When doing a GPU reset, the CSB register will be trashed and we will
lose any context-switch notifications that happened since the tasklet
was disabled. If we find that all requests on this engine were
completed, we want to make sure that the ELSP tracker is similarly empty
so that we do not feed back in the completed requests upon recovering
from the reset.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20170721123238.16428-4-chris@chris-wilson.co.uk
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Enables access to OA unit metrics for BDW, CHV, SKL and BXT which all
share (more-or-less) the same OA unit design.
Of particular note in comparison to Haswell: some OA unit HW config
state has become per-context state and as a consequence it is somewhat
more complicated to manage synchronous state changes from the cpu while
there's no guarantee of what context (if any) is currently actively
running on the gpu.
The periodic sampling frequency which can be particularly useful for
system-wide analysis (as opposed to command stream synchronised
MI_REPORT_PERF_COUNT commands) is perhaps the most surprising state to
have become per-context save and restored (while the OABUFFER
destination is still a shared, system-wide resource).
This support for gen8+ takes care to consider a number of timing
challenges involved in synchronously updating per-context state
primarily by programming all config state from the cpu and updating all
current and saved contexts synchronously while the OA unit is still
disabled.
The driver intentionally avoids depending on command streamer
programming to update OA state considering the lack of synchronization
between the automatic loading of OACTXCONTROL state (that includes the
periodic sampling state and enable state) on context restore and the
parsing of any general purpose BB the driver can control. I.e. this
implementation is careful to avoid the possibility of a context restore
temporarily enabling any out-of-date periodic sampling state. In
addition to the risk of transiently-out-of-date state being loaded
automatically; there are also internal HW latencies involved in the
loading of MUX configurations which would be difficult to account for
from the command streamer (and we only want to enable the unit when once
the MUX configuration is complete).
Since the Gen8+ OA unit design no longer supports clock gating the unit
off for a single given context (which effectively stopped any progress
of counters while any other context was running) and instead supports
tagging OA reports with a context ID for filtering on the CPU, it means
we can no longer hide the system-wide progress of counters from a
non-privileged application only interested in metrics for its own
context. Although we could theoretically try and subtract the progress
of other contexts before forwarding reports via read() we aren't in a
position to filter reports captured via MI_REPORT_PERF_COUNT commands.
As a result, for Gen8+, we always require the
dev.i915.perf_stream_paranoid to be unset for any access to OA metrics
if not root.
v5: Drain submitted requests when enabling metric set to ensure no
lite-restore erases the context image we just updated (Lionel)
v6: In addition to drain, switch to kernel context & update all
context in place (Chris)
v7: Add missing mutex_unlock() if switching to kernel context fails
(Matthew)
v8: Simplify OA period/flex-eu-counters programming by using the
batchbuffer instead of modifying ctx-image (Lionel)
v9: Back to updating the context image (due to erroneous testing,
batchbuffer programming the OA unit doesn't actually work)
(Lionel)
Pin context before updating context image (Chris)
Drop MMIO programming now that we switch to a kernel context with
right values in initial context image (Chris)
v10: Just pin_map the contexts we want to modify or let the
configuration happen on first use (Chris)
v11: Update kernel context OA config through the batchbuffer rather
than on the fly ctx-image update (Lionel)
v12: Rework OA context registers update again by swithing away from
user contexts and reconfiguring the kernel context through the
batchbuffer and updating all the other contexts' context image.
Also take care to lock slice/subslice configuration when OA is
on. (Lionel)
v13: Request rpcs updates on all engine when updating the OA config
(Lionel)
v14: Drop any kind of rpcs management now that we monitor sseu
configuration changes in a later patch (Lionel)
Remove usleep after programming the NOA configs on Gen8+, this
doesn't seem to be needed (Lionel)
v15: Respect coding style for block comments (Chris)
v16: Add missing i915_add_request() in case we fail to emit OA
configuration (Matthew)
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com> \o/
Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
Indirect Context Offset Pointer has changed for Cannonlake.
