: As Ville pointed out, based on the description the issue
    can happen when disabling the planes, similar to
    WaRsPkgCStateDisplayPMReq:hsw
    Also description of the issue was updated on commit
    message to make it more clear that we need this
    earlier.
v3: Restore comment about possibility to system hang
    to where we are sure about it, without speculation. (Ville).
v4: Remove doubled sob. Actually do v3 changes :/

Cc: Radhakrishna Sripada <radhakrishna.sripada@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NRodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171111000319.5040-1-rodrigo.vivi@intel.com

Having removed the preproduction Broxton support (see commit 0102ba1f
("drm/i915: Add early BXT sdv to the list of preproduction machines")),
we know we then always need the production Broxton workaround set and do
not need a predicate upon revision.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171114134340.5439-2-chris@chris-wilson.co.uk

We've begun excluding pre-production Broxton machines since commit
0102ba1f ("drm/i915: Add early BXT sdv to the list of preproduction
machines"), now remove the list of workaround register values for those
early machines.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20170927093325.24206-1-chris@chris-wilson.co.ukReviewed-by: NJani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171114134340.5439-1-chris@chris-wilson.co.uk

gem_workarounds reports that the SLICE_UNIT_LEVEL_CLKGATE write isn't
sticking. Commit 0a60797a ("drm/i915: Implement
ReadHitWriteOnlyDisable.") presumes that SLICE_UNIT_LEVEL_CLKGATE is a
masked register in the context image, but commit 90007bca
("drm/i915/cnl: Introduce initial Cannonlake Workarounds.") lists it as
an ordering unmasked register. The masked write will be losing the
default settings if we trust the original commit. That gem_workarounds
reports the value is lost entirely is more worrying though -- but it
clearly suggests that it is not a masked register in the context image,
so unify both w/a to use the original rmw.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103705
Fixes: 0a60797a ("drm/i915: Implement ReadHitWriteOnlyDisable.")
References: 90007bca ("drm/i915/cnl: Introduce initial Cannonlake Workarounds.")
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Rafael Antognolli <rafael.antognolli@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Oscar Mateo <oscar.mateo@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171111100336.11020-1-chris@chris-wilson.co.ukReviewed-by: NRafael Antognolli <rafael.antognolli@intel.com>

ERROR: "__udivdi3" [drivers/gpu/drm/i915/i915.ko] undefined!
ERROR: "__divdi3" [drivers/gpu/drm/i915/i915.ko] undefined!

Store the frequency in kHz and drop 64bit divisions.

v2: Use div64_u64 (Matthew)

v3: store frequency in kHz to avoid 64bit divs (Chris/Ville)

Fixes: dab91783 ("drm/i915: expose command stream timestamp frequency to userspace")
Reported-by: NMatthew Auld <matthew.auld@intel.com>
Signed-off-by: NLionel Landwerlin <lionel.g.landwerlin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171113233455.12085-3-lionel.g.landwerlin@intel.comReviewed-by: NEwelina Musial <ewelina.musial@intel.com>
Signed-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

To make looping through transcoders in intel_ddi.c more generic, let's switch
to use 'for_each_pipe()' macro to do this.

v2: Add a notion that we are dealing with transcoders instead of pipes (Jani)
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: NMika Kahola <mika.kahola@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1510216670-16848-1-git-send-email-mika.kahola@intel.comSigned-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

Since GLK, some plane configuration settings have moved to the
PLANE_COLOR_CTL register. Refactor handling of the register to work like
PLANE_CTL. This also allows us to fix the set/read of the plane Alpha
Mode for GLK+.

v2: Adjust ordering of platform checks to be newest->oldest, drop
redundant comment about alpha blending. (Ville)

v3: Move Alpha Mode bits out of skl_plane_ctl_format into
skl_plane_ctl_alpha, and drop glk_plane_ctl_format, drop initialization
of state->color_ctl on platforms that don't use it, and drop color_ctl
local var. (Ville)

v4: Consolidate skl_plane_ctl_format switch statement on formats that
return the same settings. (Ville)
Signed-off-by: NJames Ausmus <james.ausmus@intel.com>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171113181128.2926-1-james.ausmus@intel.comSigned-off-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

