Similar to commit 49b20dbf74 ("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
Since we allow removing the timeline map at runtime, there is a risk
that rq->hwsp points into a stale page. To control that risk, we hold
the RCU read lock while reading *rq->hwsp, but we missed a couple of
important barriers. First, the unpinning / removal of the timeline map
must be after all RCU readers into that map are complete, i.e. after an
rcu barrier (in this case courtesy of call_rcu()). Secondly, we must
make sure that the rq->hwsp we are about to dereference under the RCU
lock is valid. In this case, we make the rq->hwsp pointer safe during
i915_request_retire() and so we know that rq->hwsp may become invalid
only after the request has been signaled. Therefore is the request is
not yet signaled when we acquire rq->hwsp under the RCU, we know that
rq->hwsp will remain valid for the duration of the RCU read lock.
This is a very small window that may lead to either considering the
request not completed (causing a delay until the request is checked
again, any wait for the request is not affected) or dereferencing an
invalid pointer.
Fixes: 3adac4689f ("drm/i915: Introduce concept of per-timeline (context) HWSP")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: <stable@vger.kernel.org> # v5.1+
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20201218122421.18344-1-chris@chris-wilson.co.uk
(cherry picked from commit 9bb36cf660)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210118101755.476744-1-chris@chris-wilson.co.uk
Backmerging to get a common base for merging topic branches between
drm-intel-next and drm-intel-gt-next.
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
UAPI Changes:
- Deprecate I915_PMU_LAST and optimize state tracking (Tvrtko)
Avoid relying on last item ABI marker in i915_drm.h, add a
comment to mark as deprecated.
Cross-subsystem Changes:
Core Changes:
Driver Changes:
- Restore clear residuals security mitigations for Ivybridge and
Baytrail (Chris)
- Close#1858: Allow sysadmin to choose applied GPU security mitigations
through i915.mitigations=... similar to CPU (Chris)
- Fix for #2024: GPU hangs on HSW GT1 (Chris)
- Fix for #2707: Driver hang when editing UVs in Blender (Chris, Ville)
- Fix for #2797: False positive GuC loading error message (Chris)
- Fix for #2859: Missing GuC firmware for older Cometlakes (Chris)
- Lessen probability of GPU hang due to DMAR faults [reason 7,
next page table ptr is invalid] on Tigerlake (Chris)
- Fix REVID macros for TGL to fetch correct stepping (Aditya)
- Limit frequency drop to RPe on parking (Chris, Edward)
- Limit W/A 1406941453 to TGL, RKL and DG1 (Swathi)
- Make W/A 22010271021 permanent on DG1 (Lucas)
- Implement W/A 16011163337 to prevent a HS/DS hang on DG1 (Swathi)
- Only disable preemption on gen8 render engines (Chris)
- Disable arbitration around Braswell's PDP updates (Chris)
- Disable arbitration on no-preempt requests (Chris)
- Check for arbitration after writing start seqno before busywaiting (Chris)
- Retain default context state across shrinking (Venkata, CQ)
- Fix mismatch between misplaced vma check and vma insert for 32-bit
addressing userspaces (Chris, CQ)
- Propagate error for vmap() failure instead kernel NULL deref (Chris)
- Propagate error from cancelled submit due to context closure
immediately (Chris)
- Fix RCU race on HWSP tracking per request (Chris)
- Clear CMD parser shadow and GPU reloc batches (Matt A)
- Populate logical context during first pin (Maarten)
- Optimistically prune dma-resv from the shrinker (Chris)
- Fix for virtual engine ownership race (Chris)
- Remove timeslice suppression to restore fairness for virtual engines (Chris)
- Rearrange IVB/HSW workarounds properly between GT and engine (Chris)
- Taint the reset mutex with the shrinker (Chris)
- Replace direct submit with direct call to tasklet (Chris)
- Multiple corrections to virtual engine dequeue and breadcrumbs code (Chris)
- Avoid wakeref from potentially hard IRQ context in PMU (Tvrtko)
- Use raw clock for RC6 time estimation in PMU (Tvrtko)
- Differentiate OOM failures from invalid map types (Chris)