INDIRECT_CTX_OFFSET[15:6] valid value for CNL is 19h per Spec.
v2: rebased to intel_lr_indirect_ctx_offset
v3: Commit message added per Tvrtko request.
Signed-off-by: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1496781040-20888-9-git-send-email-rodrigo.vivi@intel.com
If we do not require to perform priority bumping, and we haven't yet
submitted the request, we can update its priority in situ and skip
acquiring the engine locks -- thus avoiding any contention between us
and submit/execute.
v2: Remove the stack element from the list if we can do the early
assignment.
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/20170517121007.27224-10-chris@chris-wilson.co.uk
The i915_priolist are allocated within an atomic context on a path where
we wish to minimise latency. If we use a dedicated kmem_cache, we have
the advantage of a local freelist from which to service new requests
that should keep the latency impact of an allocation small. Though
currently we expect the majority of requests to be at default priority
(and so hit the preallocate priolist), once userspace starts using
priorities they are likely to use many fine grained policies improving
the utilisation of a private slab.
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: http://patchwork.freedesktop.org/patch/msgid/20170517121007.27224-9-chris@chris-wilson.co.uk
All the requests at the same priority are executed in FIFO order. They
do not need to be stored in the rbtree themselves, as they are a simple
list within a level. If we move the requests at one priority into a list,
we can then reduce the rbtree to the set of priorities. This should keep
the height of the rbtree small, as the number of active priorities can not
exceed the number of active requests and should be typically only a few.
Currently, we have ~2k possible different priority levels, that may
increase to allow even more fine grained selection. Allocating those in
advance seems a waste (and may be impossible), so we opt for allocating
upon first use, and freeing after its requests are depleted. To avoid
the possibility of an allocation failure causing us to lose a request,
we preallocate the default priority (0) and bump any request to that
priority if we fail to allocate it the appropriate plist. Having a
request (that is ready to run, so not leading to corruption) execute
out-of-order is better than leaking the request (and its dependency
tree) entirely.
There should be a benefit to reducing execlists_dequeue() to principally
using a simple list (and reducing the frequency of both rbtree iteration
and balancing on erase) but for typical workloads, request coalescing
should be small enough that we don't notice any change. The main gain is
from improving PI calls to schedule, and the explicit list within a
level should make request unwinding simpler (we just need to insert at
the head of the list rather than the tail and not have to make the
rbtree search more complicated).
v2: Avoid use-after-free when deleting a depleted priolist
v3: Michał found the solution to handling the allocation failure
gracefully. If we disable all priority scheduling following the
allocation failure, those requests will be executed in fifo and we will
ensure that this request and its dependencies are in strict fifo (even
when it doesn't realise it is only a single list). Normal scheduling is
restored once we know the device is idle, until the next failure!
Suggested-by: Michał Wajdeczko <michal.wajdeczko@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170517121007.27224-8-chris@chris-wilson.co.uk
add/remove: 1/1 grow/shrink: 5/4 up/down: 391/-578 (-187)
function old new delta
execlists_submit_ports 262 471 +209
port_assign.isra - 136 +136
capture 6344 6359 +15
reset_common_ring 438 452 +14
execlists_submit_request 228 238 +10
gen8_init_common_ring 334 341 +7
intel_engine_is_idle 106 105 -1
i915_engine_info 2314 2290 -24
__i915_gem_set_wedged_BKL 485 411 -74
intel_lrc_irq_handler 1789 1604 -185
execlists_update_context 294 - -294
The most important change there is the improve to the
intel_lrc_irq_handler and excclist_submit_ports (net improvement since
execlists_update_context is now inlined).