GEM proxy is a kind of GEM, whose backing physical memory is pinned
and produced by guest VM and is used by host as read only. With GEM
proxy, host is able to access guest physical memory through GEM object
interface. As GEM proxy is such a special kind of GEM, a new flag
I915_GEM_OBJECT_IS_PROXY is introduced to ban host from changing the
backing storage of GEM proxy.

v3:
- update "Reviewed-by". (Joonas)

v2:
- return -ENXIO when pin and map pages of GEM proxy to kernel space.
  (Chris)

Here are the histories of this patch in "Dma-buf support for Gvt-g"
patch-set:

v14:
- return -ENXIO when gem proxy object is banned by ioctl.
  (Chris) (Daniel)

v13:
- add comments to GEM proxy. (Chris)
- don't ban GEM proxy in i915_gem_sw_finish_ioctl. (Chris)
- check GEM proxy bar after finishing i915_gem_object_wait. (Chris)
- remove GEM proxy bar in i915_gem_madvise_ioctl.

v6:
- add gem proxy barrier in the following ioctls. (Chris)
  i915_gem_set_caching_ioctl
  i915_gem_set_domain_ioctl
  i915_gem_sw_finish_ioctl
  i915_gem_set_tiling_ioctl
  i915_gem_madvise_ioctl
Signed-off-by: NTina Zhang <tina.zhang@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> #v1
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/1510555798-21079-2-git-send-email-tina.zhang@intel.comSigned-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20171114102513.22269-2-chris@chris-wilson.co.uk

-ENXIO should be returned when operations are banned from changing
backing storage of objects without backing storage.

v4:
- update "Reviewed-by". (Joonas)

v3:
- separate this patch from "Introduce GEM proxy" patch-set. (Joonas)

v2:
- update the patch description and subject to just mention objects w/o
  backing storage, instead of "GEM proxy". (Joonas)
Signed-off-by: NTina Zhang <tina.zhang@intel.com>
Reviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/1510555798-21079-1-git-send-email-tina.zhang@intel.comReviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20171114102513.22269-1-chris@chris-wilson.co.uk

Commit 78597996 (drm/i915/bxt: Fix PPS lost state after suspend
breaking eDP link training) renamed the function to
intel_power_sequencer_reset() but forgot to update comment.

Cc: Imre Deak <imre.deak@intel.com>
Signed-off-by: NLucas De Marchi <lucas.demarchi@intel.com>
Signed-off-by: NImre Deak <imre.deak@intel.com>
Reviewed-by: NImre Deak <imre.deak@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171114004638.5186-1-lucas.demarchi@intel.com

Until Haswell/Baytrail, the hardware used to have a per engine fault
register (e.g. 0x4094 - render fault register, 0x4194 - media fault
register and so on). But since Broadwell, all these registers were
combined into a singe one and the engine id stored in bits 14:12.

Not only we should not been reading (and writing to) registers that do
not exist, in platforms with VCS2 (SKL), the address that would belong
this engine (0x4494, VCS2_HW = 4) is already assigned to other register.

v2: use less controversial function names (Chris).
v3: make non-exported functions static, remove now obsolete check for
engine presence before posting_read (Chris).

References: IHD-OS-BDW-Vol 2c-11.15, page 75.
References: IHD-OS-SKL-Vol 2c-05.16, page 350.
Signed-off-by: NMichel Thierry <michel.thierry@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20171113173628.11689-1-michel.thierry@intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

From gen6, the hardware tracks address lookup failures and we should
clear those registers upon startup to prevent false positives. However,
this was happening before we have the engines defined (intel_uncore_init())
and the for_each_engine loop was just a nop. The earliest we can call
this is inside intel_engines_init_mmio().
Suggested-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NMichel Thierry <michel.thierry@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20171111004448.12360-1-michel.thierry@intel.comReviewed-by: NChris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

We use to have this fixed per generation, but starting with CNL userspace
cannot tell just off the PCI ID. Let's make this information available. This
is particularly useful for performance monitoring where much of the
normalization work is done using those timestamps (this include pipeline
statistics in both GL & Vulkan as well as OA reports).

