A little bigger than usual since it's two weeks worth.  Highlights:
- Add support for new smc firmware on some new hainan variants
- add support for SI chips that require special mc firmware
- remove workarounds for issues fixed by new mc firmware
- fix a regression in cursor handling
- various VCE fixes
- fix for UVD clockgating

* 'drm-fixes-4.10' of git://people.freedesktop.org/~agd5f/linux:
  drm/amdgpu: add support for new hainan variants
  drm/radeon: add support for new hainan variants
  drm/amdgpu: change clock gating mode for uvd_v4.
  drm/amdgpu: fix program vce instance logic error.
  drm/amdgpu: fix bug set incorrect value to vce register
  Revert "drm/amdgpu: Only update the CUR_SIZE register when necessary"
  drm/amd/powerplay: refine vce dpm update code on Cz.
  drm/amdgpu: fix vm_fault_stop on gfx6
  drm/amd/powerplay: fix vce cg logic error on CZ/St.
  drm/radeon: drop the mclk quirk for hainan
  drm/radeon: drop oland quirks
  drm/amdgpu: drop the mclk quirk for hainan
  drm/amdgpu: drop oland quirks
  drm/amdgpu/si: load special ucode for certain MC configs
  drm/radeon/si: load special ucode for certain MC configs

* 'msm-fixes-4.10-rc4' of git://people.freedesktop.org/~robclark/linux:
  drm/msm: fix potential null ptr issue in non-iommu case
  drm/msm/mdp5: rip out plane->pending tracking

A few more core fixes.

* tag 'drm-misc-fixes-2017-01-13' of git://anongit.freedesktop.org/git/drm-misc:
  drm/probe-helpers: Drop locking from poll_enable
  drm: Fix broken VT switch with video=1366x768 option
  drm: Schedule the output_poll_work with 1s delay if we have delayed event

Merge tag 'drm-intel-fixes-2017-01-19' of git://anongit.freedesktop.org/git/drm-intel into drm-fixes

More GVT-g stuff than I'd like at this stage, but then again that's
pretty new and isolated so I'm not too worried.

* tag 'drm-intel-fixes-2017-01-19' of git://anongit.freedesktop.org/git/drm-intel: (26 commits)
  drm/i915: Ignore bogus plane coordinates on SKL when the plane is not visible
  drm/i915: Remove WaDisableLSQCROPERFforOCL KBL workaround.
  drm/i915/gvt: rewrite gt reset handler using new function intel_gvt_reset_vgpu_locked
  drm/i915/gvt: fix vGPU instance reuse issues by vGPU reset function
  drm/i915/gvt: introduce intel_vgpu_reset_mmio() to reset mmio space
  drm/i915/gvt: move mmio init/clean function to mmio.c
  drm/i915/gvt: introduce intel_vgpu_reset_cfg_space to reset configuration space
  drm/i915/gvt: move cfg space inititation function to cfg_space.c
  drm/i915/gvt: introuduce intel_vgpu_reset_gtt() to reset gtt
  drm/i915/gvt: introudce intel_vgpu_reset_resource() to reset vgpu resource state
  drm/i915: Fix phys pwrite for struct_mutex-less operation
  drm/i915: Clear ret before unbinding in i915_gem_evict_something()
  drm/i915/gvt: cleanup GFP flags
  drm/i915/gvt/kvmgt: return meaningful error for vgpu creating failure
  drm/i915/gvt: cleanup opregion memory allocation code
  drm/i915/gvt: destroy the allocated idr on vgpu creating failures
  drm/i915/gvt: init/destroy vgpu_idr properly
  drm/i915/gvt: dec vgpu->running_workload_num after the workload is really done
  drm/i915/gvt: fix use after free for workload
  drm/i915/gvt: remove duplicated definition
  ...

When the plane is invisible we may have all sorts of bogus stuff
in the coordinates, which we must ignore or else we might fail the
plane update. This started to happen on SKL when I moved the plane
offset computation to happen in the check phase. Previously we
happily ignored it all since we never called the update_plane hook
with an invisible plane.