v2: Use the port_api() for guc as well (even though currently we do not
pack any counters in there, yet) and hide all port->request_count inside
the helpers.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170517121007.27224-5-chris@chris-wilson.co.uk
During execlist_context_deferred_alloc() we presumed that the context is
uninitialised (we only just allocated the state object for it!) and
chose to optimise away the later call to engine->init_context() if
engine->init_context were NULL. This breaks with GVT's contexts that are
marked as pre-initialised to avoid us annoyingly calling
engine->init_context(). The fix is to not override ce->initialised if it
is already true.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1494497262-24855-1-git-send-email-chuanxiao.dong@intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Since unifying ringbuffer/execlist submission to use
engine->pin_context, we ensure that the intel_ring is available before
we start constructing the request. We can therefore move the assignment
of the request->ring to the central i915_gem_request_alloc() and not
require it in every engine->request_alloc() callback. Another small step
towards simplification (of the core, but at a cost of handling error
pointers in less important callers of engine->pin_context).
v2: Rearrange a few branches to reduce impact of PTR_ERR() on gcc's code
generation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Oscar Mateo <oscar.mateo@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Oscar Mateo <oscar.mateo@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170504093308.4137-1-chris@chris-wilson.co.uk
Pre-calculate engine context size based on engine class and device
generation and store it in the engine instance.
v2:
- Squash and get rid of hw_context_size (Chris)
v3:
- Move after MMIO init for probing on Gen7 and 8 (Chris)
- Retained rounding (Tvrtko)
v4:
- Rebase for deferred legacy context allocation
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Oscar Mateo <oscar.mateo@intel.com>
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Cc: intel-gvt-dev@lists.freedesktop.org
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
We need to keep track of the last location we ask the hw to read up to
(RING_TAIL) separately from our last write location into the ring, so
that in the event of a GPU reset we do not tell the HW to proceed into
a partially written request (which can happen if that request is waiting
for an external signal before being executed).
v2: Refactor intel_ring_reset() (Mika)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100144
Testcase: igt/gem_exec_fence/await-hang
Fixes: 821ed7df6e ("drm/i915: Update reset path to fix incomplete requests")
Fixes: d55ac5bf97 ("drm/i915: Defer transfer onto execution timeline to actual hw submission")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170425130049.26147-1-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
As we now share the execlist_port[] tracking for both execlists/guc, we
can reset the inflight count on both and report which requests are being
restarted.
Suggested-by: Michel Thierry <michel.thierry@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michel Thierry <michel.thierry@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170425103835.31871-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
We indirectly hold the runtime-pm for the intel_lrc_irq_handler() by
virtue of dev_priv->gt.awake keeping a wakeref whilst the requests are
busy. As this is not obvious from the code, add a comment.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170411175850.2470-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Technically speaking, the context size is per engine class, not per
instance.
v2: Add MISSING_CASE (Tvrtko)
v3: Rebased
v4: Restore the interface back to hiding the class lookup (Chris)
Cc: Paulo Zanoni <paulo.r.zanoni@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Oscar Mateo <oscar.mateo@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1491905472-16189-1-git-send-email-oscar.mateo@intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Or rather it is used only by intel_ring_pin() to extract the
drm_i915_private which we can easily pass in. As this is a relatively
rare operation, save the space in the struct, and as such it is even
break even in the extra code for passing around the parameter:
add/remove: 0/0 grow/shrink: 2/3 up/down: 15/-15 (0)
function old new delta
intel_init_ring_buffer 906 918 +12
execlists_context_pin 1308 1311 +3
mock_engine 407 403 -4
intel_engine_create_ring 367 363 -4
intel_ring_pin 326 319 -7
Total: Before=1261794, After=1261794, chg +0.00%
v2: Reorder intel_init_ring_buffer to keep the ring setup together:
add/remove: 0/0 grow/shrink: 2/3 up/down: 9/-15 (-6)
function old new delta
intel_init_ring_buffer 906 912 +6
execlists_context_pin 1308 1311 +3
mock_engine 407 403 -4
intel_engine_create_ring 367 363 -4
intel_ring_pin 326 319 -7
Total: Before=1261794, After=1261788, chg -0.00%
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/20170403113426.25707-1-chris@chris-wilson.co.uk
This reverts commit 6c943de668 ("drm/i915: Skip execlists_dequeue()
early if the list is empty").