- Fix Gen9 to have 64 MOCS entries similar to Gen11 (Chris)
- Ignore repeated attempts to suspend request flow across reset (Chris)
- Remove livelock from "do_idle_maps" VT-d W/A (Chris)
- Cancel the preemption timeout early in case engine reset fails (Chris)
- Code flow optimization in the scheduling code (Chris)
- Clear the execlists timers upon reset (Chris)
- Drain the breadcrumbs just once (Chris, Matt A)
- Track the overall GT awake/busy time (Chris)
- Tweak submission tasklet flushing to avoid starvation (Chris)
- Track timelines created using the HWSP to restore on resume (Chris)
- Use cmpxchg64 for 32b compatilibity for active tracking (Chris)
- Prefer recycling an idle GGTT fence to avoid GPU wait (Chris)
- Restructure GT code organization for clearer split between GuC
and execlists (Chris, Daniele, John, Matt A)
- Remove GuC code that will remain unused by new interfaces (Matt B)
- Restructure the CS timestamp clocks code to local to GT (Chris)
- Fix error return paths in perf code (Zhang)
- Replace idr_init() by idr_init_base() in perf (Deepak)
- Fix shmem_pin_map error path (Colin)
- Drop redundant free_work worker for GEM contexts (Chris, Mika)
- Increase readability and understandability of intel_workarounds.c (Lucas)
- Defer enabling the breadcrumb interrupt to after submission (Chris)
- Deal with buddy alloc block sizes beyond 4G (Venkata, Chris)
- Encode fence specific waitqueue behaviour into the wait.flags (Chris)
- Don't cancel the breadcrumb interrupt shadow too early (Chris)
- Cancel submitted requests upon context reset (Chris)
- Use correct locks in GuC code (Tvrtko)
- Prevent use of engine->wa_ctx after error (Chris, Matt R)
- Fix build warning on 32-bit (Arnd)
- Avoid memory leak if platform would have more than 16 W/A (Tvrtko)
- Avoid unnecessary #if CONFIG_PM in PMU code (Chris, Tvrtko)
- Improve debugging output (Chris, Tvrtko, Matt R)
- Make file local variables static (Jani)
- Avoid uint*_t types in i915 (Jani)
- Selftest improvements (Chris, Matt A, Dan)
- Documentation fixes (Chris, Jose)
Signed-off-by: Dave Airlie <airlied@redhat.com>
# Conflicts:
# drivers/gpu/drm/i915/gt/intel_breadcrumbs.c
# drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h
# drivers/gpu/drm/i915/gt/intel_lrc.c
# drivers/gpu/drm/i915/gvt/mmio_context.h
# drivers/gpu/drm/i915/i915_drv.h
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210114152232.GA21588@jlahtine-mobl.ger.corp.intel.com
Remove the extraneous inlines. The only split by the compiler that
looked dubious was execlists_schedule_out, so push the code around
slightly to move all the work into the out-of-line function.
In a normal build, bloat-o-meter shows that only the
execlists_schedule_out is contentious:
add/remove: 1/0 grow/shrink: 0/2 up/down: 803/-1532 (-729)
Function old new delta
__execlists_schedule_out - 803 +803
execlists_submission_tasklet 6488 5766 -722
execlists_reset_csb.constprop 1587 777 -810
Total: Before=1605815, After=1605086, chg -0.05%
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210113151112.15212-1-chris@chris-wilson.co.uk
Since we are system_highpri_wq, we expected the heartbeat to be
scheduled promptly. However, we see delays of over 10ms upsetting our
assertions. Accept this as inevitable and bump the minimum error
threshold to 20ms (from 6 jiffies).
<6> [616.784749] rcs0: Heartbeat delay: 3570us [2802, 9188]
<6> [616.807790] bcs0: Heartbeat delay: 2111us [745, 4372]
<6> [616.853776] vcs0: Heartbeat delay: 6485us [2424, 11637]
<3> [616.859296] vcs0: Heartbeat delay was 6485us, expected less than 6000us
<3> [616.860901] i915/intel_heartbeat_live_selftests: live_heartbeat_fast failed with error -22
v2: More context from CI.
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/20210113163115.5740-2-chris@chris-wilson.co.uk
Delete GuC code unused in future patches that rewrite the GuC interface
to work with the new firmware. Most of the code deleted relates to
workqueues or execlist port. The code is safe to remove because we still
don't allow GuC submission to be enabled, even when overriding the
modparam, so it currently can't be reached.