Cc: Sivakumar Thulasimani <sivakumar.thulasimani@intel.com>
Cc: drm-intel-fixes@lists.freedesktop.org
Fixes: b63a16f6 ("drm/i915: Compute display surface offset in the plane check hook for SKL+")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98258
Testcase: igt/pm_rpm/legacy-planes
Testcase: igt/pm_rpm/universal-planes
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1478550057-24864-3-git-send-email-ville.syrjala@linux.intel.com
(cherry picked from commit a5e4c7d0)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

The WaDisableLSQCROPERFforOCL workaround has the side effect of
disabling an L3SQ optimization that has huge performance implications
and is unlikely to be necessary for the correct functioning of usual
graphic workloads.  Userspace is free to re-enable the workaround on
demand, and is generally in a better position to determine whether the
workaround is necessary than the DRM is (e.g. only during the
execution of compute kernels that rely on both L3 fences and HDC R/W
requests).

The same workaround seems to apply to BDW (at least to production
stepping G1) and SKL as well (the internal workaround database claims
that it does for all steppings, while the BSpec workaround table only
mentions pre-production steppings), but the DRM doesn't do anything
beyond whitelisting the L3SQCREG4 register so userspace can enable it
when it sees fit.  Do the same on KBL platforms.

Improves performance of the GFXBench4 gl_manhattan31 benchmark by 60%,
and gl_4 (AKA car chase) by 14% on a KBL GT2 running Mesa master --
This is followed by a regression of 35% and 10% respectively for the
same benchmarks and platform caused by my recent patch series
switching userspace to use the dataport constant cache instead of the
sampler to implement uniform pull constant loads, which caused us to
hit more heavily the L3 cache (and on platforms other than KBL had the
opposite effect of improving performance of the same two benchmarks).
The overall effect on KBL of this change combined with the recent
userspace change is respectively 4.6% and 2.6%.  SynMark2 OglShMapPcf
was affected by the constant cache changes (though it improved as it
did on other platforms rather than regressing), but is not
significantly affected by this patch (with statistical significance of
5% and sample size 20).

v2: Drop some more code to avoid unused variable warning.

Fixes: 738fa1b3 ("drm/i915/kbl: Add WaDisableLSQCROPERFforOCL")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99256Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Cc: Eero Tamminen <eero.t.tamminen@intel.com>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: beignet@lists.freedesktop.org
Cc: <stable@vger.kernel.org> # v4.7+
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
[Removed double Fixes tag]
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1484217894-20505-1-git-send-email-mika.kuoppala@intel.com
(cherry picked from commit 8726f2fa)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

New hainan parts require updated smc firmware.

Cc: Sonny Jiang <sonny.jiang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

New hainan parts require updated smc firmware.

Cc: Sonny Jiang <sonny.jiang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

use sw cg when decode. and hw cg when idle.

fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=99313
https://bugzilla.kernel.org/show_bug.cgi?id=192161Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Ack-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

need to clear bit31-29 in GRBM_GFX_INDEX,
then the program can be valid.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Set the proper bits for clockgating setup.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

gvt-fixes-2017-01-16

vGPU reset fixes from Changbin.
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

This reverts commits 7c83d7ab and
a1f49cc1.

They caused the HW cursor to disappear under various circumstances in
the wild. I wasn't able to reproduce any of them, and I'm not sure
what's going on. But those changes aren't a big deal anyway, so let's
just revert for now.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=191291
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99143Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Fixes: 9cb07b099fb ("drm/msm: support multiple address spaces")
Reported-by: Riku Voipio <riku.voipio@linaro.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>

It would race between userspace thread and commit worker.  Ie. vblank
irq would trigger event and userspace could begin the next atomic
update, before the commit worker had a chance to clear the pending
flag.

If we do end up needing something to prevent userspace from trying
another pageflip before getting vblank event, it should probably be
implemented as a pending_planes bitmask, similar to pending_crtcs.  See
start_atomic() and end_atomic().
Signed-off-by: Rob Clark <robdclark@gmail.com>

GT reset and FLR share some operations and they are both implemented in
our new function intel_gvt_reset_vgpu_locked(). This patch rewrite the
gt reset handler using this new function.

Besides, this new implementation fixed the old issue in GT reset. The
old implementation reset GGTT entries which is illegal. We only clear
GGTT entries at PCI level reset.
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

Our function tests found several issues related to reusing vGPU
instance. They are qemu reboot failure, guest tdr after reboot, host
hang when reboot guest. All these issues are caused by dirty status
inherited from last VM.