The validity of using READ_ONCE there depends upon having a mb to
coordinate the assignment of engine->execlist_first inside
submit_request() and checking prior to taking the spinlock in
execlists_dequeue(). We wrote "the update to TASKLET_SCHED incurs a
memory barrier making this cross-cpu checking safe", but failed to
notice that this mb was *conditional* on the execlists being ready, i.e.
there wasn't the required mb when it was most necessary!
We could install an unconditional memory barrier to fixup the
READ_ONCE():
diff --git a/drivers/gpu/drm/i915/intel_lrc.c
b/drivers/gpu/drm/i915/intel_lrc.c
index 7dd732cb9f57..1ed164b16d44 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -616,6 +616,7 @@ static void execlists_submit_request(struct
drm_i915_gem_request *request)
if (insert_request(&request->priotree, &engine->execlist_queue))
{
engine->execlist_first = &request->priotree.node;
+ smp_wmb();
if (execlists_elsp_ready(engine))
But we have opted to remove the race as it should be rarely effective,
and saves us having to explain the necessary memory barriers which we
quite clearly failed at.
Reported-and-tested-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Fixes: 6c943de668 ("drm/i915: Skip execlists_dequeue() early if the list is empty")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170329100052.29505-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Unlocking is dangerous. In this case we combine an early update to the
out-of-queue request, because we know that it will be inserted into the
correct FIFO priority-ordered slot when it becomes ready in the future.
However, given sufficient enthusiasm, it may become ready as we are
continuing to reschedule, and so may gazump the FIFO if we have since
dropped its spinlock. The result is that it may be executed too early,
before its dependencies.
v2: Move all work into the second phase over the topological sort. This
removes the shortcut on the out-of-rbtree request to ensure that we only
adjust its priority after adjusting all of its dependencies.
Fixes: 20311bd350 ("drm/i915/scheduler: Execute requests in order of priorities")
Testcase: igt/gem_exec_whisper
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: <stable@vger.kernel.org> # v4.10+
Link: http://patchwork.freedesktop.org/patch/msgid/20170327202143.7972-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Whilst I like having the assertions clearly visible in the code, they
are quite repetitious! As we find new limits we want to incorporate into
the set of assertions, it make sense to refactor them to a common
routine.
v2: Add a guc holdout.
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/20170327131412.20293-1-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
If the request->wa_tail is 0 (because it landed exactly on the end of
the ringbuffer), when we reconstruct request->tail following a reset we
fill in an illegal value (-8 or 0x001ffff8). As a result, RING_HEAD is
never able to catch up with RING_TAIL and the GPU spins endlessly. If
the ring contains a couple of breadcrumbs, even our hangcheck is unable
to catch the busy-looping as the ACTHD and seqno continually advance.
v2: Move the wrap into a common intel_ring_wrap().
Fixes: a3aabe86a3 ("drm/i915/execlists: Reinitialise context image after GPU hang")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: <stable@vger.kernel.org> # v4.10+
Link: http://patchwork.freedesktop.org/patch/msgid/20170327130009.4678-1-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
I noticed that gcc was spilling the CSB to the stack, so rearrange the
code to be more compact. Spilling in this function is slightly more
interesting due to the mmio reads acting as memory barriers and so
end up flushing the stack spills. Still miniscule to having to do at
least the pair of uncached reads :(
function old new delta
intel_lrc_irq_handler 1039 878 -161
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170325201053.21306-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
We only need to care about the ordering of the clearing of the bit with
the uncached CSB read in order to correctly detect a new interrupt
before the read completes. The uncached read itself acts as a full
memory barrier, so we do not need to enforce another in the form of a
locked clear_bit.
v2: Clarify why the split and unlocked test/clear is harmless.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170323134803.10418-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Storing the position of the breadcrumb of the last retired request as
a separate last_retired_head is superfluous as we always copy that into
head prior to recalculation of the intel_ring.space.
