This moves ppgtt root hook out of scan and shadow function,
as it's only required at dispatch time. Also make sure this
checks against shadow mm to be ready, otherwise bail to fail
earlier.
Reviewed-by: Xiong Zhang <xiong.y.zhang@intel.com>
Cc: Xiong Zhang <xiong.y.zhang@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
There is a really hairy resolution involving amdgpu fixes, that I'd rather confirm here.
Also some misc fixes are landed by me, but the pr has them as well.
Signed-off-by: Dave Airlie <airlied@redhat.com>
We don't want to busywait on the GPU if we have other work to do. If we
give non-busywaiting workloads higher (initial) priority than workloads
that require a busywait, we will prioritise work that is ready to run
immediately. We then also have to be careful that we don't give earlier
semaphores an accidental boost because later work doesn't wait on other
rings, hence we keep a history of semaphore usage of the dependency chain.
v2: Stop rolling the bits into a chain and just use a flag in case this
request or any of our dependencies use a semaphore. The rolling around
was contagious as Tvrtko was heard to fall off his chair.
Testcase: igt/gem_exec_schedule/semaphore
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/20190301170901.8340-4-chris@chris-wilson.co.uk
Having introduced per-context seqno, we now have a means to identity
progress across the system without feel of rollback as befell the
global_seqno. That is we can program a MI_SEMAPHORE_WAIT operation in
advance of submission safe in the knowledge that our target seqno and
address is stable.
However, since we are telling the GPU to busy-spin on the target address
until it matches the signaling seqno, we only want to do so when we are
sure that busy-spin will be completed quickly. To achieve this we only
submit the request to HW once the signaler is itself executing (modulo
preemption causing us to wait longer), and we only do so for default and
above priority requests (so that idle priority tasks never themselves
hog the GPU waiting for others).
As might be reasonably expected, HW semaphores excel in inter-engine
synchronisation microbenchmarks (where the 3x reduced latency / increased
throughput more than offset the power cost of spinning on a second ring)
and have significant improvement (can be up to ~10%, most see no change)
for single clients that utilize multiple engines (typically media players
and transcoders), without regressing multiple clients that can saturate
the system or changing the power envelope dramatically.
v3: Drop the older NEQ branch, now we pin the signaler's HWSP anyway.
v4: Tell the world and include it as part of scheduler caps.
Testcase: igt/gem_exec_whisper
Testcase: igt/benchmarks/gem_wsim
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/20190301170901.8340-3-chris@chris-wilson.co.uk
In preparation for enabling HW semaphores, we need to keep in flight
timeline HWSP alive until its use across entire system has completed,
as any other timeline active on the GPU may still refer back to the
already retired timeline. We both have to delay recycling available
cachelines and unpinning old HWSP until the next idle point.
An easy option would be to simply keep all used HWSP until the system as
a whole was idle, i.e. we could release them all at once on parking.
However, on a busy system, we may never see a global idle point,
essentially meaning the resource will be leaked until we are forced to
do a GC pass. We already employ a fine-grained idle detection mechanism
for vma, which we can reuse here so that each cacheline can be freed
immediately after the last request using it is retired.
v3: Keep track of the activity of each cacheline.
v4: cacheline_free() on canceling the seqno tracking
v5: Finally with a testcase to exercise wraparound
v6: Pack cacheline into empty bits of page-aligned vaddr
v7: Use i915_utils to hide the pointer casting around bit manipulation
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190301170901.8340-2-chris@chris-wilson.co.uk
On unwinding the active request we give it a small (limited to internal
priority levels) boost to prevent it from being gazumped a second time.
However, this means that it can be promoted to above the request that
triggered the preemption request, causing a preempt-to-idle cycle for no
change. We can avoid this if we take the boost into account when
checking if the preemption request is valid.
v2: After preemption the active request will be after the preemptee if
they end up with equal priority.
v3: Tvrtko pointed out that this, the existing logic, makes
I915_PRIORITY_WAIT non-preemptible. Document this interesting quirk!
v4: Prove Tvrtko was right about WAIT being non-preemptible and test it.
v5: Except not all priorities were made equal, and the WAIT not preempting
is only if we start off as !NEWCLIENT.
v6: More commentary after coming to an understanding about what I had
forgotten to say.
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/20190301170901.8340-1-chris@chris-wilson.co.uk
A simple mutex used for guarding the flow of requests in and out of the
timeline. In the short-term, it will be used only to guard the addition
of requests into the timeline, taken on alloc and released on commit so
that only one caller can construct a request into the timeline
(important as the seqno and ring pointers must be serialised). This will
be used by observers to ensure that the seqno/hwsp is stable. Later,
when we have reduced retiring to only operate on a single timeline at a
time, we can then use the mutex as the sole guard required for retiring.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190301110547.14758-2-chris@chris-wilson.co.uk
VI planes support coarse scaling which helps to overcome VI scaler
limitations. While exact working of coarse scaling isn't known, it seems
that it just skips programmed amount of rows and columns. This is
especially useful for downscaling very big planes (4K down to 1080p).
