Merge tag 'drm-intel-gt-next-2021-01-21-1' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

Cross-subsystem Changes:

- Includes gvt-gt-next-2021-01-18 + header check fix for GVT

Driver Changes:

- Fix for #2955: Clear potentially malicious register state before
  executing clear residuals security mitigation (Chris)
- Fixes that lead to marking per-engine-reset as supported on Gen7
  (Chris)
- Remove per-client stats from debugfs/i915_gem_objects) (Tvrtko, Chris)
- Add arbitration check before semaphore wait (Chris)
- Apply interactive priority to explicit flip fences (Chris)
- Make GEM errors non-fatal by default to help capturing logs during
  development (Chris)
- Fix object page offset within a region in error capture (CQ, Matt A)
- Close race between enable_breadcrumbs and cancel_breadcrumbs (Chris)
- Almagamate clflushes on suspend/freeze to speed up S/R (Chris)
- Protect used framebuffers from casual eviction (Chris)

- Fix the sgt.pfn sanity check (Kui, Matt A)
- Reduce locking around i915_request.lock and ctx->engines_mutex (Chris)
- Simplify tracking for engine->fw_active and stats.active (Chris)
- Constrain pool objects by mapping type (Chris, Matt A)
- Use shrinkable status for unknown swizzle quirks (Chris)
- Do not suspend bonded requests if one hangs (Chris)
- Restore "Skip over completed active execlists" optimization (Chris)

- Move stolen node into GEM object union (Chris)
. Split gem_create into own file (Matt A)
- Convert object_create into object_init in LMEM region code (Matt A)
- Reduce test_and_set_bit to set_bit in i915_request_submit() (Chris)
- Mark up protected uses of 'i915_request_completed' (Chris)
- Remove extraneous inline modifiers (Chris)
- Add function to define defaults for GuC/HuC enable (John)

- Improve code locality by moving closer to single user (Matt A, Chris)
- Compiler warning fixes (Matt A, Chris)
- Selftest / CI improvements (Chris)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210121150747.GA58732@jlahtine-mobl.ger.corp.intel.com

drm/i915/gvt: Add missing forward decl of intel_vgpu for HDRTEST

./drivers/gpu/drm/i915/gvt/cmd_parser.h:53:44: error: ‘struct intel_vgpu’ declared inside parameter list will not be visible outside of this definition or declaration

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210121132558.2893-1-chris@chris-wilson.co.uk

Merge tag 'gvt-gt-next-2021-01-18' of https://github.com/intel/gvt-linux into drm-intel-gt-next

gvt-gt-next-2021-01-18

- GVT cmd parser enhancement against guest context (Yan)

Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
From: Zhenyu Wang <zhenyuw@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210118050739.GY15982@zhen-hp.sh.intel.com

drm/i915/gt: Skip over completed active execlists, again

Now that we are careful to always force-restore contexts upon rewinding
(where necessary), we can restore our optimisation to skip over
completed active execlists when dequeuing.

Referenecs: 35f3fd8182ba ("drm/i915/execlists: Workaround switching back to a completed context")
References: 8ab3a3812aa9 ("drm/i915/gt: Incrementally check for rewinding")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210120121718.26435-2-chris@chris-wilson.co.uk

drm/i915/gt: Do not suspend bonded requests if one hangs

Treat the dependency between bonded requests as weak and leave the
remainder of the pair on the GPU if one hangs.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210120121718.26435-1-chris@chris-wilson.co.uk

drm/i915/gem: Drop lru bumping on display unpinning

Simplify the frontbuffer unpin by removing the lock requirement. The LRU
bumping was primarily to protect the GTT from being evicted and from
frontbuffers being eagerly shrunk. Now we protect frontbuffers from the
shrinker, and we avoid accidentally evicting from the GTT, so the
benefit from bumping LRU is no more, and we can save more time by not.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119214336.1463-6-chris@chris-wilson.co.uk

drm/i915/gem: Protect used framebuffers from casual eviction

In the shrinker, we protect framebuffers from light reclaim as we
typically expect framebuffers to be reused in the near future (and with
low latency requirements). We can apply the same logic to the GGTT
eviction and defer framebuffers to the second pass only used if the
caller is desperate enough to wait for space to become available.
In most cases, the caller will use a smaller partial vma instead of
trying to force the object into the GGTT if doing so will cause other
users to be evicted.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119214336.1463-5-chris@chris-wilson.co.uk

drm/i915/gem: Use shrinkable status for unknown swizzle quirks

Give obj->mm.quirked a name much more reflective of its purpose
(i915_gem_object_has_tiling_quirk) and move it from the obj->mm field as
it doesn't denote a quirk of the backing store, but a quirk in the
object in its treatment of the backing pages, similar to tiling modes.

Then instead of abusing the pinned status of the buffer to protect it
from the shrinker, we can instead hide the buffer from the shrinker so
it is never considered for being swapped.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119214336.1463-4-chris@chris-wilson.co.uk

drm/i915/gem: Move stolen node into GEM object union

The obj->stolen is currently used to identify an object allocated from
stolen memory. This dates back to when there were just 1.5 types of
objects, an object backed by shmemfs and an object backed by shmemfs
with a contiguous physical address. Now that we have several different
types of objects, we no longer want to treat stolen objects as a special
case.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119214336.1463-3-chris@chris-wilson.co.uk

drm/i915/gem: Almagamate clflushes on freeze

When flushing objects larger than the CPU cache it is preferrable to use
a single wbinvd() rather than overlapping clflush(). At runtime, we
avoid wbinvd() due to its system-wide latencies, but during
singlethreaded suspend, no one will observe the imposed latency and we
can opt for the faster wbinvd to clear all objects in a single hit.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119214336.1463-2-chris@chris-wilson.co.uk

drm/i915/gem: Almagamate clflushes on suspend

When flushing objects larger than the CPU cache it is preferrable to use
a single wbinvd() rather than overlapping clflush(). At runtime, we
avoid wbinvd() due to its system-wide latencies, but during
singlethreaded suspend, no one will observe the imposed latency and we
can opt for the faster wbinvd to clear all objects in a single hit.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119214336.1463-1-chris@chris-wilson.co.uk

Merge tag 'amd-drm-next-5.12-2021-01-20' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-5.12-2021-01-20:

amdgpu:
- Fix non-x86 build
- W=1 fixes from Lee Jones
- Enable GPU reset on Navy Flounder
- Kernel doc fixes
- SMU workload profile fixes for APUs
- Display updates
- SR-IOV fixes
- Vangogh SMU feature enablment and bug fixes
- GPU reset support for Vangogh
- Misc cleanups

Conflicts:
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_mpc.c

Resolve the conflict by picking the initialization value from amd from
f03e80d2e82c ("drm/amd/display: Initialize stack variable") over the
one Linus picked in 61d791365b72 ("drm/amd/display: avoid
uninitialized variable warning"). It shouldn't matter.