This patch fix all these issues by resetting a virtual GPU before VM
use it. The reset logical is put into a low level function
_intel_gvt_reset_vgpu(), which supports Device Model Level Reset, Full
GT Reset and Per-Engine Reset.

vGPU Device Model Level Reset (DMLR) simulates the PCI reset to reset
the whole vGPU to default state as when it is created, including GTT,
execlist, scratch pages, cfg space, mmio space, pvinfo page, scheduler
and fence registers. The ultimate goal of vGPU DMLR is that reuse a
vGPU instance by different virtual machines. When we reassign a vGPU
to a virtual machine we must issue such reset first.

Full GT Reset and Per-Engine GT Reset are soft reset flow for GPU engines
(Render, Blitter, Video, Video Enhancement). It is defined by GPU Spec.
Unlike the FLR, GT reset only reset particular resource of a vGPU per
the reset request. Guest driver can issue a GT reset by programming
the virtual GDRST register to reset specific virtual GPU engine or all
engines.

Since vGPU DMLR and GT reset can share some code so we implement both
these two into one single function intel_gvt_reset_vgpu_locked(). The
parameter dmlr is to identify if we will do FLR or GT reset. The
parameter engine_mask is to specific the engines that need to be
resetted. If value ALL_ENGINES is given for engine_mask, it means
the caller requests a full gt reset that we will reset all virtual
GPU engines.
Signed-off-by: Changbin Du <changbin.du@intel.com>
Reviewed-by: Jike Song <jike.song@intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

This patch introduces a new function intel_vgpu_reset_mmio() to
reset vGPU MMIO space (virtual registers of the vGPU). The default
values are loaded as firmware during gvt inititiation.
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

Move the mmio space inititation function setup_vgpu_mmio()
and cleanup function clean_vgpu_mmio() in vgpu.c to dedicated
source file mmio.c, and rename them as intel_vgpu_init_mmio()
and intel_vgpu_clean_mmio() respectively.
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

This patch introduces a new function intel_vgpu_reset_cfg_space()
to reset vGPU configuration space. This function will unmap gttmmio
and aperture if they are mapped before. Then entire cfg space will
be restored to default values.

Currently we only do such reset when vGPU is not owned by any VM
so we simply restore entire cfg space to default value, not following
the PCIe FLR spec that some fields should remain unchanged.
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

Move the configuration space inititation function setup_vgpu_cfg_space()
in vgpu.c to dedicated source file cfg_space.c, and rename the function
as intel_vgpu_init_cfg_space().
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

This patch introduces a new function intel_vgpu_reset_gtt() to reset
the all GTT related status, including GGTT, PPGTT, scratch page. This
function can free all shadowed PPGTT, clear all GGTT entry, and clear
scratch page to all zero. After this, we can ensure no gtt related
information can be leakaged from one VM to anothor one when assign
vgpu instance across different VMs (not simultaneously).
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

This patch introudces a new function intel_vgpu_reset_resource() to
reset allocated vgpu resources by intel_vgpu_alloc_resource(). So far
we only need clear the fence registers. The function _clear_vgpu_fence()
will reset both virtual and physical fence registers to 0.
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>

Program HardMin based on the vce_arbiter.ecclk
if ecclk is 0, disable ECLK DPM 0. Otherwise VCE
could hang if switching SCLK from DPM 0 to 6/7
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Signed-off-by: Flora Cui <Flora.Cui@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

can fix Bug 191281: vce ib test failed.

when vce idle, set vce clock gate, so the clock
in vce domain will be disabled.
when need to encode, disable vce clock gate,
enable the clocks to vce engine.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

fixed by the new 58 mc firmware.
Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Fixed by the new 58 MC firmware.
Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

fixed by the new 58 mc firmware.
Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Fixed by the new 58 MC firmware.
Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Special MC ucode is required for these memory configurations.
Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Special MC ucode is required for these memory configurations.
Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

It was only needed to protect the connector_list walking, see

commit 8c4ccc4a
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Thu Jul 9 23:44:26 2015 +0200

    drm/probe-helper: Grab mode_config.mutex in poll_init/enable

Unfortunately the commit message of that patch fails to mention that
the new locking check was for the connector_list.

But that requirement disappeared in

commit c36a3254
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Thu Dec 15 16:58:43 2016 +0100

    drm: Convert all helpers to drm_connector_list_iter

and so we can drop this again.

This fixes a locking inversion on nouveau, where the rpm code needs to
re-enable. But in other places the rpm_get() calls are nested within
the big modeset locks.

While at it, also improve the kerneldoc for these two functions a
notch.

v2: Update the kerneldoc even more to explain that these functions
can't be called concurrently, or bad things happen (Chris).