Horizontal coarse scaling is currently used to fit one line to VI scaler
buffer.
Vertical coarse scaling is used to assure that VI scaler is actually
capable of processing framebuffer in one frame time.
Signed-off-by: Jernej Skrabec <jernej.skrabec@siol.net>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228200329.11128-4-jernej.skrabec@siol.net
As vGPU shadow ctx is loaded with guest context state, arbitrarily
submitting request in error workload dispatch path would cause trouble.
So don't try to submit in error path now like in previous code.
This is to fix VM failure when GPU hang happens.
Fixes: f0e9943725 ("drm/i915/gvt: Fix workload request allocation before request add")
Reviewed-by: Xiong Zhang <xiong.y.zhang@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
There is one corner case that workload_thread may pick and dispatch one
workload of vgpu after it's already deactivated. Below is the scenario:
1. deactive_vgpu got the vgpu_lock, it found pending workload was
submitted, then it released the vgpu_lock and wait for vgpu idle.
2. before deactive_vgpu got the vgpu_lock back, workload_thread might pick
one new valid workload, then it was blocked by the vgpu_lock.
3. deactive_vgpu got the vgpu_lock again, finished the last processes of
deactivating, then release the vgpu_lock.
4. workload_thread got the vgpu_lock, then it will try to dispatch the
fetched workload. It's not expected one workload of deactivated vgpu is
dispatched.
The solution is to add condition check of the vgpu's active flag and stop
to schedule when it's inactive.
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Weinan Li <weinan.z.li@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
WAIT is occasionally suppressed by virtue of preempted requests being
promoted to NEWCLIENT if they have not all ready received that boost.
Make this consistent for all WAIT boosts that they are not allowed to
preempt executing contexts and are merely granted the right to be at the
front of the queue for the next execution slot. This is in keeping with
the desire that the WAIT boost be a minor tweak that does not give
excessive promotion to its user and open ourselves to trivial abuse.
The problem with the inconsistent WAIT preemption becomes more apparent
as the preemption is propagated across the engines, where one engine may
preempt and the other not, and we be relying on the exact execution
order being consistent across engines (e.g. using HW semaphores to
coordinate parallel execution).
v2: Also protect GuC submission from false preemption loops.
v3: Build bug safeguards and better debug messages for st.
v4: Do the priority bumping in unsubmit (i.e. on preemption/reset
unwind), applying it earlier during submit causes out-of-order execution
combined with execute fences.
v5: Call sw_fence_fini for our dummy request (Matthew)
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>
Cc: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228220639.3173-1-chris@chris-wilson.co.uk
ASoC: More changes for v5.1
Another batch of changes for ASoC, no big core changes - it's mainly
small fixes and improvements for individual drivers.
- A big refresh and cleanup of the Samsung drivers, fixing a number of
issues which allow the driver to be used with a wider range of
userspaces.
- Fixes for the Intel drivers to make them more standard so less likely
to get bitten by core issues.
- New driver for Cirrus Logic CS35L26.
Daniel writes:
mei-hdcp driver
mei driver for the me hdcp client, for use by drm/i915.
Including the following prep work:
- whitelist hdcp client in mei bus
- merge to include char-misc-next
- drm/i915 side of the mei_hdcp/i915 component interface
- component prep work (including one patch touching i915&snd-hda)
* tag 'topic/mei-hdcp-2019-02-26' of git://anongit.freedesktop.org/drm/drm-intel: (23 commits)
misc/mei/hdcp: Component framework for I915 Interface
misc/mei/hdcp: Closing wired HDCP2.2 Tx Session
misc/mei/hdcp: Enabling the HDCP authentication
misc/mei/hdcp: Verify M_prime
misc/mei/hdcp: Repeater topology verification and ack
misc/mei/hdcp: Prepare Session Key
misc/mei/hdcp: Verify L_prime
misc/mei/hdcp: Initiate Locality check
misc/mei/hdcp: Store the HDCP Pairing info
misc/mei/hdcp: Verify H_prime
misc/mei/hdcp: Verify Receiver Cert and prepare km
misc/mei/hdcp: Initiate Wired HDCP2.2 Tx Session
misc/mei/hdcp: Define ME FW interface for HDCP2.2
misc/mei/hdcp: Client driver for HDCP application
mei: bus: whitelist hdcp client
drm/audio: declaration of struct device
drm: helper functions for hdcp2 seq_num to from u32
drm/i915: MEI interface definition
drm/i915: header for i915 - MEI_HDCP interface
drm/i915: enum port definition is moved into i915_drm.h
...
We currently use a worker queued from an rcu callback to determine when
a how grace period has elapsed while we remained idle. We use this idle
delay to infer that we will be idle for a while and this is a suitable
point at which we can trim our global memory caches.