From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210120060951.22600-1-alexander.deucher@amd.com
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>

drm/i915/region: don't leak the object on error

Sanity check the object size before allocating a new gem object.

Fixes: 97d553963250 ("drm/i915/region: convert object_create into object_init")
Testcase: igt/gem_create/create-massive
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210120104714.112812-1-matthew.auld@intel.com

drm/i915/uc: Add function to define defaults for GuC/HuC enable

There is a module parameter for controlling what GuC/HuC features are
enabled. Setting to -1 means 'use the default'. However, the default
was not well defined, out of date and needs to be different across
platforms.

The default is now to disable both GuC and HuC on legacy platforms
where legacy means TGL/RKL and anything prior to Gen12. For new
platforms, the default is to load HuC but not enable GuC submission
as that has not landed yet.

Signed-off-by: John Harrison <John.C.Harrison@Intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210113220724.2484897-1-John.C.Harrison@Intel.com

Merge tag 'drm-misc-next-2021-01-19' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

drm-misc-next for v5.12:

UAPI Changes:
- Fix fourcc macro for amlogic video fbc.

Cross-subsystem Changes:
- Export pci_rebar_bytes_to_size.
- Add a PCI quirk to increase bar0 for RX 5600 XT Pulse to max possible size.
- Convert devicetree bindings to use the OF graph schema.
- Update s6e63m0 bindings.
- Make omapfb2 DSI_CM incompatible with drm/omap2 DSI-CM because of
  module conflicts.
- Add Zack Rusin as vmwgfx maintainer.
- Add CONFIG_DMABUF_DEBUG for validating dma-buf users don't loo kat struct page when importing or detaching.

Core Changes:
- Remove references to drm_device.pdev
- Fix regression in ttm_bo_move_to_lru_tail().
- Assorted docbook updates.
- Do not send dp-mst hotplug events on error when probing.
- Move some agp macros to agpsupport.c, so it's not always compiled.
- Move drm_need_swiotlb.h to drm_cache.c
- Only build drm_memory.o for legacy drivers, and move CONFIG_DRM_VM to legacy.
- Nuke drm_device.hose
- Warn when the ttm resource manager is non-empty when disabling.
- Assorted small fixes.

Driver Changes:
- Small assorted fixes in radeon, v3d, hisilicon, mipi-dbi, panfrost, hibmc, vc4, amdgpu, vkms, vmwgfx.
- Move hisilicon to use simple encode.
- Add writeback connector to vkms.
- Add support for BT2020 to DE3.
- Use gem prime mmap helpers in vc4, and move the mmap function upwards.
- Use managed drm device, and cleanup error paths and display registers in vmwgfx.
- Use correct bus_format and connector_type for innolux_n116bge.
- Fix a lot of warnings with W=1 (Lee Jones)

Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/5c3ad775-48ce-33ee-e4c6-a5e1e540f845@linux.intel.com

drm/i915/gt: Close race between enable_breadcrumbs and cancel_breadcrumbs

If we enable_breadcrumbs for a request while that request is being
removed from HW; we may see that the request is active as we take the
ce->signal_lock and proceed to attach the request to ce->signals.
However, during unsubmission after marking the request as inactive, we
see that the request has not yet been added to ce->signals and so skip
the removal. Pull the check during cancel_breadcrumbs under the same
spinlock as enabling so that we the two tests are consistent in
enable/cancel.

Otherwise, we may insert a request onto ce->signals that we expect should
not be there:

  intel_context_remove_breadcrumbs:488 GEM_BUG_ON(!__i915_request_is_complete(rq))

While updating, we can note that we are always called with
irqs-disabled, due to the engine->active.lock being held at the single
caller, and so remove the irqsave/restore making it symmetric to
enable_breadcrumbs.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2931
Fixes: c18636f76344 ("drm/i915: Remove requirement for holding i915_request.lock for breadcrumbs")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Andi Shyti <andi.shyti@intel.com>
Cc: <stable@vger.kernel.org> # v5.10+
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119162057.31097-1-chris@chris-wilson.co.uk

drm/i915/display: Apply interactive priority to explicit flip fences

Currently, if a modeset/pageflip needs to wait for render completion to
an object, we boost the priority of that rendering above all other work.
We can apply the same interactive priority boosting to explicit fences
that we can unwrap into a native i915_request (i.e. sync_file).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119204454.10343-1-chris@chris-wilson.co.uk

drm/i915/gem: Make i915_gem_object_flush_write_domain() static

flush_write_domain() is only used within the GEM domain management code,
so move it to i915_gem_domain.c and drop the export.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119144912.12653-5-chris@chris-wilson.co.uk

drm/i915/pool: constrain pool objects by mapping type

In a few places we always end up mapping the pool object with the FORCE
constraint(to prevent hitting -EBUSY) which will destroy the cached
mapping if it has a different type. As a simple first step, make the
mapping type part of the pool interface, where the behaviour is to only
give out pool objects which match the requested mapping type.

Suggested-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119133106.66294-4-matthew.auld@intel.com

drm/i915: move i915_map_type into i915_gem_object_types.h

Looks like it belongs there anyway, otherwise we have to include the
entirety of i915_gem_object.h just to get at the enum.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119133106.66294-3-matthew.auld@intel.com

drm/i915/error: Fix object page offset within a region

io_mapping_map_wc() expects the offset to be relative to the iomapping
base address. Currently we just pass in the physical address for the
page which only works if the region.start starts at zero.

Signed-off-by: CQ Tang <cq.tang@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119133106.66294-2-matthew.auld@intel.com

drm/i915: Fix the sgt.pfn sanity check

For the device local-memory case, sgt.pfn will always be equal to zero,
since we instead use sgt.dma. Also, for device local-memory it is
perfectly valid for it to start from zero anyway, so no need to add a
new check for that either.