v2: Use variables for 24MHz/19.2MHz values (Ewelina)
    Renamed function & coding style (Sagar)

v3: Fix frequency read on Broadwell (Sagar)
    Fix missing divide by 4 on <= gen4 (Sagar)
Signed-off-by: NLionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: NRafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: NSagar Arun Kamble <sagar.a.kamble@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110190845.32574-7-lionel.g.landwerlin@intel.com

This adds new registers to the whitelist to configs emitted from userspace.
Signed-off-by: NLionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: NMatthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110190845.32574-6-lionel.g.landwerlin@intel.com

This name was added with the whitelisting of registers for building up OA
configs. It is contained in a range gen8 whitelist :

   addr >= RPM_CONFIG0.reg && addr <= NOA_CONFIG(8).reg

Hence why the name isn't used anywhere.

v2: Fix register name again RPC->RCP (Matthew)
Signed-off-by: NLionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: NMatthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110190845.32574-5-lionel.g.landwerlin@intel.com

Gen8/9 aren't very different and we can merge some of this code.
Signed-off-by: NLionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: NMatthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110190845.32574-4-lionel.g.landwerlin@intel.com

We can enable GT3 as well as GT2.
Signed-off-by: NLionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: NMatthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110190845.32574-3-lionel.g.landwerlin@intel.com

We were missing some registers and also can name one for which we only had
the offset.
Signed-off-by: NLionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: NMatthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110190845.32574-2-lionel.g.landwerlin@intel.com

Lock the bare minimum, instead of the entire world, and
use interruptible locking because we can.
Signed-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110113503.16253-6-maarten.lankhorst@linux.intel.comReviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Some parameters use CHECK_BOOL, but should really use
CHECK_BOOL_INCOMPLETE. We cannot currently check whether
the inherited infoframes and audio are set up correctly.
Signed-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110113503.16253-4-maarten.lankhorst@linux.intel.com
[mlankhorst: Add danvet's comment about why this is needed.]
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

Add PIPE_CONF_CHECK_BOOL for boolean options, which are printed with yesno.
Signed-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110113503.16253-3-maarten.lankhorst@linux.intel.comReviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

The flag just tells us IPS can be enabled, if the primary plane
is not enabled it means IPS might not be. This never triggered
in CI because we don't have a haswell ULT there, but can be
reproduced easily with kms_atomic_transitions.plane-all-modeset-transition
Signed-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110113503.16253-2-maarten.lankhorst@linux.intel.com
[mlankhorst: Remove from haswell_get_pipe_config too. (danvet)]
Reviewed-by: NDaniel Vetter <daniel.vetter@ffwll.ch>

GEN8_CONFIG0 (0xD00) is a protected by a lock (bit 31) which is set by
the BIOS, so there is no way we can enable the three chicken bits
mandated by the WA (the BIOS should be doing it instead).

v2: Rebased
v3: Standalone patch

References: b033bb6d ("drm/i915/gen9: Enable must set chicken bits in config0 reg")
Signed-off-by: NOscar Mateo <oscar.mateo@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1510185589-9100-2-git-send-email-oscar.mateo@intel.comReviewed-by: NMika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>

Now that we always execute a context switch upon module load, there is
no need to queue a delayed task for doing so. The purpose of the delayed
task is to enable GT powersaving, for which we need the HW state to be
valid (i.e. having loaded a context and initialised basic state). We
used to defer this operation as historically it was slow (due to slow
register polling, fixed with commit 1758b90e ("drm/i915: Use a hybrid
scheme for fast register waits")) but now we have a requirement to save
the default HW state.

v2: Load the kernel context (to provide the power context) upon resume.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171112112738.1463-3-chris@chris-wilson.co.uk

We no longer use intel_crtc->wm.active for watermarks any more,
which was incorrect. But this uncovered a bug in sanitize_watermarks(),
which meant that we wrote the correct watermarks, but the next
update would still use the wrong hw watermarks for calculating.
This caused all further updates to fail with -EINVAL and the
log would reveal an error like the one below:

[   10.043902] [drm:ilk_validate_wm_level.part.8 [i915]] Sprite WM0 too large 56 (max 0)
[   10.043960] [drm:ilk_validate_pipe_wm [i915]] LP0 watermark invalid
[   10.044030] [drm:intel_crtc_atomic_check [i915]] No valid intermediate pipe watermarks are possible
Signed-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Fixes: b6b178a7 ("drm/i915: Calculate ironlake intermediate watermarks correctly, v2.")
Cc: stable@vger.kernel.org #v4.8+
Link: https://patchwork.freedesktop.org/patch/msgid/20171110113503.16253-1-maarten.lankhorst@linux.intel.comSigned-off-by: NMaarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>

Move all of the i915->mm initialisation to a private function that can
be reused by the mock i915 device to save forgetting any more steps.