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/20170321102552.24357-1-chris@chris-wilson.co.uk
Rather than impose the cost of a locked test before queuing a new
request, reduce it to a simple test_bit() with a following clear_bit()
prior to doing the CSB check. This ensure that if an interrupt does
occur whilst reading from the CSB, we still detect it (the interrupt
would trigger a rescheduling of the tasklet anyway).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170321113320.2603-1-chris@chris-wilson.co.uk
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
When switching back to execlists, we also now need to restore the
tasklet handler.
Reported-by: Oscar Mateo <oscar.mateo@intel.com>
Fixes: 31de73501a ("drm/i915/scheduler: emulate a scheduler for guc")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Cc: Oscar Mateo <oscar.mateo@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170318102859.24101-1-chris@chris-wilson.co.uk
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Do an early read of the execlists' queue before we take the spinlock and
start checking. This is safe as the first writer to the execlists queue
will cause the tasklet to be run again after a memory barrier.
v2: Keep guc in sync with execlists queue changes
v3: Explain the mb between the tasklet running on one cpu and the
execlist_first update and schedule from a second cpu.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Michał Winiarski <michal.winiarski@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170317120716.17191-1-chris@chris-wilson.co.uk
This should be impossible, but let's assert that we do not pin a context
4 billion times before retiring!
v2: Fix the assertion -- the patch had just one job to do!
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170316171628.3228-1-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
It turns out that we may want to restore the original
engine->submit_request (and engine->schedule) callbacks from more than
just the guc <-> execlists transition. Move this to a vfunc so we can
have a common interface.
v2: Move initial selection to intel_engines_init_common(), repaint vfunc
with engine->set_default_submission (and a similar colour for the
helper).
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170316171305.12972-2-chris@chris-wilson.co.uk
GVTg has introduced the context status notifier to schedule the GVTg
workload. At that time, the notifier is bound to GVTg context only,
so GVTg is not aware of host workloads.
Now we are going to improve GVTg's guest workload scheduler policy,
and add Guc emulation support for new Gen graphics. Both these two
features require acknowledgment for all contexts running on hardware.
(But will not alter host workload.) So here try to make some change.
The change is simple:
1. Move the context status notifier head from i915_gem_context to
intel_engine_cs. Which means there is a notifier head per engine
instead of per context. Execlist driver still call notifier for
each context sched-in/out events of current engine.
2. At GVTg side, it binds a notifier_block for each physical engine
at GVTg initialization period. Then GVTg can hear all context
status events.
In this patch, GVTg do nothing for host context event, but later
will add a function there. But in any case, the notifier callback is
a noop if this is no active vGPU.
Since intel_gvt_init() is called at early initialization stage and
require the status notifier head has been initiated, I initiate it in
intel_engine_setup().
v2: remove a redundant newline. (chris)
Fixes: 3c7ba6359d ("drm/i915: Introduce execlist context status change notification")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100232
Signed-off-by: Changbin Du <changbin.du@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170313024711.28591-1-changbin.du@intel.com
Acked-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This emulates execlists on top of the GuC in order to defer submission of
requests to the hardware. This deferral allows time for high priority
requests to gazump their way to the head of the queue, however it nerfs
the GuC by converting it back into a simple execlist (where the CPU has
to wake up after every request to feed new commands into the GuC).