Signed-off-by: Kui Wen <kui.wen@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119133106.66294-1-matthew.auld@intel.com

drm/amdgpu: fix build error without x86 kconfig (v2)

This patch is to fix below build error while we are using the kconfig
without x86.

drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu11/vangogh_ppt.c: In function
'vangogh_get_smu_metrics_data':
drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu11/vangogh_ppt.c:300:10:
error: 'boot_cpu_data' undeclared (first use in this function); did you
mean 'boot_cpuid'?
  300 |          boot_cpu_data.x86_max_cores * sizeof(uint16_t));
      |          ^~~~~~~~~~~~~
      |          boot_cpuid
drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu11/vangogh_ppt.c: In function
'vangogh_read_sensor':
drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu11/vangogh_ppt.c:1320:11:
error: 'boot_cpu_data' undeclared (first use in this function); did you
mean 'boot_cpuid'?
 1320 |   *size = boot_cpu_data.x86_max_cores * sizeof(uint16_t);
      |           ^~~~~~~~~~~~~
      |           boot_cpuid
drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu11/vangogh_ppt.c: In function
'vangogh_od_edit_dpm_table':
drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu11/vangogh_ppt.c:1460:19:
error: 'boot_cpu_data' undeclared (first use in this function); did you
mean 'boot_cpuid'?
 1460 |   if (input[0] >= boot_cpu_data.x86_max_cores) {
      |                   ^~~~~~~~~~~~~
      |                   boot_cpuid

v2: fix #ifdef and add comment for APU only

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/i915/gem: Remove per-client stats from debugfs/i915_gem_objects

Rather than trying to avoid the use-after-free possible with the current
context chasing, simply remove presentation of the per-client stats from
debugfs. While we know from bug reports that this debugfs/i915_gem_objects
has been used by chromeos (and chrome itself) for debug purposes, google
suggests that it is unparsed, so we are free to invoke debugfs is not
ABI and remove details from it.

Suggested-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210118110854.1873-1-chris@chris-wilson.co.uk

drm/i915: Mark per-engine-reset as supported on gen7

The benefit of only resetting a single engine is that we leave other
streams of userspace work intact across a hang; vital for process
isolation. We had wired up individual engine resets for gen6, but only
enabled it from gen8; now let's turn it on for the forgotten gen7. gen6
is still a mystery as how to unravel some global state that appears to
be reset along with an engine (in particular the ppgtt enabling in
GFX_MODE).

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119110802.22228-6-chris@chris-wilson.co.uk

drm/i915/gt: Pull ring submission resume under its caller forcewake

Take advantage of calling xcs_resume under a forcewake by using direct
mmio access. In particular, we can avoid the sleeping variants to allow
resume to be called from softirq context, required for engine resets.

v2: Keep the posting read at the start of resume as a guardian memory
barrier.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119110802.22228-5-chris@chris-wilson.co.uk

drm/i915/gt: Disable the ring before resetting HEAD/TAIL

During the reset of ring submission, we first stop the engine by
clearing the HEAD/TAIL and marking the ring as disabled. However, it
would be safer to disable the ring (after emptying) before resetting the
HEAD/TAIL.

Suggested-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119110802.22228-4-chris@chris-wilson.co.uk

drm/i915/gt: Lift stop_ring() to reset_prepare

Push the sleeping stop_ring() out of the reset resume function to reset
prepare; we are not allowed to sleep in the former.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119110802.22228-3-chris@chris-wilson.co.uk

drm/i915/selftests: Prepare the selftests for engine resets with ring submission

The engine resets selftests kick the tasklets, safe up until now as only
execlists supported engine resets.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119110802.22228-2-chris@chris-wilson.co.uk

drm/i915/gt: One more flush for Baytrail clear residuals

CI reports that Baytail requires one more invalidate after CACHE_MODE
for it to be happy.

Fixes: ace44e13e577 ("drm/i915/gt: Clear CACHE_MODE prior to clearing residuals")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210119110802.22228-1-chris@chris-wilson.co.uk

drm/i915: Make GEM errors non-fatal by default

While immensely convenient for developing to only tackle the first
error, and not be flooded by repeated or secondiary issues, many more
casual testers are not setup to remotely capture debug traces. For those
testers, it is more beneficial to keep the system running in the remote
chance that they are able to extract the original debug logs.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114113434.8229-2-chris@chris-wilson.co.uk

drm/i915: Add DEBUG_GEM to the recommended CI config

Now that i915 compiles cleanly with Werror, we can enforce enabling
DEBUG_GEM when selecting the default debug config.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jani Nikula <jani.nikula@intel.com>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114113434.8229-1-chris@chris-wilson.co.uk

drm/amdgpu: Remove accidentally added small unused hunk.

Commit 8f66090b7bb7 ("drm/amdgpu: Remove references to struct drm_device.pdev")
accidentally adds an initialization of adev, which isn't used anywhere. This
caused a compiler warning on the unused variable, remove it.

Fixes: 8f66090b7bb7 ("drm/amdgpu: Remove references to struct drm_device.pdev")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>

drm/sched: Cancel and flush all outstanding jobs before finish.

To avoid any possible use after free.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/414814/
CC: stable@vger.kernel.org
Signed-off-by: Christian König <christian.koenig@amd.com>

drm: Unamp the entire device address space on device unplug

Invalidate all BOs CPU mappings once device is removed.

v3: Move the code from TTM into drm_dev_unplug

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/414809/
Signed-off-by: Christian König <christian.koenig@amd.com>

drm: Upcast struct drm_device.dev to struct pci_device; replace pdev

We have DRM drivers based on USB, SPI and platform devices. All of them
are fine with storing their device reference in struct drm_device.dev.
PCI devices should be no exception. Therefore struct drm_device.pdev is
deprecated.

Instead upcast from struct drm_device.dev with to_pci_dev(). PCI-specific
code can use dev_is_pci() to test for a PCI device. This patch changes
the DRM core code and documentation accordingly.

v4:
* split-off pdev deprecation into separate patch

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Tested-by: Andy Lavr <andy.lavr@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210118131420.15874-2-tzimmermann@suse.de

drm/drm_agpsupport: Strip out obviously wrong descriptions and demote to standard headers

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/drm_agpsupport.c:61: warning: Function parameter or member 'dev' not described in 'drm_agp_info'
 drivers/gpu/drm/drm_agpsupport.c:61: warning: Function parameter or member 'info' not described in 'drm_agp_info'
 drivers/gpu/drm/drm_agpsupport.c:105: warning: Function parameter or member 'dev' not described in 'drm_agp_acquire'
 drivers/gpu/drm/drm_agpsupport.c:132: warning: Function parameter or member 'dev' not described in 'drm_agp_acquire_ioctl'
 drivers/gpu/drm/drm_agpsupport.c:132: warning: Function parameter or member 'data' not described in 'drm_agp_acquire_ioctl'
 drivers/gpu/drm/drm_agpsupport.c:132: warning: Function parameter or member 'file_priv' not described in 'drm_agp_acquire_ioctl'
 drivers/gpu/drm/drm_agpsupport.c:145: warning: Function parameter or member 'dev' not described in 'drm_agp_release'
 drivers/gpu/drm/drm_agpsupport.c:171: warning: Function parameter or member 'dev' not described in 'drm_agp_enable'
 drivers/gpu/drm/drm_agpsupport.c:171: warning: Function parameter or member 'mode' not described in 'drm_agp_enable'
 drivers/gpu/drm/drm_agpsupport.c:203: warning: Function parameter or member 'dev' not described in 'drm_agp_alloc'
 drivers/gpu/drm/drm_agpsupport.c:203: warning: Function parameter or member 'request' not described in 'drm_agp_alloc'
 drivers/gpu/drm/drm_agpsupport.c:256: warning: Function parameter or member 'dev' not described in 'drm_agp_lookup_entry'
 drivers/gpu/drm/drm_agpsupport.c:256: warning: Function parameter or member 'handle' not described in 'drm_agp_lookup_entry'
 drivers/gpu/drm/drm_agpsupport.c:279: warning: Function parameter or member 'dev' not described in 'drm_agp_unbind'
 drivers/gpu/drm/drm_agpsupport.c:279: warning: Function parameter or member 'request' not described in 'drm_agp_unbind'
 drivers/gpu/drm/drm_agpsupport.c:318: warning: Function parameter or member 'dev' not described in 'drm_agp_bind'
 drivers/gpu/drm/drm_agpsupport.c:318: warning: Function parameter or member 'request' not described in 'drm_agp_bind'
 drivers/gpu/drm/drm_agpsupport.c:363: warning: Function parameter or member 'dev' not described in 'drm_agp_free'
 drivers/gpu/drm/drm_agpsupport.c:363: warning: Function parameter or member 'request' not described in 'drm_agp_free'
 drivers/gpu/drm/drm_agpsupport.c:404: warning: Function parameter or member 'dev' not described in 'drm_agp_init'