For example,
<7>[ 1542.046332] [IGT] drv_selftest: starting subtest mock_objects
<4>[ 1542.123924] Setting dangerous option mock_selftests - tainting kernel
<6>[ 1542.167941] i915: Performing mock selftests with st_random_seed=0x246f5ab5 st_timeout=1000
<4>[ 1542.178012] INFO: trying to register non-static key.
<4>[ 1542.178027] the code is fine but needs lockdep annotation.
<4>[ 1542.178032] turning off the locking correctness validator.
<4>[ 1542.178041] CPU: 3 PID: 6008 Comm: kworker/3:7 Tainted: G     U          4.14.0-rc8-CI-CI_DRM_3332+ #1
<4>[ 1542.178049] Hardware name:                  /NUC6CAYB, BIOS AYAPLCEL.86A.0040.2017.0619.1722 06/19/2017
<4>[ 1542.178144] Workqueue: events __i915_gem_free_work [i915]
<4>[ 1542.178152] Call Trace:
<4>[ 1542.178163]  dump_stack+0x68/0x9f
<4>[ 1542.178170]  register_lock_class+0x3fd/0x580
<4>[ 1542.178177]  ? unwind_next_frame+0x14/0x20
<4>[ 1542.178184]  ? __save_stack_trace+0x73/0xd0
<4>[ 1542.178191]  __lock_acquire+0xa4/0x1b00
<4>[ 1542.178254]  ? __i915_gem_free_work+0x28/0xa0 [i915]
<4>[ 1542.178261]  ? __lock_acquire+0x4ab/0x1b00
<4>[ 1542.178268]  lock_acquire+0xb0/0x200
<4>[ 1542.178273]  ? lock_acquire+0xb0/0x200
<4>[ 1542.178336]  ? __i915_gem_free_work+0x28/0xa0 [i915]
<4>[ 1542.178344]  _raw_spin_lock+0x32/0x50
<4>[ 1542.178405]  ? __i915_gem_free_work+0x28/0xa0 [i915]
<4>[ 1542.178468]  __i915_gem_free_work+0x28/0xa0 [i915]
<4>[ 1542.178476]  process_one_work+0x221/0x650
<4>[ 1542.178483]  worker_thread+0x4e/0x3c0
<4>[ 1542.178489]  kthread+0x114/0x150
<4>[ 1542.178494]  ? process_one_work+0x650/0x650
<4>[ 1542.178499]  ? kthread_create_on_node+0x40/0x40
<4>[ 1542.178506]  ret_from_fork+0x27/0x40

v2: Fish out i915->mm.object_stat_lock which was being inited over in
i915_drv.c (Matthew)
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110232447.21618-1-chris@chris-wilson.co.ukReviewed-by: NMatthew Auld <matthew.auld@intel.com>

Update the kerneldoc parameter name to match the real parameter name.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20171109140644.10805-1-chris@chris-wilson.co.ukReviewed-by: NMatthew Auld <matthew.auld@intel.com>

As we now record the default HW state and so only emit the "golden"
renderstate once to prepare the HW, there is no advantage in keeping the
renderstate batch around as it will never be used again.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110142634.10551-8-chris@chris-wilson.co.uk

Take a copy of the HW state after a reset upon module loading by
executing a context switch from a blank context to the kernel context,
thus saving the default hw state over the blank context image.
We can then use the default hw state to initialise any future context,
ensuring that each starts with the default view of hw state.

v2: Unmap our default state from the GTT after stealing it from the
context. This should stop us from accidentally overwriting it via the
GTT (and frees up some precious GTT space).

Testcase: igt/gem_ctx_isolation
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110142634.10551-7-chris@chris-wilson.co.uk

In the next few patches, we will want to both copy out of the context
image and write a valid image into a new context. To be completely safe,
we should then couple in our domain tracking to ensure that we don't
have any issues with stale data remaining in unwanted cachelines.

