Ideally, we want to automagically have the GPU respond to the
instantaneous load by reclocking itself. However, reclocking occurs
relatively slowly, and to the client waiting for a result from the GPU,
too late. To compensate and reduce the client latency, we allow the
first wait from a client to boost the GPU clocks to maximum. This
overcomes the lag in autoreclocking, at the expense of forcing the GPU
clocks too high. So to offset the excessive power usage, we currently
allow a client to only boost the clocks once before we detect the GPU
is idle again. This works reasonably for say the first frame in a
benchmark, but for many more synchronous workloads (like OpenCL) we find
the GPU clocks remain too low. By noting a wait which would idle the GPU
(i.e. we just waited upon the last known request), we can give that
client the idle boost credit (for their next wait) without the 100ms
delay required for us to detect the GPU idle state. The intention is to
boost clients that are stalling in the process of feeding the GPU more
work (and who in doing so let the GPU idle), without granting boost
credits to clients that are throttling themselves (such as compositors).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: "Zou, Nanhai" <nanhai.zou@intel.com>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1467616119-4093-4-git-send-email-chris@chris-wilson.co.uk
Rather than persistently postponing the idle-work everytime somebody
calls i915_gem_retire_requests() (potentially ensuring that we never
reach the idle state), queue the work the first time we detect all
requests are complete. Then if in 100ms, more requests have been queued,
we will abort the idle-worker and wait again until all the new requests
have been completed.
Of course, this does depend upon the idle worker cancelling itself
gracefully from the previous patch.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467616119-4093-2-git-send-email-chris@chris-wilson.co.uk
The retire worker is a low frequency task that makes sure we retire
outstanding requests if userspace is being lax. We only need to start it
once as it remains active until the GPU is idle, so do a cheap test
before the more expensive queue_work(). A consequence of this is that we
need correct locking in the worker to make the hot path of request
submission cheap. To keep the symmetry and keep hangcheck strictly bound
by the GPU's wakelock, we move the cancel_sync(hangcheck) to the idle
worker before dropping the wakelock.
v2: Guard against RCU fouling the breadcrumbs bottom-half whilst we kick
the waiter.
v3: Remove the wakeref assertion squelching (now we hold a wakeref for
the hangcheck, any rpm error there is genuine).
v4: To prevent excess work when retiring requests, we split the busy
flag into two, a boolean to denote whether we hold the wakeref and a
bitmask of active engines.
v5: Reorder cancelling hangcheck upon idling to avoid a race where we
might cancel a hangcheck after being preempted by a new task
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
References: https://bugs.freedesktop.org/show_bug.cgi?id=88437
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467616119-4093-1-git-send-email-chris@chris-wilson.co.uk
alloc_workqueue replaces deprecated create_workqueue().
create_workqueue has been replaced with alloc_workqueue with max_active
as 0 since there is no need for throttling the number of active work items.
WQ_MEM_RECLAIM has not been set to because kfd_process_wq will not be
used in memory reclaim path.
kfd_process_wq is used for delay destruction. A work item embedded in
kfd_process gets queued to kfd_process_wq and when it executes it
destroys and frees the containing kfd_process and thus itself.
This requires a dedicated workqueue because a work item once queued, may
get freed at any point of time and any external entity cannot
flush the work item. So, in order to wait for such a work item,
it needs to be put on a dedicated workqueue.
kfd_module_exit() calls kfd_process_destroy_wq which ensures that all
pending work items are finished before the module is removed.
flush_workqueue is unnecessary since destroy_workqueue() itself calls
drain_workqueue() which flushes repeatedly till the workqueue
becomes empty.
Hence flush_workqueue has been removed.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Oded Gabbay <oded.gabbay@gmail.com>
smatch complains of:
drivers/gpu/drm/i915/intel_fbdev.c:403
intel_fb_initial_config() warn: should '1 << i' be a 64 bit type?
drivers/gpu/drm/i915/intel_fbdev.c:422 intel_fb_initial_config() warn:
should '1 << i' be a 64 bit type?
drivers/gpu/drm/i915/intel_fbdev.c:501 intel_fb_initial_config() warn:
should '1 << i' be a 64 bit type?