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Faith <faith@valinux.com>
Cc: Gareth Hughes <gareth@valinux.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-24-lee.jones@linaro.org

drm/gma500/mdfld_intel_display: Demote three kernel-doc formatting abuses

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/mdfld_intel_display.c:103: warning: Function parameter or member 'dev' not described in 'psb_intel_panel_fitter_pipe'
 drivers/gpu/drm/gma500/mdfld_intel_display.c:273: warning: Function parameter or member 'crtc' not described in 'mdfld_crtc_dpms'
 drivers/gpu/drm/gma500/mdfld_intel_display.c:273: warning: Function parameter or member 'mode' not described in 'mdfld_crtc_dpms'
 drivers/gpu/drm/gma500/mdfld_intel_display.c:609: warning: Function parameter or member 'crtc' not described in 'mdfldFindBestPLL'
 drivers/gpu/drm/gma500/mdfld_intel_display.c:609: warning: Function parameter or member 'target' not described in 'mdfldFindBestPLL'
 drivers/gpu/drm/gma500/mdfld_intel_display.c:609: warning: Function parameter or member 'refclk' not described in 'mdfldFindBestPLL'
 drivers/gpu/drm/gma500/mdfld_intel_display.c:609: warning: Function parameter or member 'best_clock' not described in 'mdfldFindBestPLL'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-23-lee.jones@linaro.org

drm/gma500/mdfld_dsi_output: Demote a couple of kernel-doc formatting abuses

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/mdfld_dsi_output.c:71: warning: Function parameter or member 'dev' not described in 'mdfld_dsi_gen_fifo_ready'
 drivers/gpu/drm/gma500/mdfld_dsi_output.c:71: warning: Function parameter or member 'gen_fifo_stat_reg' not described in 'mdfld_dsi_gen_fifo_ready'
 drivers/gpu/drm/gma500/mdfld_dsi_output.c:71: warning: Function parameter or member 'fifo_stat' not described in 'mdfld_dsi_gen_fifo_ready'
 drivers/gpu/drm/gma500/mdfld_dsi_output.c:95: warning: Function parameter or member 'dsi_config' not described in 'mdfld_dsi_brightness_init'
 drivers/gpu/drm/gma500/mdfld_dsi_output.c:95: warning: Function parameter or member 'pipe' not described in 'mdfld_dsi_brightness_init'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: jim liu <jim.liu@intel.com>
Cc: Jackie Li <yaodong.li@intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-22-lee.jones@linaro.org

drm/gma500/oaktrail_crtc: Demote unworthy kernel-doc headers

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/oaktrail_crtc.c:184: warning: Function parameter or member 'limit' not described in 'mrst_lvds_find_best_pll'
 drivers/gpu/drm/gma500/oaktrail_crtc.c:184: warning: Function parameter or member 'crtc' not described in 'mrst_lvds_find_best_pll'
 drivers/gpu/drm/gma500/oaktrail_crtc.c:184: warning: Function parameter or member 'target' not described in 'mrst_lvds_find_best_pll'
 drivers/gpu/drm/gma500/oaktrail_crtc.c:184: warning: Function parameter or member 'refclk' not described in 'mrst_lvds_find_best_pll'
 drivers/gpu/drm/gma500/oaktrail_crtc.c:184: warning: Function parameter or member 'best_clock' not described in 'mrst_lvds_find_best_pll'
 drivers/gpu/drm/gma500/oaktrail_crtc.c:215: warning: Function parameter or member 'crtc' not described in 'oaktrail_crtc_dpms'
 drivers/gpu/drm/gma500/oaktrail_crtc.c:215: warning: Function parameter or member 'mode' not described in 'oaktrail_crtc_dpms'
 drivers/gpu/drm/gma500/oaktrail_crtc.c:345: warning: Function parameter or member 'dev' not described in 'oaktrail_panel_fitter_pipe'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-21-lee.jones@linaro.org

drm/gma500/oaktrail_lvds: Demote one documentation header fix another

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/oaktrail_lvds.c:38: warning: Function parameter or member 'dev' not described in 'oaktrail_lvds_set_power'
 drivers/gpu/drm/gma500/oaktrail_lvds.c:38: warning: Function parameter or member 'gma_encoder' not described in 'oaktrail_lvds_set_power'
 drivers/gpu/drm/gma500/oaktrail_lvds.c:38: warning: Function parameter or member 'on' not described in 'oaktrail_lvds_set_power'
 drivers/gpu/drm/gma500/oaktrail_lvds.c:291: warning: Function parameter or member 'mode_dev' not described in 'oaktrail_lvds_init'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: Jesse Barnes <jesse.barnes@intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-20-lee.jones@linaro.org

drm/gma500/oaktrail_lvds_i2c: Remove unused variables 'tmp'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/oaktrail_lvds_i2c.c: In function ‘get_clock’:
 drivers/gpu/drm/gma500/oaktrail_lvds_i2c.c:69:11: warning: variable ‘tmp’ set but not used [-Wunused-but-set-variable]
 drivers/gpu/drm/gma500/oaktrail_lvds_i2c.c: In function ‘get_data’:
 drivers/gpu/drm/gma500/oaktrail_lvds_i2c.c:83:11: warning: variable ‘tmp’ set but not used [-Wunused-but-set-variable]