Historically, we omitted the .write=true from the call to set-gtt-domain
in i915_switch_context() in order to avoid a stall between every request
as we would want to wait for the previous context write from the gpu.
Since then, we limit the set-gtt-domain to only occur when we first bind
the vma, so once in use we will never stall, and we are sure to flush
the context following a load from swap.

Equally we never applied the lessons learnt from ringbuffer submission
to execlists; so time to apply the flush of the lrc after load as well.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Acked-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NMika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110142634.10551-6-chris@chris-wilson.co.uk

intel_modeset_gem_init() now only sets up the legacy overlay, so let's
remove the function and call the setup directly during driver load. This
should help us find a better point in the initialisation sequence for it
later.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NMika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110142634.10551-5-chris@chris-wilson.co.uk

Despite its name intel_init_clock_gating applies both display clock gating
workarounds; GT mmio workarounds and the occasional GT power context
workaround. Worse, sometimes it includes a context register workaround
which we need to apply before we record the default HW state for all
contexts.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: NVille Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110142634.10551-4-chris@chris-wilson.co.uk

GT powersaving is tightly coupled to the request infrastructure. To
avoid complications with the order of initialisation in the next patch
(where we want to send requests to hw during GEM init) move the
powersaving initialisation into the purview of i915_gem_init().
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: NJoonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110142634.10551-3-chris@chris-wilson.co.uk

In the next few patches, we will have a hard requirement that we emit a
context-switch to the perma-pinned i915->kernel_context (so that we can
save the HW state using that context-switch). As the first context
itself may be classed as a kernel context, we want to be explicit in our
comparison. For an extra-layer of finesse, we can check the last
unretired context on the engine; as well as the last retired context
when idle.

v2: verbose verbosity
v3: Always force the switch, even when the engine is idle, and update
the assert that this happens before suspend.
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> #v1
Reviewed-by: NMika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110142634.10551-2-chris@chris-wilson.co.uk

We want to be able to report back to userspace details about an engine's
class, and in return for userspace to be able to request actions
regarding certain classes of engines. To isolate the uABI from any
variations between hw generations, we define an abstract class for the
engines and internally map onto the hw.

v2: Remove MAX from the uABI; keep it internal if we need it, but don't
let userspace make the mistake of using it themselves.
v3: s/OTHER/INVALID/
  The use of OTHER is ill-defined, so remove it from the uABI as any
  future new type of engine can define a class to suit it. But keep a
  reserved value for an invalid class, so that we can always
  unambiguously express when something doesn't belong to the
  classification.
Signed-off-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> #v2
Reviewed-by: NLionel Landwerlin <lionel.g.landwerlin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110142634.10551-1-chris@chris-wilson.co.uk

lockdep spotted that the mock tests were using the i915->mm.obj_lock
without first initialiasing it:

>[ 1303.217043] [IGT] drv_selftest: starting subtest mock_objects
<4>[ 1303.240898] Setting dangerous option mock_selftests - tainting kernel
<6>[ 1303.253665] i915: Performing mock selftests with st_random_seed=0xd87ea6c6 st_timeout=1000
<4>[ 1303.254812] INFO: trying to register non-static key.
<4>[ 1303.254816] the code is fine but needs lockdep annotation.
<4>[ 1303.254818] turning off the locking correctness validator.
<4>[ 1303.254820] CPU: 4 PID: 13112 Comm: drv_selftest Tainted: G     U  W       4.14.0-rc8-CI-Patchwork_7058+ #1
<4>[ 1303.254823] Hardware name: MSI MS-7924/Z97M-G43(MS-7924), BIOS V1.12 02/15/2016
<4>[ 1303.254825] Call Trace:
<4>[ 1303.254829]  dump_stack+0x68/0x9f
<4>[ 1303.254832]  register_lock_class+0x3fd/0x580
<4>[ 1303.254835]  ? ___slab_alloc.constprop.29+0x157/0x3d0
<4>[ 1303.254837]  ? ___slab_alloc.constprop.29+0x157/0x3d0
<4>[ 1303.254840]  ? sg_kmalloc+0x1e/0x50
<4>[ 1303.254842]  ? debug_smp_processor_id+0x17/0x20
<4>[ 1303.254845]  __lock_acquire+0xa4/0x1b00
<4>[ 1303.254884]  ? __i915_gem_object_set_pages+0x116/0x1f0 [i915]
<4>[ 1303.254887]  ? __this_cpu_preempt_check+0x13/0x20
<4>[ 1303.254889]  ? sg_kmalloc+0x1e/0x50
<4>[ 1303.254891]  lock_acquire+0xb0/0x200
<4>[ 1303.254893]  ? lock_acquire+0xb0/0x200
<4>[ 1303.254917]  ? __i915_gem_object_set_pages+0x116/0x1f0 [i915]
<4>[ 1303.254920]  _raw_spin_lock+0x32/0x50
<4>[ 1303.254944]  ? __i915_gem_object_set_pages+0x116/0x1f0 [i915]
<4>[ 1303.254967]  __i915_gem_object_set_pages+0x116/0x1f0 [i915]
<4>[ 1303.254991]  i915_gem_object_get_pages_phys+0x286/0x2b0 [i915]
<4>[ 1303.255015]  ____i915_gem_object_get_pages+0x20/0x60 [i915]
<4>[ 1303.255039]  i915_gem_object_attach_phys+0x137/0x1a0 [i915]
<4>[ 1303.255063]  igt_phys_object+0x45/0x120 [i915]
<4>[ 1303.255094]  __i915_subtests+0x40/0xd0 [i915]
<4>[ 1303.255099]  ? work_on_cpu_safe+0x60/0x60
<4>[ 1303.255131]  i915_gem_object_mock_selftests+0x34/0x50 [i915]
Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110151919.18451-1-chris@chris-wilson.co.ukReviewed-by: NMatthew Auld <matthew.william.auld@gmail.com>
Reviewed-by: NTvrtko Ursulin <tvrtko.ursulin@intel.com>

So it appears that commit 5427f207 ("drm/i915: Bump wait-times for
the final CS interrupt before parking") was a little over optimistic in
its belief that it had successfully waited for all residual activity on
the engines before parking. Numerous sightings in CI since then of

<7>[   52.542886] [IGT] core_auth: executing
<3>[   52.561013] [drm:intel_engines_park [i915]] *ERROR* vcs0 is not idle before parking
<7>[   52.561215] intel_engines_park vcs0
<7>[   52.561229] intel_engines_park 	current seqno 98, last 98, hangcheck 0 [-247449 ms], inflight 0
<7>[   52.561238] intel_engines_park 	Reset count: 0
<7>[   52.561266] intel_engines_park 	Requests:
<7>[   52.561363] intel_engines_park 	RING_START: 0x00000000 [0x00000000]
<7>[   52.561377] intel_engines_park 	RING_HEAD:  0x00000000 [0x00000000]
<7>[   52.561390] intel_engines_park 	RING_TAIL:  0x00000000 [0x00000000]
<7>[   52.561406] intel_engines_park 	RING_CTL:   0x00000000
<7>[   52.561422] intel_engines_park 	RING_MODE:  0x00000200 [idle]
<7>[   52.561442] intel_engines_park 	ACTHD:  0x00000000_00000000
<7>[   52.561459] intel_engines_park 	BBADDR: 0x00000000_00000000
<7>[   52.561474] intel_engines_park 	Execlist status: 0x00000301 00000000
<7>[   52.561489] intel_engines_park 	Execlist CSB read 5 [5 cached], write 5 [5 from hws], interrupt posted? no
<7>[   52.561500] intel_engines_park 		ELSP[0] idle
<7>[   52.561510] intel_engines_park 		ELSP[1] idle
<7>[   52.561519] intel_engines_park 		HW active? 0x0
<7>[   52.561608] intel_engines_park Idle? yes
<7>[   52.561617] intel_engines_park

on Braswell, which indicates that the engine just needs that little bit
longer after flushing the tasklet to settle. So give it a few more
milliseconds before declaring an err and applying the emergency brake.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103479Signed-off-by: NChris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171110112550.28909-1-chris@chris-wilson.co.ukReviewed-by: NMika Kuoppala <mika.kuoppala@linux.intel.com>