Since we wrote that, this mechanism now exists as rcu_work, and having
converted the idle shrinkers over to using that, we can remove our own
variant.
v2: Say goodbye to gt.epoch as well.
v3: Remove the misplaced and redundant comment before parking globals
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/20190228102035.5857-3-chris@chris-wilson.co.uk
As kmem_caches share the same properties (size, allocation/free behaviour)
for all potential devices, we can use global caches. While this
potential has worse fragmentation behaviour (one can argue that
different devices would have different activity lifetimes, but you can
also argue that activity is temporal across the system) it is the
default behaviour of the system at large to amalgamate matching caches.
The benefit for us is much reduced pointer dancing along the frequent
allocation paths.
v2: Defer shrinking until after a global grace period for futureproofing
multiple consumers of the slab caches, similar to the current strategy
for avoiding shrinking too early.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190228102035.5857-1-chris@chris-wilson.co.uk
The prepare_fb call always happens on new_plane_state.
The drm_atomic_helper_cleanup_planes checks to see if
plane state pointer has changed when deciding to call cleanup_fb on
either the new_plane_state or the old_plane_state.
For a non-async atomic commit the state pointer is swapped, so this
helper calls prepare_fb on the new_plane_state and cleanup_fb on the
old_plane_state. This makes sense, since we want to prepare the
framebuffer we are going to use and cleanup the the framebuffer we are
no longer using.
For the async atomic update helpers this differs. The async atomic
update helpers perform in-place updates on the existing state. They call
drm_atomic_helper_cleanup_planes but the state pointer is not swapped.
This means that prepare_fb is called on the new_plane_state and
cleanup_fb is called on the new_plane_state (not the old).
In the case where old_plane_state->fb == new_plane_state->fb then
there should be no behavioral difference between an async update
and a non-async commit. But there are issues that arise when
old_plane_state->fb != new_plane_state->fb.
The first is that the new_plane_state->fb is immediately cleaned up
after it has been prepared, so we're using a fb that we shouldn't
be.
The second occurs during a sequence of async atomic updates and
non-async regular atomic commits. Suppose there are two framebuffers
being interleaved in a double-buffering scenario, fb1 and fb2:
- Async update, oldfb = NULL, newfb = fb1, prepare fb1, cleanup fb1
- Async update, oldfb = fb1, newfb = fb2, prepare fb2, cleanup fb2
- Non-async commit, oldfb = fb2, newfb = fb1, prepare fb1, cleanup fb2
We call cleanup_fb on fb2 twice in this example scenario, and any
further use will result in use-after-free.
The simple fix to this problem is to block framebuffer changes
in the drm_atomic_helper_async_check function for now.
v2: Move check by itself, add a FIXME (Daniel)
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Cc: <stable@vger.kernel.org> # v4.14+
Fixes: fef9df8b59 ("drm/atomic: initial support for asynchronous plane update")
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Vetter <daniel@ffwll.ch>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Link: https://patchwork.freedesktop.org/patch/275364/
Signed-off-by: Dave Airlie <airlied@redhat.com>
Clang warns when an expression that equals zero is used as a null
pointer constant (in lieu of NULL):
drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm.c:4435:3:
warning: expression which evaluates to zero treated as a null pointer
constant of type 'const enum color_transfer_func *'
[-Wnon-literal-null-conversion]
TRANSFER_FUNC_UNKNOWN,
^~~~~~~~~~~~~~~~~~~~~
1 warning generated.
This warning is caused by commit bb47de7366 ("drm/amdgpu: Set FreeSync
state using drm VRR properties") and it could be solved by using NULL
instead of TRANSFER_FUNC_UNKNOWN or casting TRANSFER_FUNC_UNKNOWN as a
pointer. However, after looking into it, there doesn't appear to be a
good reason to pass app_tf by reference as it is never mutated along the
way. This is the only code path in which app_tf is used:
mod_freesync_build_vrr_infopacket ->
build_vrr_infopacket_v2 ->
build_vrr_infopacket_fs2_data
Neither mod_freesync_build_vrr_infopacket or build_vrr_infopacket_v2
modify app_tf's value and build_vrr_infopacket_fs2_data expects just
the value so we can avoid dereferencing anything by just passing in
app_tf's value to mod_freesync_build_vrr_infopacket and
build_vrr_infopacket_v2.
There is no functional change because build_vrr_infopacket_fs2_data
doesn't do anything if TRANSFER_FUNC_UNKNOWN is passed to it, the same
as not calling build_vrr_infopacket_fs2_data at all like before this
change when NULL was used for app_tf.
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
For UCLK_FMAX OD feature, SMU overwrites the highest UCLK DPM level freq.
Therefore it can only take values that are greater than the second highest
DPM level freq.
Signed-off-by: Evan Quan <evan.quan@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch uses REG32_PCIE wrapper instead of writting pci_index2 and reading
pci_data2 for psp. This sequence should be protected by pcie_idx_lock.
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>
This patch uses REG32_PCIE wrapper instead of writting pci_index2 and reading
pci_data2 for powerplay. This sequence should be protected by pcie_idx_lock.
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>