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Jan Safrata <jan.nikitenko@gmail.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-19-lee.jones@linaro.org

drm/gma500/cdv_intel_dp: Demote one unworthy header fix another

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/cdv_intel_dp.c:315: warning: Function parameter or member 'encoder' not described in 'is_edp'
 drivers/gpu/drm/gma500/cdv_intel_dp.c:315: warning: Excess function parameter 'intel_dp' description in 'is_edp'
 drivers/gpu/drm/gma500/cdv_intel_dp.c:1698: warning: Function parameter or member 'connector' not described in 'cdv_intel_dp_detect'
 drivers/gpu/drm/gma500/cdv_intel_dp.c:1698: warning: Function parameter or member 'force' not described in 'cdv_intel_dp_detect'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Keith Packard <keithp@keithp.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-18-lee.jones@linaro.org
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-17-lee.jones@linaro.org

drm/gma500/cdv_intel_lvds: Demote unworthy headers to standard blocks and fix another

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/cdv_intel_lvds.c:83: warning: Function parameter or member 'dev' not described in 'cdv_intel_lvds_set_backlight'
 drivers/gpu/drm/gma500/cdv_intel_lvds.c:83: warning: Function parameter or member 'level' not described in 'cdv_intel_lvds_set_backlight'
 drivers/gpu/drm/gma500/cdv_intel_lvds.c:107: warning: Function parameter or member 'dev' not described in 'cdv_intel_lvds_set_power'
 drivers/gpu/drm/gma500/cdv_intel_lvds.c:107: warning: Function parameter or member 'encoder' not described in 'cdv_intel_lvds_set_power'
 drivers/gpu/drm/gma500/cdv_intel_lvds.c:107: warning: Function parameter or member 'on' not described in 'cdv_intel_lvds_set_power'
 drivers/gpu/drm/gma500/cdv_intel_lvds.c:298: warning: Function parameter or member 'connector' not described in 'cdv_intel_lvds_get_modes'
 drivers/gpu/drm/gma500/cdv_intel_lvds.c:480: warning: Function parameter or member 'mode_dev' not described in 'cdv_intel_lvds_init'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: Jesse Barnes <jesse.barnes@intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-17-lee.jones@linaro.org

drm/gma500/cdv_intel_display: Demote kernel-doc abuse in 'cdv_intel_panel_fitter_pipe()'s header

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/cdv_intel_display.c:559: warning: Function parameter or member 'dev' not described in 'cdv_intel_panel_fitter_pipe'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-16-lee.jones@linaro.org

drm/gma500/cdv_intel_crt: Demote kernel-doc abuse in 'cdv_intel_crt_detect_hotplug()'s header

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/cdv_intel_crt.c:138: warning: Function parameter or member 'connector' not described in 'cdv_intel_crt_detect_hotplug'
 drivers/gpu/drm/gma500/cdv_intel_crt.c:138: warning: Function parameter or member 'force' not described in 'cdv_intel_crt_detect_hotplug'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-15-lee.jones@linaro.org

drm/drm_dp_helper: Fix spelling issue and add missing description for 'enable_frl_ready_hpd'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/drm_dp_helper.c:965: warning: Function parameter or member 'color_spc' not described in 'drm_dp_downstream_rgb_to_ycbcr_conversion'
 drivers/gpu/drm/drm_dp_helper.c:965: warning: Excess function parameter 'colorspc' description in 'drm_dp_downstream_rgb_to_ycbcr_conversion'
 drivers/gpu/drm/drm_dp_helper.c:2678: warning: Function parameter or member 'enable_frl_ready_hpd' not described in 'drm_dp_pcon_frl_prepare'

Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-14-lee.jones@linaro.org

drm/gma500/psb_irq: Demote kernel-doc abuse

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/psb_irq.c:134: warning: Function parameter or member 'dev' not described in 'mid_pipe_event_handler'
 drivers/gpu/drm/gma500/psb_irq.c:134: warning: Function parameter or member 'pipe' not described in 'mid_pipe_event_handler'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-13-lee.jones@linaro.org

drm/gma500/psb_intel_sdvo: Demote kernel-doc abuses

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/psb_intel_sdvo.c:230: warning: Function parameter or member 'psb_intel_sdvo' not described in 'psb_intel_sdvo_write_sdvox'
 drivers/gpu/drm/gma500/psb_intel_sdvo.c:230: warning: Function parameter or member 'val' not described in 'psb_intel_sdvo_write_sdvox'
 drivers/gpu/drm/gma500/psb_intel_sdvo.c:598: warning: Function parameter or member 'psb_intel_sdvo' not described in 'psb_intel_sdvo_get_trained_inputs'
 drivers/gpu/drm/gma500/psb_intel_sdvo.c:598: warning: Function parameter or member 'input_1' not described in 'psb_intel_sdvo_get_trained_inputs'
 drivers/gpu/drm/gma500/psb_intel_sdvo.c:598: warning: Function parameter or member 'input_2' not described in 'psb_intel_sdvo_get_trained_inputs'
 drivers/gpu/drm/gma500/psb_intel_sdvo.c:1831: warning: Function parameter or member 'dev_priv' not described in 'psb_intel_sdvo_select_ddc_bus'
 drivers/gpu/drm/gma500/psb_intel_sdvo.c:1831: warning: Function parameter or member 'sdvo' not described in 'psb_intel_sdvo_select_ddc_bus'
 drivers/gpu/drm/gma500/psb_intel_sdvo.c:1831: warning: Function parameter or member 'reg' not described in 'psb_intel_sdvo_select_ddc_bus'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Jesse Barnes <jesse.barnes@intel.com>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-12-lee.jones@linaro.org

drm/gma500/mmu: Actually return an error if one occurs

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/mmu.c: In function ‘psb_mmu_insert_pfn_sequence’:
 drivers/gpu/drm/gma500/mmu.c:651:6: warning: variable ‘ret’ set but not used [-Wunused-but-set-variable]

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181601.3432599-7-lee.jones@linaro.org

drm/gma500/power: Remove excess param description 'state'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/power.c:190: warning: Excess function parameter 'state' description in 'gma_power_suspend'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Benjamin Defnet <benjamin.r.defnet@intel.com>
Cc: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
Cc: Alan Cox <alan@linux.intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-39-lee.jones@linaro.org

drm/gma500/mmu: Remove unused function 'psb_get_default_pd_addr'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/mmu.c:420:10: warning: no previous prototype for ‘psb_get_default_pd_addr’ [-Wmissing-prototypes]

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-35-lee.jones@linaro.org

drm/gma500/gem: Include our own header containing prototypes

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/gem.c:59:5: warning: no previous prototype for ‘psb_gem_create’ [-Wmissing-prototypes]

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-33-lee.jones@linaro.org

drm/gma500/gem: Do not rely on consumers to include drm/drm_device.h before gem.h

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/gem.h:13:57: warning: ‘struct drm_device’ declared inside parameter list will not be visible outside of this definition or declaration