The defines + structs for the process descriptor and workqueue remain.
Although the new GuC interface does not require either of these for the
normal submission path multi-lrc submission does. The usage of the
process descriptor and workqueue for multi-lrc will be quite different
from the code that is deleted in this patch. A future patch will
implement multi-lrc submission.
v2: add a code in the commit message about the code being safe to
remove (Chris)
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: John Harrison <john.c.harrison@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/20210113021236.8164-2-daniele.ceraolospurio@intel.com
During igt_reset_nop_engine, it was observed that an unexpected failed
engine reset lead to us busywaiting on the stop-ring semaphore (set
during the reset preparations) on the first request afterwards. There was
no explicit MI_ARB_CHECK in this sequence as the presumption was that
the failed MI_SEMAPHORE_WAIT would itself act as an arbitration point.
It did not in this circumstance, so force it.
This patch is based on the assumption that the MI_SEMAPHORE_WAIT failure
to arbitrate is a rare Tigerlake bug, similar to the lite-restore vs
semaphore issues previously seen in the CS. The explicit MI_ARB_CHECK
should always ensure that there is at least one arbitration point in the
request before the MI_SEMAPHORE_WAIT to trigger the IDLE->ACTIVE event.
Upon processing that event, we will clear the stop-ring flag and release
the semaphore from its busywait.
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/20210112100759.32698-2-chris@chris-wilson.co.uk
On the off chance that we need to arbitrate before launching the
payload, perform the check after we signal the request is ready to
start. Assuming instantaneous processing of the CS event, the request
will then be treated as having started when we make the decisions as to
how to process that CS event.
v2: More commentary about the users of i915_request_started() as a
reminder about why we are marking the initial breadcrumb.
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/20210112100759.32698-1-chris@chris-wilson.co.uk
The reason why we did not enable preemption on Broadwater was due to
missing GPGPU workarounds. Since this only applies to rcs0, only
restrict rcs0 (and our global capabilities).
While this does not affect exposing a preemption capability to
userspace, it does affect our internal decisions on whether to use
timeslicing and semaphores between individual 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/20210108204026.20682-6-chris@chris-wilson.co.uk
We use the completion of the last active breadcrumb to retire the
requests along a timeline. This is purely opportunistic as nothing
guarantees that any particular timeline is terminated by a breadcrumb;
except for parking the engine where we explicitly add a breadcrumb so
that we park quickly and do an explicit retire upon signaling to reduce
the latency dramatically (avoiding a retire worker roundtrip).
With scheduling, we anticipate retiring completed timelines as a matter
of course. Performing the same action from inside the breadcrumbs is
intended to provide similar functionality for legacy ringbuffer
submission.
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/20210108204026.20682-5-chris@chris-wilson.co.uk
Before we mark the virtual engine as no longer inflight, flush any
ongoing signaling that may be using the ce->signal_link along the
previous breadcrumbs. On switch to a new physical engine, that link will
be inserted into the new set of breadcrumbs, causing confusion to an
ongoing iterator.
This patch undoes a last minute mistake introduced into commit
bab0557c8d ("drm/i915/gt: Remove virtual breadcrumb before transfer"),
whereby instead of unconditionally applying the flush, it was only
applied if the request itself was going to be reused.
v2: Generalise and cancel all remaining ce->signals
Fixes: bab0557c8d ("drm/i915/gt: Remove virtual breadcrumb before transfer")
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/20210108204026.20682-4-chris@chris-wilson.co.uk
In the next^W future patch, we remove the strict priority system and
continuously re-evaluate the relative priority of tasks. As such we need
to enable the timeslice whenever there is more than one context in the
pipeline. This simplifies the decision and removes some of the tweaks to
suppress timeslicing, allowing us to lift the timeslice enabling to a
common spot at the end of running the submission tasklet.
One consequence of the suppression is that it was reducing fairness
between virtual engines on an over saturated system; undermining the
principle for timeslicing.
v2: Commentary
v3: Commentary for the right cancel_timer()
v4: Add tracing for why we need a timeslice
Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2802
Testcase: igt/gem_exec_balancer/fairslice
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>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210107132322.28373-1-chris@chris-wilson.co.uk