We are prepared to iterate over a u64 but don't limit the number of
connectors we try to configure to a maximum of 64.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/1467470166-31717-5-git-send-email-chris@chris-wilson.co.uk
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
The patchset contains a new helper in drm_fb_cma_helper.c for suspend/
resume when using cma backed framebuffers.
* 'for-next' of http://git.agner.ch/git/linux-drm-fsl-dcu:
drm/fsl-dcu: disable vblank events on CRTC disable
drm/fsl-dcu: implement suspend/resume using atomic helpers
drm/fsl-dcu: use clk helpers
drm/fsl-dcu: move layer initialization to plane file
drm/fsl-dcu: store layer registers in soc_data
drm/fb_cma_helper: add suspend helper
here's a batch of i915 fixes for 4.7.
* tag 'drm-intel-fixes-2016-06-30' of git://anongit.freedesktop.org/drm-intel:
drm/i915: Fix missing unlock on error in i915_ppgtt_info()
drm/i915: Removing PCI IDs that are no longer listed as Kabylake.
drm/i915: Add more Kabylake PCI IDs.
drm/i915: Avoid early timeout during AUX transfers
drm/i915/hsw: Avoid early timeout during LCPLL disable/restore
drm/i915/lpt: Avoid early timeout during FDI PHY reset
drm/i915/bxt: Avoid early timeout during PLL enable
drm/i915: Refresh cached DP port register value on resume
Just a few more late fixes for Polaris cards.
* 'drm-fixes-4.7' of git://people.freedesktop.org/~agd5f/linux:
drm/amd/powerplay: workaround for UVD clock issue
drm/amdgpu: add ACLK_CNTL setting for polaris10
drm/amd/powerplay: fix issue uvd dpm can't enabled on Polaris11.
drm/amd/powerplay: Workaround for Memory EDC Error on Polaris10.
drm/amd/powerplay: Update CKS on/ CKS off voltage offset calculation
drm/amd/powerplay: disable FFC.
drm/amd/powerplay: add some definition for FFC feature on polaris.
Since the tests can and do explicitly check debugfs/i915_ring_missed_irqs
for the handling of a "missed interrupt", adding it to the dmesg at INFO
is just noise. When it happens for real, we still class it as an ERROR.
Note that I have chose to remove it entirely because when we detect the
"missed interrupt" is irrelevant and the message contains no more
information than we glean from looking in debugfs.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-20-git-send-email-chris@chris-wilson.co.uk
With only a single callsite for intel_engine_cs->irq_get and ->irq_put,
we can reduce the code size by moving the common preamble into the
caller, and we can also eliminate the reference counting.
For completeness, as we are no longer doing reference counting on irq,
rename the get/put vfunctions to enable/disable respectively and are
able to review the use of posting reads. We only require the
serialisation with hardware when enabling the interrupt (i.e. so we
cannot miss an interrupt by going to sleep before the hardware truly
enables it).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-18-git-send-email-chris@chris-wilson.co.uk
If we convert the tracing over from direct use of ring->irq_get() and
over to the breadcrumb infrastructure, we only have a single user of the
ring->irq_get and so we will be able to simplify the driver routines
(eliminating the redundant validation and irq refcounting).
Process context is preferred over softirq (or even hardirq) for a couple
of reasons:
- we already utilize process context to have fast wakeup of a single
client (i.e. the client waiting for the GPU inspects the seqno for
itself following an interrupt to avoid the overhead of a context
switch before it returns to userspace)
- engine->irq_seqno() is not suitable for use from an softirq/hardirq
context as we may require long waits (100-250us) to ensure the seqno
write is posted before we read it from the CPU
A signaling framework is a requirement for enabling dma-fences.
v2: Move to a signaling framework based upon the waiter.
v3: Track the first-signal to avoid having to walk the rbtree everytime.
v4: Mark the signaler thread as RT priority to reduce latency in the
indirect wakeups.
v5: Make failure to allocate the thread fatal.
v6: Rename kthreads to i915/signal:%u
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-16-git-send-email-chris@chris-wilson.co.uk
We have testcases to ensure that seqno wraparound works fine, so we can
forgo forcing everyone to encounter seqno wraparound during early
uptime. seqno wraparound incurs a full GPU stall so not forcing it
will eliminate one jitter from the early system. Using the testcases, we
have very deterministic testing which given how difficult it would be to
debug an issue (GPU hang) stemming from a wraparound using pure
postmortem analysis I see no value in forcing a wrap during boot.
