Space reservation is already safe with respect to the ring->size
modulus, but hardware only expects to see values in the range
0...ring->size-1 (inclusive) and so requires the modulus to prevent us
writing the value ring->size instead of 0. As this is only required for
the register itself, we can defer the modulus to the register update and
not perform it after every command packet. We keep the
intel_ring_advance() around in the code to provide demarcation for the
end-of-packet (which then can be compared against intel_ring_begin() as
the number of dwords emitted must match the reserved space).
v2: Assert that the ring size is a power-of-two to match assumptions in
the code. Simplify the comment before writing the tail value to explain
why the modulus is necessary.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Dave Gordon <david.s.gordon@intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1470174640-18242-13-git-send-email-chris@chris-wilson.co.uk
Rather than passing a complete set of GPU cache domains for either
invalidation or for flushing, or even both, just pass a single parameter
to the engine->emit_flush to determine the required operations.
engine->emit_flush(GPU, 0) -> engine->emit_flush(EMIT_INVALIDATE)
engine->emit_flush(0, GPU) -> engine->emit_flush(EMIT_FLUSH)
engine->emit_flush(GPU, GPU) -> engine->emit_flush(EMIT_FLUSH | EMIT_INVALIDATE)
This allows us to extend the behaviour easily in future, for example if
we want just a command barrier without the overhead of flushing.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dave Gordon <david.s.gordon@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1470174640-18242-8-git-send-email-chris@chris-wilson.co.uk
Space for flushing the GPU cache prior to completing the request is
preallocated and so cannot fail - the GPU caches will always be flushed
along with the completed request. This means we no longer have to track
whether the GPU cache is dirty between batches like we had to with the
outstanding_lazy_seqno.
With the removal of the duplication in the per-backend entry points for
emitting the obsolete lazy flush, we can then further unify the
engine->emit_flush.
v2: Expand a bit on the legacy of gpu_caches_dirty
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/1469432687-22756-18-git-send-email-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1470174640-18242-7-git-send-email-chris@chris-wilson.co.uk
The state stored in this struct is not only the information about the
buffer object, but the ring used to communicate with the hardware. Using
buffer here is overly specific and, for me at least, conflates with the
notion of buffer objects themselves.
s/struct intel_ringbuffer/struct intel_ring/
s/enum intel_ring_hangcheck/enum intel_engine_hangcheck/
s/describe_ctx_ringbuf()/describe_ctx_ring()/
s/intel_ring_get_active_head()/intel_engine_get_active_head()/
s/intel_ring_sync_index()/intel_engine_sync_index()/
s/intel_ring_init_seqno()/intel_engine_init_seqno()/
s/ring_stuck()/engine_stuck()/
s/intel_cleanup_engine()/intel_engine_cleanup()/
s/intel_stop_engine()/intel_engine_stop()/
s/intel_pin_and_map_ringbuffer_obj()/intel_pin_and_map_ring()/
s/intel_unpin_ringbuffer()/intel_unpin_ring()/
s/intel_engine_create_ringbuffer()/intel_engine_create_ring()/
s/intel_ring_flush_all_caches()/intel_engine_flush_all_caches()/
s/intel_ring_invalidate_all_caches()/intel_engine_invalidate_all_caches()/
s/intel_ringbuffer_free()/intel_ring_free()/
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/1469432687-22756-15-git-send-email-chris@chris-wilson.co.uk
Link: http://patchwork.freedesktop.org/patch/msgid/1470174640-18242-4-git-send-email-chris@chris-wilson.co.uk
'ring' is an old deprecated term for a GPU engine. Chris Wilson wants to
use the name for what is currently known as an intel_ringbuffer, but it
will be dreadfully confusing if some rings are ringbuffers but other
rings are still engines. So this patch changes the names of a bunch of
parameters called 'ring' to either 'engine' or 'engine_id' according to
what they actually are.
Signed-off-by: Dave Gordon <david.s.gordon@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1469034967-15840-3-git-send-email-david.s.gordon@intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Add this workaround to prevent hang when in place compression
is used.
References: HSD#2135774
Cc: stable@vger.kernel.org
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Arun Siluvery <arun.siluvery@linux.intel.com>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Ringbuffers are now being written to either through LLC or WC paths, so
treating them as simply iomem is no longer adequate. However, for the
older !llc hardware, the hardware is documentated as treating the TAIL
register update as serialising, so we can relax the barriers when filling
the rings (but even if it were not, it is still an uncached register write
and so serialising anyway.).
