Some users require that when a master batch is executed on one particular
engine, a companion batch is run simultaneously on a specific slave
engine. For this purpose, we introduce virtual engine bonding, allowing
maps of master:slaves to be constructed to constrain which physical
engines a virtual engine may select given a fence on a master engine.
For the moment, we continue to ignore the issue of preemption deferring
the master request for later. Ideally, we would like to then also remove
the slave and run something else rather than have it stall the pipeline.
With load balancing, we should be able to move workload around it, but
there is a similar stall on the master pipeline while it may wait for
the slave to be executed. At the cost of more latency for the bonded
request, it may be interesting to launch both on their engines in
lockstep. (Bubbles abound.)
Opens: Also what about bonding an engine as its own master? It doesn't
break anything internally, so allow the silliness.
v2: Emancipate the bonds
v3: Couple in delayed scheduling for the selftests
v4: Handle invalid mutually exclusive bonding
v5: Mention what the uapi does
v6: s/nbond/num_bonds/
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/20190521211134.16117-9-chris@chris-wilson.co.uk
The handling of the no-preemption priority level imposes the restriction
that we need to maintain the implied ordering even though preemption is
disabled. Otherwise we may end up with an AB-BA deadlock across multiple
engine due to a real preemption event reordering the no-preemption
WAITs. To resolve this issue we currently promote all requests to WAIT
on unsubmission, however this interferes with the timeslicing
requirement that we do not apply any implicit promotion that will defeat
the round-robin timeslice list. (If we automatically promote the active
request it will go back to the head of the queue and not the tail!)
So we need implicit promotion to prevent reordering around semaphores
where we are not allowed to preempt, and we must avoid implicit
promotion on unsubmission. So instead of at unsubmit, if we apply that
implicit promotion on adding the dependency, we avoid the semaphore
deadlock and we also reduce the gains made by the promotion for user
space waiting. Furthermore, by keeping the earlier dependencies at a
higher level, we reduce the search space for timeslicing without
altering runtime scheduling too badly (no dependencies at all will be
assigned a higher priority for rrul).
v2: Limit the bump to external edges (as originally intended) i.e.
between contexts and out to the user.
Testcase: igt/gem_concurrent_blit
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/20190515130052.4475-3-chris@chris-wilson.co.uk
(cherry picked from commit 6e7eb7a807)
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Add a live selftest to excercise rotated/remapped vmas. We simply
write through the rotated/remapped vma, and confirm that the data
appears in the right page when read through the normal vma.
Not sure what the fallout of making all rotated/remapped vmas
mappable/fenceable would be, hence I just hacked it in the test.
v2: Grab rpm reference (Chris)
GEM_BUG_ON(view.type not as expected) (Chris)
Allow CAN_FENCE for rotated/remapped vmas (Chris)
Update intel_plane_uses_fence() to ask for a fence
only for normal vmas on gen4+
v3: Deal with intel_wakeref_t
v4: Rebase
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190509122159.24376-4-ville.syrjala@linux.intel.com
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Replace the racy continuation check within retire_work with a definite
kill-switch on idling. The race was being exposed by gem_concurrent_blit
where the retire_worker would be terminated too early leaving us
spinning in debugfs/i915_drop_caches with nothing flushing the
retirement queue.
Although that the igt is trying to idle from one child while submitting
from another may be a contributing factor as to why it runs so slowly...
v2: Use the non-sync version of cancel_delayed_work(), we only need to
stop it from being scheduled as we independently check whether now is
the right time to be parking.
Testcase: igt/gem_concurrent_blit
Fixes: 79ffac8599 ("drm/i915: Invert the GEM wakeref hierarchy")
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/20190507121108.18377-3-chris@chris-wilson.co.uk
The original intent for the delay before running the idle_work was to
provide a hysteresis to avoid ping-ponging the device runtime-pm. Since
then we have also pulled in some memory management and general device
management for parking. But with the inversion of the wakeref handling,
GEM is no longer responsible for the wakeref and by the time we call the
idle_work, the device is asleep. It seems appropriate now to drop the
delay and just run the worker immediately to flush the cached GEM state
before sleeping.