Advancing the global next_seqno after a GPU reset is equally pointless.
References? https://bugs.freedesktop.org/show_bug.cgi?id=95023
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-15-git-send-email-chris@chris-wilson.co.uk
On Ironlake, there is no command nor register to ensure that the write
from a MI_STORE command is completed (and coherent on the CPU) before the
command parser continues. This means that the ordering between the seqno
write and the subsequent user interrupt is undefined (like gen6+). So to
ensure that the seqno write is completed after the final user interrupt
we need to delay the read sufficiently to allow the write to complete.
This delay is undefined by the bspec, and empirically requires 75us even
though a register read combined with a clflush is less than 500ns. Hence,
the delay is due to an on-chip buffer rather than the latency of the write
to memory.
Note that the render ring controls this by filling the PIPE_CONTROL fifo
with stalling commands that force the earliest pipe-control with the
seqno to be completed before the command parser continues. Given that we
need a barrier operation for BSD, we may as well forgo the extra
per-batch latency by using a common per-interrupt barrier.
Studying the impact of adding the usleep shows that in both sequences of
and individual synchronous no-op batches is negligible for the media
engine (where the write now is unordered with the interrupt). Converting
the render engine over from the current glutton of pie-controls over to
the per-interrupt delays speeds up both the sequential and individual
synchronous no-ops by 20% and 60%, respectively. This speed up holds
even when looking at the throughput of small copies (4KiB->4MiB), both
serial and synchronous, by about 20%. This is because despite adding a
significant delay to the interrupt, in all likelihood we will see the
seqno write without having to apply the barrier (only in the rare corner
cases where the write is delayed on the last required is the delay
necessary).
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=94307
Testcase: igt/gem_sync #ilk
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-12-git-send-email-chris@chris-wilson.co.uk
When waiting for an interrupt (waiting for the engine to complete some
work), we know we are the only waiter to be woken on this engine. We also
know when the GPU has nearly completed our request (or at least started
processing it), so after being woken and we detect that the GPU is
active and working on our request, allow us the bottom-half (the first
waiter who wakes up to handle checking the seqno after the interrupt) to
spin for a very short while to reduce client latencies.
The impact is minimal, there was an improvement to the realtime-vs-many
clients case, but exporting the function proves useful later. However,
it is tempting to adjust irq_seqno_barrier to include the spin. The
problem is first ensuring that the "start-of-request" seqno is coherent
as we use that as our basis for judging when it is ok to spin. If we
could, spinning there could dramatically shorten some sleeps, and allow
us to make the barriers more conservative to handle missed seqno writes
on more platforms (all gen7+ are known to have the occasional issue, at
least).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-7-git-send-email-chris@chris-wilson.co.uk
One particularly stressful scenario consists of many independent tasks
all competing for GPU time and waiting upon the results (e.g. realtime
transcoding of many, many streams). One bottleneck in particular is that
each client waits on its own results, but every client is woken up after
every batchbuffer - hence the thunder of hooves as then every client must
do its heavyweight dance to read a coherent seqno to see if it is the
lucky one.
Ideally, we only want one client to wake up after the interrupt and
check its request for completion. Since the requests must retire in
order, we can select the first client on the oldest request to be woken.
Once that client has completed his wait, we can then wake up the
next client and so on. However, all clients then incur latency as every
process in the chain may be delayed for scheduling - this may also then
cause some priority inversion. To reduce the latency, when a client
is added or removed from the list, we scan the tree for completed
seqno and wake up all the completed waiters in parallel.