v2: Drop hack status - though iirc there is still a lockdep inversion
between fence and engine->timeline->lock (which is impossible as the
nesting only occurs on different fences - hopefully just requires some
judicious lockdep annotation)
v3: Apply lockdep nesting to enabling signaling on the request, using
the pattern we already have in __i915_gem_request_submit();
v4: Replaying requests after a hang also now needs the timeline
spinlock, to disable the interrupts at least
v5: Hold wq lock for completeness, and emit a tracepoint for enabling signal
v6: Reorder interrupt checking for a happier gcc.
v7: Only signal the tasklet after a user-interrupt if using guc scheduling
v8: Restore lost update of rq through the i915_guc_irq_handler (Tvrtko)
v9: Avoid re-initialising the engine->irq_tasklet from inside a reset
v10: Hook up the execlists-style tracepoints
v11: Clear the execlists irq_posted bit after taking over the interrupt/tasklet
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170316125619.6856-1-chris@chris-wilson.co.uk
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
The replay bit of the ring mode register is not a valid bit for Gen8+.
Do not write to this bit.
Signed-off-by: Kelvin Gardiner <kelvin.gardiner@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Ceraolo Spurio, Daniele <daniele.ceraolospurio@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
[Joonas: Fixed commit message line to be under 72 chars]
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1487963724-4824-1-git-send-email-kelvin.gardiner@intel.com
Two new tracepoints placed at the call sites where requests are
actually passed to the GPU enable userspace to track engine
utilisation.
These tracepoints are only enabled when the
DRM_I915_LOW_LEVEL_TRACEPOINTS Kconfig option is enabled.
v2: Fix compilation with !CONFIG_DRM_I915_LOW_LEVEL_TRACEPOINTS.
v3: Name global seqno consistently across tracepoints.
v4: Remove port info from request out tracepoint. (Chris Wilson)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Compact the name of the macro and reg_state variable, and cache
some data in local variables to make the function more compact
and more readable.
v2: Fixup some checkpatch warnings.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170221095839.30525-1-tvrtko.ursulin@linux.intel.com
The hardware requires that the tail pointer only advance in qword units,
so assert that the value we write is aligned to qwords, and similarly
enforce this restriction onto the request->tail.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170217163833.731-1-chris@chris-wilson.co.uk
Reviewed-by: Michał Winiarski <michal.winiarski@intel.com>
We have a few open coded instances in the execlists code and an
almost suitable helper in intel_ringbuf.c
We can consolidate to a single helper if we change the existing
helper to emit directly to ring buffer memory and move the space
reservation outside it.
v2: Drop memcpy for memset. (Chris Wilson)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170216122325.31391-2-tvrtko.ursulin@linux.intel.com
Use the "*batch++ = " style as in the ring emission for better
readability and also simplify the logic a bit by consolidating
the offset and size calculations and overflow checking. The
latter is a programming error so it is not required to check
for it after each write to the object, but rather do it once the
whole state has been written and fail the driver if something
went wrong.
v2: Rebase.
v3: Keep track of offsets and sizes in bytes for simplicity
and rename function pointer variable to _fn suffix.
(Chris Wilson)
v4: Fix size calc broken in v3 and add alignment warning. (Chris Wilson)
v5: Fix return code.
v6: I added an exit from loop in v5 but forgot to put back
the object teardown.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> (v5)
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
This removes the usage of intel_ring_emit in favour of
directly writing to the ring buffer.
intel_ring_emit was preventing the compiler for optimising
fetch and increment of the current ring buffer pointer and
therefore generating very verbose code for every write.
It had no useful purpose since all ringbuffer operations
are started and ended with intel_ring_begin and
intel_ring_advance respectively, with no bail out in the
middle possible, so it is fine to increment the tail in
intel_ring_begin and let the code manage the pointer
itself.
Useless instruction removal amounts to approximately
two and half kilobytes of saved text on my build.
Not sure if this has any measurable performance
implications but executing a ton of useless instructions
on fast paths cannot be good.
v2:
* Change return from intel_ring_begin to error pointer by
popular demand.