Cc: Dave Airlie <airlied@gmail.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Lyude <lyude@redhat.com>
Reviewed-by: Lyude <lyude@redhat.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170111090117.5134-1-daniel.vetter@ffwll.ch

Since commit fe115628 ("drm/i915: Implement pwrite without
struct-mutex") the lowlevel pwrite calls are now called without the
protection of struct_mutex, but pwrite_phys was still asserting that it
held the struct_mutex and later tried to drop and relock it.

Fixes: fe115628 ("drm/i915: Implement pwrite without struct-mutex")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: <drm-intel-fixes@lists.freedesktop.org>
Link: http://patchwork.freedesktop.org/patch/msgid/20170106152240.5793-1-chris@chris-wilson.co.ukReviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
(cherry picked from commit 10466d2a)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Missed when rebasing patches, I failed to set ret to zero before
starting the unbind loop (which depends upon ret being zero).
Reported-by: Matthew Auld <matthew.william.auld@gmail.com>
Fixes: 9332f3b1 ("drm/i915: Combine loops within i915_gem_evict_something")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.william.auld@gmail.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170105155940.10033-1-chris@chris-wilson.co.ukReviewed-by: Matthew Auld <matthew.william.auld@gmail.com>
Cc: <stable@vger.kernel.org> # v4.9+
(cherry picked from commit 121dfbb2)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

I noticed that the VT switch doesn't work any longer with a Dell
laptop with 1366x768 eDP when the machine is connected with a DP
monitor.  It behaves as if VT were switched, but the graphics remain
frozen.  Actually the keyboard works, so I could switch back to VT7
again.

I tried to track down the problem, and encountered a long story until
we reach to this error:

- The machine is booted with video=1366x768 option (the distro
  installer seems to add it as default).
- Recently, drm_helper_probe_single_connector_modes() deals with
  cmdline modes, and it tries to create a new mode when no
  matching mode is found.
- The drm_mode_create_from_cmdline_mode() creates a mode based on
  either CVT of GFT according to the given cmdline mode; in our case,
  it's 1366x768.
- Since both CVT and GFT can't express the width 1366 due to
  alignment, the resultant mode becomes 1368x768, slightly larger than
  the given size.
- Later on, the atomic commit is performed, and in
  drm_atomic_check_only(), the size of each plane is checked.
- The size check of 1366x768 fails due to the above, and eventually
  the whole VT switch fails.

Back in the history, we've had a manual fix-up of 1368x768 in various
places via c09dedb7 ("drm/edid: Add a workaround for 1366x768 HD
panel"), but they have been all in drm_edid.c at probing the modes
from EDID.  For addressing the problem above, we need a similar hack
to the mode newly created from cmdline, manually adjusting the width
when the expected size is 1366 while we get 1368 instead.

Fixes: eaf99c74 ("drm: Perform cmdline mode parsing during...")
Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Link: http://patchwork.freedesktop.org/patch/msgid/20170109145614.29454-1-tiwai@suse.deReviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Instead of scheduling the work to handle the initial delayed event, use 1s
delay.

This delay should not be needed, but Optimus/nouveau will fail in a
mysterious way if the delayed event is handled as soon as possible like it
is done in drm_helper_probe_single_connector_modes() in case the poll
was enabled before.

Reverting 339fd362 would give back the 10 sec (!) delay to handle the
delayed event. Adding 1sec delay to the poll_work is enough to work around
the issue in Optimus setups and gives shorter response on handling the
initial delayed event.

Fixes: 339fd362 ("drm: drm_probe_helper: Fix output_poll_work scheduling")
Cc: stable@vger.kernel.org   # v4.9
Signed-off-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
[danvet: Add FIXME to the comment to make it stick out more.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20170109143158.21917-1-peter.ujfalusi@ti.com

GVT-g fixes from Zhenya, "Please pull GVT-g device model fixes for
rc4. This is based on rc3 with new vfio/mdev interface change."
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

Single drm bridge fix.

* tag 'drm-misc-fixes-2017-01-09' of git://anongit.freedesktop.org/git/drm-misc:
  drm/bridge: analogix dp: Fix runtime PM state on driver bind

In gvt, almost all memory allocations are in sleepable contexts. It's
fault-prone to use GFP_ATOMIC everywhere. Replace it with GFP_KERNEL
wherever possible.
Signed-off-by: Jike Song <jike.song@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>