For simplicity, let's ignore the iomem annotation.
v2: Remove iomem from ringbuffer->virtual_address
v3: And for good measure add iomem elsewhere to keep sparse happy
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> #v2
Link: http://patchwork.freedesktop.org/patch/msgid/1469005202-9659-8-git-send-email-chris@chris-wilson.co.uk
Link: http://patchwork.freedesktop.org/patch/msgid/1469017917-15134-7-git-send-email-chris@chris-wilson.co.uk
dma-buf provides a generic fence class for interoperation between
drivers. Internally we use the request structure as a fence, and so with
only a little bit of interfacing we can rebase those requests on top of
dma-buf fences. This will allow us, in the future, to pass those fences
back to userspace or between drivers.
v2: The fence_context needs to be globally unique, not just unique to
this device.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jesse Barnes <jbarnes@virtuousgeek.org>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1469002875-2335-4-git-send-email-chris@chris-wilson.co.uk
Engine contains dev_priv so need to pass it in.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Chris Wilson <chris-wilson.co.uk>
Use more of the shared engine setup data for legacy engine
initialization. This time to simplify the irq initialization
code.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris-wilson.co.uk>
With the unified common engine setup done, and the execlist engine
initialization loop clearly split into two phases, we can eliminate
the separate legacy engine initialization code.
v2: Fix cleanup path for legacy.
v3: Rename constructors. (Chris Wilson)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Reviewed-by: Chris Wilson <chris-wilson.co.uk>
intel_lrc.c has a table of "logical rings" (meaning engines), while
intel_ringbuffer.c has separately open-coded initialisation for each
engine. We can deduplicate this somewhat by using the same first-stage
engine-setup function for both modes.
So here we expose the function that transfers information from the
static table of (all) known engines to the dev_priv->engine array of
engines available on this device (adjusting the names along the way)
and then embed calls to it in both the LRC and the legacy-mode setup.
Signed-off-by: Dave Gordon <david.s.gordon@intel.com>
Reviewed-by: Chris Wilson <chris-wilson.co.uk>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Prior to gen6 we didn't have per-ring IMR registers, which means that
since commit 61ff75ac20 ("drm/i915: Simplify enabling
user-interrupts with L3-remapping") we're now masking off all interrupts
when init_render_ring() gets called. That's rather rude. Let's limit
the ring IMR frobbing to machines that actually have the per-ring IMR
registers.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: 61ff75ac20 ("drm/i915: Simplify enabling user-interrupts with L3-remapping")
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1468340687-3596-1-git-send-email-ville.syrjala@linux.intel.com
Reviewd-by: Chris Wilson <chris@chris-wilson.co.uk>
Since drm_i915_private is now a subclass of drm_device we do not need to
chase the drm_i915_private->dev backpointer and can instead simply
access drm_i915_private->drm directly.
text data bss dec hex filename
1068757 4565 416 1073738 10624a drivers/gpu/drm/i915/i915.ko
1066949 4565 416 1071930 105b3a drivers/gpu/drm/i915/i915.ko
Created by the coccinelle script:
@@
struct drm_i915_private *d;
identifier i;
@@
(
- d->dev->i
+ d->drm.i
|
- d->dev
+ &d->drm
)
and for good measure the dev_priv->dev backpointer was removed entirely.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467711623-2905-4-git-send-email-chris@chris-wilson.co.uk
Since we now subclass struct drm_device, we can save pointer dances by
noting the equivalence of struct drm_device and struct drm_i915_private,
i.e. by using to_i915().
text data bss dec hex filename
1073824 4562 416 1078802 107612 drivers/gpu/drm/i915/i915.ko
1068976 4562 416 1073954 106322 drivers/gpu/drm/i915/i915.ko
Created by the coccinelle script:
@@
expression E;
identifier p;
@@
- struct drm_i915_private *p = E->dev_private;
+ struct drm_i915_private *p = to_i915(E);
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Dave Gordon <david.s.gordon@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467628477-25379-1-git-send-email-chris@chris-wilson.co.uk
Now that we have (near) universal GPU recovery code, we can inject a
real hang from userspace and not need any fakery. Not only does this
mean that the testing is far more realistic, but we can simplify the
kernel in the process.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Arun Siluvery <arun.siluvery@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467616119-4093-7-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
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
After the elimination of using the scratch page for Ironlake's
breadcrumb, we no longer need to kmap the object. We therefore can move
it into the high unmappable space and do not need to force the object to
be coherent (i.e. snooped on !llc platforms).
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/1467390209-3576-9-git-send-email-chris@chris-wilson.co.uk
By using the same address for storing the HWS on every platform, we can
remove the platform specific vfuncs and reduce the get-seqno routine to
a single read of a cached memory location.
v2: Fix semaphore_passed() to look at the signaling engine (not the
waiter's)
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/1467390209-3576-8-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
Consolidate the block of default vfuncs for dispatching the batchbuffer.
Just a minor tweak on top of Tvrtko's great job of tidying up the vfunc
initialisation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467361093-20209-1-git-send-email-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Replace per-engine initialization with a common half-programatic,
half-data driven code for ease of maintenance and compactness.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Store the semaphore offset in a temporary variable to avoid
having to get the VMA offset twice.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Replace the macro initializer with a programatic loop which
results in smaller code and hopefully just as clear.
v2: Rebase.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
The object needs to be created before semaphores can be initialized
on any ring and it makes sense to pull it out to this semaphore
dedicated helper.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
v2: Put dispatch_execbuffer before add_request. (Chris Wilson)
v3: Fix add_request and irq_seqno_barrier for gen8+.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