* Move tail increment to intel_ring_advance to enable some
error checking.
v3:
* Move tail advance back into intel_ring_begin.
* Rebase and tidy.
v4:
* Complete rebase after a few months since v3.
v5:
* Remove unecessary cast and fix !debug compile. (Chris Wilson)
v6:
* Make intel_ring_offset take request as well.
* Fix recording of request postfix plus a sprinkle of asserts.
(Chris Wilson)
v7:
* Use intel_ring_offset to get the postfix. (Chris Wilson)
* Convert GVT code as well.
v8:
* Rename *out++ to *cs++.
v9:
* Fix GVT out to cs conversion in GVT.
v10:
* Rebase for new intel_ring_begin in selftests.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Zhi Wang <zhi.a.wang@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170214113242.29241-1-tvrtko.ursulin@linux.intel.com
After a brief discussion, we settled on a naming convention for the
conditional GEM debugging data that should be clearer to the casual
user: GEM_DEBUG
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170207102319.10910-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Now that we have fast top-down insertion into the drm_mm, we can use it
for frequent runtime operations like insertion of the context object,
whereas before we limited it to the one-off insertion of the pinned
kernel context. Keeping the active context objects out of the mappable
region of the global GTT (except under memory pressure) improves our
ability to allocate mappable aperture region without triggering a GPU
stall.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170210101422.1598-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Once the address space has been created (using 3 or 4 levels of page
tables), we should use that to program the appropriate type into the
contexts. This gives us the flexibility to handle different types of
address spaces at runtime.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170209144036.23664-1-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Following a reset, the context and page directory registers are lost.
However, the queue of requests that we resubmit after the reset may
depend upon them - the registers are restored from a context image, but
that restore may be inhibited and may simply be absent from the request
if it was in the middle of a sequence using the same context. If we
prime the CCID/PD registers with the first request in the queue (even
for the hung request), we prevent invalid memory access for the
following requests (and continually hung engines).
v2: Magic BIT(8), reserved for future use but still appears unused.
v3: Some commentary on handling innocent vs guilty requests
v4: Add a wait for PD_BASE fetch. The reload appears to be instant on my
Ivybridge, but this bit probably exists for a reason.
Fixes: 821ed7df6e ("drm/i915: Update reset path to fix incomplete requests")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170207152437.4252-1-chris@chris-wilson.co.uk
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
In commit 86aa7e760a ("drm/i915: Assert that the context-switch
completion matches our context") I added a read to the irq tasklet
handler that compared the on-chip status with that of our sw tracking,
using an unguarded read of the request pointer to get the context and
beyond. Whilst we hold a reference to the request, we do not hold
anything on the context and if we are unlucky it may be reaped from a
second thread retiring the request (since it may retire the request as
soon as the breadcrumb is complete, even before we finish processing the
context switch) as we try to read from the context pointer.
Avoid the racy read from underneath the request by storing the expected
result in the execlist_port[].
v2: Include commentary about port[].request being unprotected.
Fixes: 86aa7e760a ("drm/i915: Assert that the context-switch completion matches our context")
Reported-by: Mika Kuoppala <mika.kuoppala@intel.com>
Testcase: igt/gem_ctx_create
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170206170502.30944-2-chris@chris-wilson.co.uk
execlist_update_context() will try to update PDPs in a context before a
ELSP submission only for full PPGTT mode, while PDPs was populated during
context initialization. Now the latter code path is removed. Let
execlist_update_context() also cover !FULL_PPGTT mode.
Fixes: 34869776c7 ("drm/i915: check ppgtt validity when init reg state")
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Cc: Michal Winiarski <michal.winiarski@intel.com>
Cc: Michel Thierry <michel.thierry@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Cc: Zhiyuan Lv <zhiyuan.lv@intel.com>
Signed-off-by: Zhi Wang <zhi.a.wang@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1486377436-15380-1-git-send-email-zhi.a.wang@intel.com
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Chris Wilson