Using igt/benchmarks/gem_latency, we can demonstrate this effect. The
benchmark measures the number of GPU cycles between completion of a
batch and the client waking up from a call to wait-ioctl. With many
concurrent waiters, with each on a different request, we observe that
the wakeup latency before the patch scales nearly linearly with the
number of waiters (before external factors kick in making the scaling much
worse). After applying the patch, we can see that only the single waiter
for the request is being woken up, providing a constant wakeup latency
for every operation. However, the situation is not quite as rosy for
many waiters on the same request, though to the best of my knowledge this
is much less likely in practice. Here, we can observe that the
concurrent waiters incur extra latency from being woken up by the
solitary bottom-half, rather than directly by the interrupt. This
appears to be scheduler induced (having discounted adverse effects from
having a rbtree walk/erase in the wakeup path), each additional
wake_up_process() costs approximately 1us on big core. Another effect of
performing the secondary wakeups from the first bottom-half is the
incurred delay this imposes on high priority threads - rather than
immediately returning to userspace and leaving the interrupt handler to
wake the others.
To offset the delay incurred with additional waiters on a request, we
could use a hybrid scheme that did a quick read in the interrupt handler
and dequeued all the completed waiters (incurring the overhead in the
interrupt handler, not the best plan either as we then incur GPU
submission latency) but we would still have to wake up the bottom-half
every time to do the heavyweight slow read. Or we could only kick the
waiters on the seqno with the same priority as the current task (i.e. in
the realtime waiter scenario, only it is woken up immediately by the
interrupt and simply queues the next waiter before returning to userspace,
minimising its delay at the expense of the chain, and also reducing
contention on its scheduler runqueue). This is effective at avoid long
pauses in the interrupt handler and at avoiding the extra latency in
realtime/high-priority waiters.
v2: Convert from a kworker per engine into a dedicated kthread for the
bottom-half.
v3: Rename request members and tweak comments.
v4: Use a per-engine spinlock in the breadcrumbs bottom-half.
v5: Fix race in locklessly checking waiter status and kicking the task on
adding a new waiter.
v6: Fix deciding when to force the timer to hide missing interrupts.
v7: Move the bottom-half from the kthread to the first client process.
v8: Reword a few comments
v9: Break the busy loop when the interrupt is unmasked or has fired.
v10: Comments, unnecessary churn, better debugging from Tvrtko
v11: Wake all completed waiters on removing the current bottom-half to
reduce the latency of waking up a herd of clients all waiting on the
same request.
v12: Rearrange missed-interrupt fault injection so that it works with
igt/drv_missed_irq_hang
v13: Rename intel_breadcrumb and friends to intel_wait in preparation
for signal handling.
v14: RCU commentary, assert_spin_locked
v15: Hide BUG_ON behind the compiler; report on gem_latency findings.
v16: Sort seqno-groups by priority so that first-waiter has the highest
task priority (and so avoid priority inversion).
v17: Add waiters to post-mortem GPU hang state.
v18: Return early for a completed wait after acquiring the spinlock.
Avoids adding ourselves to the tree if the is already complete, and
skips the awkward question of why we don't do completion wakeups for
waits earlier than or equal to ourselves.
v19: Prepare for init_breadcrumbs to fail. Later patches may want to
allocate during init, so be prepared to propagate back the error code.
Testcase: igt/gem_concurrent_blit
Testcase: igt/benchmarks/gem_latency
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: "Rogozhkin, Dmitry V" <dmitry.v.rogozhkin@intel.com>
Cc: "Gong, Zhipeng" <zhipeng.gong@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Dave Gordon <david.s.gordon@intel.com>
Cc: "Goel, Akash" <akash.goel@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com> #v18
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-6-git-send-email-chris@chris-wilson.co.uk
Currently __i915_wait_request uses a per-engine wait_queue_t for the dual
purpose of waking after the GPU advances or for waking after an error.
In the future, we may add even more wake sources and require greater
separation, but for now we can conceptually simplify wakeups by separating
the two sources. In particular, this allows us to use different wait-queues
(e.g. one on the engine advancement, a global one for errors and one on
each requests) without any hassle.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-5-git-send-email-chris@chris-wilson.co.uk
We can forgo queuing the hangcheck from the start of every request to
until we wait upon a request. This reduces the overhead of every
request, but may increase the latency of detecting a hang. However, if
nothing every waits upon a hang, did it ever hang? It also improves the
robustness of the wait-request by ensuring that the hangchecker is
indeed running before we sleep indefinitely (and thereby ensuring that
we never actually sleep forever waiting for a dead GPU).