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-32-lee.jones@linaro.org

drm/gma500/psb_intel_modes: Provide missing descriptions for function param 'adapter'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/psb_intel_modes.c:17: warning: Function parameter or member 'adapter' not described in 'psb_intel_ddc_probe'
 drivers/gpu/drm/gma500/psb_intel_modes.c:51: warning: Function parameter or member 'adapter' not described in 'psb_intel_ddc_get_modes'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Jesse Barnes <jesse.barnes@intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-31-lee.jones@linaro.org

drm/gma500/psb_intel_lvds: Supply description for 'mode_dev'

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/psb_intel_lvds.c:635: warning: Function parameter or member 'mode_dev' not described in 'psb_intel_lvds_init'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: Jesse Barnes <jesse.barnes@intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-26-lee.jones@linaro.org

drm/gma500/psb_intel_display: Demote kernel-doc formatting abuse

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/psb_intel_display.c:79: warning: Function parameter or member 'dev' not described in 'psb_intel_panel_fitter_pipe'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-25-lee.jones@linaro.org

drm/gma500/gma_device: Include our own header containing prototypes

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/gma_device.c:10:6: warning: no previous prototype for ‘gma_get_core_freq’ [-Wmissing-prototypes]

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-23-lee.jones@linaro.org

drm/gma500/gma_display: Demote kernel-doc abuses to standard comment blocks

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/gma_display.c:27: warning: Function parameter or member 'crtc' not described in 'gma_pipe_has_type'
 drivers/gpu/drm/gma500/gma_display.c:27: warning: Function parameter or member 'type' not described in 'gma_pipe_has_type'
 drivers/gpu/drm/gma500/gma_display.c:190: warning: Function parameter or member 'crtc' not described in 'gma_crtc_dpms'
 drivers/gpu/drm/gma500/gma_display.c:190: warning: Function parameter or member 'mode' not described in 'gma_crtc_dpms'
 drivers/gpu/drm/gma500/gma_display.c:573: warning: Function parameter or member 'crtc' not described in 'gma_crtc_save'
 drivers/gpu/drm/gma500/gma_display.c:616: warning: Function parameter or member 'crtc' not described in 'gma_crtc_restore'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-22-lee.jones@linaro.org

drm/gma500/power: Fix a bunch of *dev documentation issues

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/power.c:101: warning: Function parameter or member 'pdev' not described in 'gma_resume_display'
 drivers/gpu/drm/gma500/power.c:155: warning: Function parameter or member 'pdev' not described in 'gma_resume_pci'
 drivers/gpu/drm/gma500/power.c:155: warning: Excess function parameter 'dev' description in 'gma_resume_pci'
 drivers/gpu/drm/gma500/power.c:189: warning: Function parameter or member '_dev' not described in 'gma_power_suspend'
 drivers/gpu/drm/gma500/power.c:189: warning: Excess function parameter 'pdev' description in 'gma_power_suspend'
 drivers/gpu/drm/gma500/power.c:189: warning: Excess function parameter 'state' description in 'gma_power_suspend'
 drivers/gpu/drm/gma500/power.c:216: warning: Function parameter or member '_dev' not described in 'gma_power_resume'
 drivers/gpu/drm/gma500/power.c:216: warning: Excess function parameter 'pdev' description in 'gma_power_resume'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Benjamin Defnet <benjamin.r.defnet@intel.com>
Cc: Rajesh Poornachandran <rajesh.poornachandran@intel.com>
Cc: Alan Cox <alan@linux.intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-20-lee.jones@linaro.org

drm/gma500/mmu: Make local function 'psb_mmu_pt_alloc_map_lock()' static

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/mmu.c:316:20: warning: no previous prototype for ‘psb_mmu_pt_alloc_map_lock’ [-Wmissing-prototypes]

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-19-lee.jones@linaro.org

drm/gma500/intel_i2c: Remove superflouous parameter description and rename another

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/intel_i2c.c:108: warning: Excess function parameter 'output' description in 'psb_intel_i2c_create'
 drivers/gpu/drm/gma500/intel_i2c.c:153: warning: Function parameter or member 'chan' not described in 'psb_intel_i2c_destroy'
 drivers/gpu/drm/gma500/intel_i2c.c:153: warning: Excess function parameter 'output' description in 'psb_intel_i2c_destroy'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-18-lee.jones@linaro.org

drm/gma500/intel_bios: Demote non-conformant kernel-doc header

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/intel_bios.c:581: warning: Function parameter or member 'dev' not described in 'psb_intel_destroy_bios'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-17-lee.jones@linaro.org

drm/gma500/gem: Add and rename some function parameter descriptions

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/gem.c:57:5: warning: no previous prototype for ‘psb_gem_create’ [-Wmissing-prototypes]
 drivers/gpu/drm/gma500/gem.c:59: warning: Function parameter or member 'stolen' not described in 'psb_gem_create'
 drivers/gpu/drm/gma500/gem.c:59: warning: Function parameter or member 'align' not described in 'psb_gem_create'
 drivers/gpu/drm/gma500/gem.c:110: warning: Function parameter or member 'file' not described in 'psb_gem_dumb_create'
 drivers/gpu/drm/gma500/gem.c:110: warning: Excess function parameter 'drm_file' description in 'psb_gem_dumb_create'
 drivers/gpu/drm/gma500/gem.c:135: warning: Excess function parameter 'vma' description in 'psb_gem_fault'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-16-lee.jones@linaro.org

drm/gma500/framebuffer: Fix some possible doc-rot issues

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/gma500/framebuffer.c:171: warning: Function parameter or member 'obj' not described in 'psb_framebuffer_init'
 drivers/gpu/drm/gma500/framebuffer.c:171: warning: Excess function parameter 'gt' description in 'psb_framebuffer_init'
 drivers/gpu/drm/gma500/framebuffer.c:212: warning: Function parameter or member 'obj' not described in 'psb_framebuffer_create'
 drivers/gpu/drm/gma500/framebuffer.c:212: warning: Excess function parameter 'gt' description in 'psb_framebuffer_create'
 drivers/gpu/drm/gma500/framebuffer.c:262: warning: Function parameter or member 'fb_helper' not described in 'psbfb_create'
 drivers/gpu/drm/gma500/framebuffer.c:262: warning: Excess function parameter 'fbdev' description in 'psbfb_create'

Cc: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-15-lee.jones@linaro.org

drm/mga/mga_ioc32: Fix some formatting issues in 'mga_compat_ioctl's header

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/mga/mga_ioc32.c:171: warning: Function parameter or member 'filp' not described in 'mga_compat_ioctl'
 drivers/gpu/drm/mga/mga_ioc32.c:171: warning: Function parameter or member 'cmd' not described in 'mga_compat_ioctl'
 drivers/gpu/drm/mga/mga_ioc32.c:171: warning: Function parameter or member 'arg' not described in 'mga_compat_ioctl'

Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Lee Jones <lee.jones@linaro.org>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-7-lee.jones@linaro.org

drm/r128/r128_ioc32: Fix formatting issues in 'r128_compat_ioctl()'s header

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/r128/r128_ioc32.c:182: warning: Function parameter or member 'filp' not described in 'r128_compat_ioctl'
 drivers/gpu/drm/r128/r128_ioc32.c:182: warning: Function parameter or member 'cmd' not described in 'r128_compat_ioctl'
 drivers/gpu/drm/r128/r128_ioc32.c:182: warning: Function parameter or member 'arg' not described in 'r128_compat_ioctl'

Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Lee Jones <lee.jones@linaro.org>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-6-lee.jones@linaro.org

drm/mga/mga_ioc32: Document headers do not make good kernel-doc candidates

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/mga/mga_ioc32.c:2: warning: Cannot understand  * file mga_ioc32.c

Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-3-lee.jones@linaro.org

drm/r128/r128_ioc32: Document headers do not make good kernel-doc candidates

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/r128/r128_ioc32.c:2: warning: Cannot understand  * file r128_ioc32.c

Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115181313.3431493-2-lee.jones@linaro.org

dma-buf: Add debug option

We have too many people abusing the struct page they can get at but
really shouldn't in importers. Aside from that the backing page might
simply not exist (for dynamic p2p mappings) looking at it and using it
e.g. for mmap can also wreak the page handling of the exporter
completely. Importers really must go through the proper interface like
dma_buf_mmap for everything.

I'm semi-tempted to enforce this for dynamic importers since those
really have no excuse at all to break the rules.

Unfortuantely we can't store the right pointers somewhere safe to make
sure we oops on something recognizable, so best is to just wrangle
them a bit by flipping all the bits. At least on x86 kernel addresses
have all their high bits sets and the struct page array is fairly low
in the kernel mapping, so flipping all the bits gives us a very high
pointer in userspace and hence excellent chances for an invalid
dereference.

v2: Add a note to the @map_dma_buf hook that exporters shouldn't do
fancy caching tricks, which would blow up with this address scrambling
trick here (Chris)

Enable by default when CONFIG_DMA_API_DEBUG is enabled.

v3: Only one copy of the mangle/unmangle code (Christian)

v4: #ifdef, not #if (0day)

v5: sg_table can also be an ERR_PTR (Chris, Christian)

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Stevens <stevensd@chromium.org>
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Link: https://patchwork.freedesktop.org/patch/msgid/20210115164739.3958206-1-daniel.vetter@ffwll.ch

drm/ttm: WARN_ON non-empty lru when disabling a resource manager

ttm_resource_manager->use_type is only used for runtime changes by
vmwgfx. I think ideally we'd push this functionality into drivers -
ttm itself does not provide any locking to guarantee this is safe, so
the only way this can work at runtime is if the driver does provide
additional guarantees. vwmgfx does that through the
vmw_private->reservation_sem. Therefore supporting this feature in
shared code feels a bit misplaced.

As a first step add a WARN_ON to make sure the resource manager is
empty. This is just to make sure I actually understand correctly what
vmwgfx is doing, and to make sure an eventual subsequent refactor
doesn't break anything.

This check should also be useful for other drivers, to make sure they
haven't leaked anything.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Cc: Christian Koenig <christian.koenig@amd.com>
Cc: Huang Rui <ray.huang@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201211162942.3399050-3-daniel.vetter@ffwll.ch

drm/vmwgfx: Always evict vram _before_ disabling it

Other way round is a bit inconsistent (but not buggy in any kind).
This is prep work so that ttm_resource_manager_set_used can assert
that the resource manager is empty.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Cc: VMware Graphics <linux-graphics-maintainer@vmware.com>
Cc: Roland Scheidegger <sroland@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201211162942.3399050-2-daniel.vetter@ffwll.ch

drm/vmwgfx: Drop svga_lock

This isn't actually protecting anything becuase:
- when running, ttm_resource_manager->use_type is protected through
  vmw_private->reservation_semaphore against concurrent execbuf or
  well anything else that might evict or reserve buffers
- during suspend/resume there's nothing else running, hence
  vmw_pm_freeze and vmw_pm_restore do not need to take the same lock.
- this also holds for the SVGA_REG_ENABLE register write

Hence it is safe to just remove that spinlock.

Signed-off-by: Daniel Vetter <daniel.vetter@intel.com>
Reviewed-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Cc: VMware Graphics <linux-graphics-maintainer@vmware.com>
Cc: Roland Scheidegger <sroland@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201211162942.3399050-1-daniel.vetter@ffwll.ch

drm/i915/gt: Add arbitration check before semaphore wait

Similar to commit 49b20dbf7497 ("drm/i915/gt: Perform an arbitration
check before busywaiting"), also add a check prior to the busywait
on gen8+, as we have now seen (because we added a selftest to add fault
injection into the engine resets) the same engine reset failure leading
to an indefinite wait on the ring-stop semaphore. So not a Tigerlake
specific bug after all, though it still seems odd behaviour for the
busywait as we do get the arbitration point elsewhere on a miss.

Testcase: igt_reset_fail_engine
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210117110418.3361-1-chris@chris-wilson.co.uk

drm/i915/gt: Clear CACHE_MODE prior to clearing residuals

Since we do a bare context switch with no restore, the clear residual
kernel runs on dirty state, and we must be careful to avoid executing
with bad state from context registers inherited from a malicious client.

Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2955
Fixes: 008ead6ef8f5 ("drm/i915/gt: Restore clear-residual mitigations for Ivybridge, Baytrail")
Fixes: 09aa9e45863e ("drm/i915/gt: Restore clear-residual mitigations for Ivybridge, Baytrail")
Testcase: igt/gem_ctx_isolation # ivb,vlv
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Reviewed-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210117093015.29143-1-chris@chris-wilson.co.uk

drm/panel: panel-simple: add bus-format and connector-type to Innolux n116bge

The Innolux n116bge panel has an eDP connector and 3*6 bits bus format.

Signed-off-by: Heiko Stuebner <heiko.stuebner@theobroma-systems.com>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210109130951.3448435-1-heiko@sntech.de

vgaarb: Remove unneeded semicolons

Remove superfluous semicolons after function definitions.