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/20190507121108.18377-2-chris@chris-wilson.co.uk
We switched to a tree of per-engine HW context to accommodate the
introduction of virtual engines. However, we plan to also support
multiple instances of the same engine within the GEM context, defeating
our use of the engine as a key to looking up the HW context. Just
allocate a logical per-engine instance and always use an index into the
ctx->engines[]. Later on, this ctx->engines[] may be replaced by a user
specified map.
v2: Add for_each_gem_engine() helper to iterator within the engines lock
v3: intel_context_create_request() helper
v4: s/unsigned long/unsigned int/ 4 billion engines is quite enough.
v5: Push iterator locking to caller
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/20190426163336.15906-7-chris@chris-wilson.co.uk
In the current scheme, on submitting a request we take a single global
GEM wakeref, which trickles down to wake up all GT power domains. This
is undesirable as we would like to be able to localise our power
management to the available power domains and to remove the global GEM
operations from the heart of the driver. (The intent there is to push
global GEM decisions to the boundary as used by the GEM user interface.)
Now during request construction, each request is responsible via its
logical context to acquire a wakeref on each power domain it intends to
utilize. Currently, each request takes a wakeref on the engine(s) and
the engines themselves take a chipset wakeref. This gives us a
transition on each engine which we can extend if we want to insert more
powermangement control (such as soft rc6). The global GEM operations
that currently require a struct_mutex are reduced to listening to pm
events from the chipset GT wakeref. As we reduce the struct_mutex
requirement, these listeners should evaporate.
Perhaps the biggest immediate change is that this removes the
struct_mutex requirement around GT power management, allowing us greater
flexibility in request construction. Another important knock-on effect,
is that by tracking engine usage, we can insert a switch back to the
kernel context on that engine immediately, avoiding any extra delay or
inserting global synchronisation barriers. This makes tracking when an
engine and its associated contexts are idle much easier -- important for
when we forgo our assumed execution ordering and need idle barriers to
unpin used contexts. In the process, it means we remove a large chunk of
code whose only purpose was to switch back to the kernel context.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Imre Deak <imre.deak@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190424200717.1686-5-chris@chris-wilson.co.uk
Start acquiring the logical intel_context and using that as our primary
means for request allocation. This is the initial step to allow us to
avoid requiring struct_mutex for request allocation along the
perma-pinned kernel context, but it also provides a foundation for
breaking up the complex request allocation to handle different scenarios
inside execbuf.
For the purpose of emitting a request from inside retirement (see the
next patch for engine power management), we also need to lift control
over the timeline mutex to the caller.
v2: Note that the request carries the active reference upon construction.
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/20190424200717.1686-4-chris@chris-wilson.co.uk
The RING_NONPRIV allows us to add registers to a whitelist that allows
userspace to modify them. Ideally such registers should be safe and
saved within the context such that they do not impact system behaviour
for other users. This selftest verifies that those registers we do add
are (a) then writable by userspace and (b) only affect a single client.
Opens:
- Is GEN9_SLICE_COMMON_ECO_CHICKEN1 really write-only?
v2: Remove the blatant copy-paste.
v3: Emulate userspace register writes via the batch again.
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/20190424110941.9869-1-chris@chris-wilson.co.uk
When we introduced preemption, we chose to keep it disabled for gen8 as
supporting preemption inside GPGPU user batches required various w/a in
userspace. Since then, the desire to preempt long queues of requests
between batches (e.g. within busywaiting semaphores) has grown. So allow
arbitration within the busywaits and between requests, but disable
arbitration within user batches so that we can preempt between requests
and not risk breaking GPGPU.
However, since this preemption is much coarser and doesn't interfere
with userspace, we decline to include it amongst the scheduler
capabilities. (This is also required for us to skip over the preemption
selftests that expect to be able to preempt user batches.)
Michal suggested that we could perhaps allow preemption inside gen8
userspace batches if we can satisfy ourselves that the default
preemption settings are viable with existing userspace (principally
OpenCL which already should carry any known workaround). We could then
merge the two code paths back into one, even dropping the artifical
has-preemption device feature flag.
Testcase: igt/gem_exec_scheduler/semaphore-user
References: beecec9017 ("drm/i915/execlists: Preemption!")
Fixes: e886196469 ("drm/i915: Use HW semaphores for inter-engine synchronisation on gen8+")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michal Winiarski <michal.winiarski@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Michal Winiarski <michal.winiarski@intel.com> #irc
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190329134024.5254-1-chris@chris-wilson.co.uk