As pointed out by Tvrtko, it is possible for a GPU hang to go unnoticed
for as long as nobody is waiting for the GPU. Though this rare, during
that time we may be consuming more power than if we had promptly
recovered, and in the most extreme case we may exhaust all memory before
forcing the hangcheck. Something to be wary off in future.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467390209-3576-2-git-send-email-chris@chris-wilson.co.uk
When the surface backing a framebuffer doesn't match the framebuffer's
dimensions, the screen target code would test the framebuffer dimensions
rather than the surface dimensions when deciding whether to bind the
surface as a screen target directly. This causes a screen target -
surface dimension mismatch and a subsequent device error.
Fix this by testing against the surface dimension.
v2: Fix review comments by Sinclair Yeh.
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Cc: <stable@vger.kernel.org>
For the Screen Object display unit, we need to reserve a
guest-invisible region equal to the size of the framebuffer for
the host. This region can only be reserved in VRAM, whereas
the guest-visible framebuffer can be reserved in either VRAM or
GMR.
As such priority should be given to the guest-invisible
region otherwise in a limited VRAM situation, we can fail to
allocate this region.
This patch makes it so that vmw_sou_backing_alloc() is called
before the framebuffer is pinned.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Cc: <stable@vger.kernel.org>
---
This is the last patch of a 3-patch series to fix console black
screen issue on Ubuntu 16.04 server
In certain scenarios, e.g. when fbdev is enabled, we can get into
a situation where a vmw_framebuffer_pin() is called on a buffer
that is already pinned.
When this happens, ttm_bo_validate() will unintentially remove the
TTM_PL_FLAG_NO_EVICT flag, thus unpinning it, and leaving no way
to actually pin the buffer again.
To prevent this, if a buffer is already pinned, then instead of
calling ttm_bo_validate(), just make sure the proposed placement is
compatible with the existing placement.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Cc: <stable@vger.kernel.org>
---
This is the 2nd patch in a 3-patch series to fix a console black
screen issue on Ubuntu 16.04 server. This fixes a BUG_ON()
condition where a pinned buffer gets accidentally put onto the
LRU list.
There are cases where it is desired to see if a proposed placement
is compatible with a buffer object before calling ttm_bo_validate().
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Cc: <stable@vger.kernel.org>
---
This is the first of a 3-patch series to fix a black screen
issue observed on Ubuntu 16.04 server.
Offer an option for advanced users who want larger modes at 16bpp.
This becomes necessary after the fix: "Work around mode set
failure in 2D VMs." Without this patch, there would be no way
for existing advanced users to get to a high res mode, and the
regression is they will likely get a black screen after a software
update on their current VM.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Cc: <stable@vger.kernel.org>
In a low-memory 2D VM, fbdev can take up a large percentage of
available memory, making them unavailable for other DRM clients.
Since we do not take fbdev into account when filtering modes,
we end up claiming to support more modes than we actually do.
As a result, users get a black screen when setting a mode too
large for current available memory. In a low-memory VM
configuration, users can get a black screen for a mode as low
as 1024x768.
The current mode filtering mechanism keys off of
SVGA_REG_SUGGESTED_GBOBJECT_MEM_SIZE_KB, i.e. the maximum amount
of surface memory we have. Since this value is a performance
suggestion, not a hard limit, and since there should not be much
of a performance impact for a 2D VM, rather than filtering out
more modes, we will just allow ourselves to exceed the SVGA's
performance suggestion.
Also changed assumed bpp to 32 from 16 to make sure we can
actually support all the modes listed.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Cc: <stable@vger.kernel.org>
Discovered by static code analysis tool. If for some reason communication
with the host fails more than preset number of retries, return an error
instead of return garbage.
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
This function is useful in both eDP and DP modes, so split it out in
anticipation of adding DP support.
Signed-off-by: Thierry Reding <treding@nvidia.com>