Signed-off-by: Yue Zou <zouyue3@huawei.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210118010356.214491-1-zouyue3@huawei.com

drm: Set vm_ops to GEM object's values during mmap

The GEM mmap code relies on the GEM object's mmap callback to set the
VMA's vm_ops field. This is easily forgotten and already led to a memory
leak in the CMA helpers. Instead set the vm_ops field in the DRM core
code to the GEM object's value. Drivers with different needs can override
this in their mmap callback.

v2:
* support (vm_ops == NULL) if mmap is given; required by VRAM
  helpers

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: f5ca8eb6f9bd ("drm/cma-helper: Implement mmap as GEM CMA object functions")
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reported-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Tested-by: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Eric Anholt <eric@anholt.net>
Cc: dri-devel@lists.freedesktop.org
Link: https://patchwork.freedesktop.org/patch/msgid/20210115093038.10345-1-tzimmermann@suse.de

drm/i915/gt: Convert stats.active to plain unsigned int

As context-in/out is now always serialised, we do not have to worry
about concurrent enabling/disable of the busy-stats and can reduce the
atomic_t active to a plain unsigned int, and the seqlock to a seqcount.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115142331.24458-3-chris@chris-wilson.co.uk

drm/i915/gt: Extract busy-stats for ring-scheduler

Lift the busy-stats context-in/out implementation out of intel_lrc, so
that we can reuse it for other scheduler implementations.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115142331.24458-2-chris@chris-wilson.co.uk

drm/i915/gt: Drop atomic for engine->fw_active tracking

Since schedule-in/out is now entirely serialised by the tasklet bitlock,
we do not need to worry about concurrent in/out operations and so reduce
the atomic operations to plain instructions.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115142331.24458-1-chris@chris-wilson.co.uk

drm/vkms: Fix missing kmalloc allocation failure check

Currently the kmalloc allocation for config is not being null
checked and could potentially lead to a null pointer dereference.
Fix this by adding the missing null check.

Addresses-Coverity: ("Dereference null return value")
Fixes: 2df7af93fdad ("drm/vkms: Add vkms_config type")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Sumera Priyadarsini <sylphrenadin@gmail.com>
Signed-off-by: Melissa Wen <melissa.srw@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115130911.71073-1-colin.king@canonical.com

drm/amdgpu/display: these forward declarations are not used

Remove them.

Signed-off-by: ZhiJie.Zhang <zhangzhijie@loongson.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix the coding style issue of integrated_info

Follow the coding style of size of a struct, changing to:

p = kmalloc(sizeof(*p), ...);

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: update driver if version for navy_flounder

It's in accordance with pmfw 65.22.0 for navy_flounder.

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix the system memory page fault because of copy overflow

The buffer is allocated with the size of pointer and copy with the size of
data structure. Then trigger the system memory page fault. Use the
orignal data structure to get the object size.

Fixes: 3a00c04212d1 ("drm/amd/display/dc/core/dc_link: Move some local data from the stack to the heap")
Signed-off-by: Huang Rui <ray.huang@amd.com>
Cc: Lee Jones <lee.jones@linaro.org>
Reviewed-by: Jinzhou.Su <Jinzhou.Su@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: updated fw attestation interface

updated interface to be compatible with latest FW

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: toggle on DF Cstate after finishing xgmi injection

Fixes: 5c23e9e05e42 ("drm/amdgpu: Update RAS XGMI error inject sequence")
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/i915: add back static declaration

warning: symbol '__i915_gem_object_create_stolen' was not declared. Should it be static?
warning: symbol '_i915_gem_object_stolen_init' was not declared. Should it be static?

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210115101329.880667-1-matthew.auld@intel.com

drm/amdgpu: Remove unused variable

Remove unused space_needed variable.

Fixes: 453f617a30aa ("drm/amdgpu: Resize BAR0 to the maximum available size, even if it doesn't cover VRAM")
Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/413807/

drm/vc4: Initialize vc4_drm_driver with CMA helper defaults

The function vc4_prime_import_sg_table() is an otherwise empty wrapper
around CMA's drm_gem_cma_prime_import_sg_table(). Removing it in favor
of the latter allows to initialize vc4_drm_driver with CMA's initializer
macro.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114084949.29014-1-tzimmermann@suse.de

drm/hisilicon/hibmc: Remove hibmc_ttm.c

The file is not in use. It got re-added by a rebased patch. Removing
it.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: 4d4dad21cc7b ("drm/hibmc: Remove references to struct drm_device.pdev")
Reviewed-by: Tian Tao <tiantao6@hisilicon.com>
Reported-by: Tian Tao <tiantao6@hisilicon.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Xinliang Liu <xinliang.liu@linaro.org>
Cc: Tian Tao <tiantao6@hisilicon.com>
Cc: John Stultz <john.stultz@linaro.org>
Cc: Xinwei Kong <kong.kongxinwei@hisilicon.com>
Cc: Chen Feng <puck.chen@hisilicon.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Gong junjie <gongjunjie2@huawei.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210113113107.12005-1-tzimmermann@suse.de

drm/i915/region: convert object_create into object_init

Give more flexibility to the caller, if they already have an allocated
object, in case they wish to apply some transformation to the object
prior to handing it over to the region specific initialisation step,
like in gem_create_ext where we would like to first apply the extensions
to the object.

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114182402.840247-3-matthew.auld@intel.com

drm/i915/gem: sanity check object size in gem_create

Depending on the regions min_page_size we might need to adjust the
object size, ensure this matches our expectations.

Suggested-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114182402.840247-2-matthew.auld@intel.com

drm/i915/gem: split gem_create into own file

In preparation for gem_create_ext break out the gem_create uAPI, so that
we don't clutter i915_gem.c once we start adding various extensions

Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114182402.840247-1-matthew.auld@intel.com

drm/i915: Reduce test_and_set_bit to set_bit in i915_request_submit()

Avoid the full blown memory barrier of test_and_set_bit() by noting the
completed request and removing it from the lists.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114135612.13210-5-chris@chris-wilson.co.uk

drm/i915/gem: Reduce ctx->engines_mutex for get_engines()

Take a snapshot of the ctx->engines, so we can avoid taking the
ctx->engines_mutex for a mere read in get_engines().

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114135612.13210-4-chris@chris-wilson.co.uk

drm/i915/gem: Reduce ctx->engine_mutex for reading the clone source

When cloning the engines from the source context, we need to ensure that
the engines are not freed as we copy them, and that the flags we clone
from the source correspond with the engines we copy across. To do this
we need only take a reference to the src->engines, rather than hold the
src->engine_mutex, so long as we verify that nothing changed under the
read.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114135612.13210-3-chris@chris-wilson.co.uk

drm/i915: Drop i915_request.lock serialisation around await_start

Originally, we used the signal->lock as a means of following the
previous link in its timeline and peeking at the previous fence.
However, we have replaced the explicit serialisation with a series of
very careful probes that anticipate the links being deleted and the
fences recycled before we are able to acquire a strong reference to it.
We do not need the signal->lock crutch anymore, nor want the contention.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114135612.13210-2-chris@chris-wilson.co.uk

drm/i915: Mark up protected uses of 'i915_request_completed'

When we know that we are inside the timeline mutex, or inside the
submission flow (under active.lock or the holder's rcu lock), we know
that the rq->hwsp is stable and we can use the simpler direct version.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Andi Shyti <andi.shyti@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114135612.13210-1-chris@chris-wilson.co.uk