diff options
Diffstat (limited to 'drivers/gpu/drm/i915/i915_request.c')
| -rw-r--r-- | drivers/gpu/drm/i915/i915_request.c | 346 |
1 files changed, 277 insertions, 69 deletions
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index c2a5c48c7541..0a3d94517d0a 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -22,16 +22,31 @@ * */ -#include <linux/prefetch.h> #include <linux/dma-fence-array.h> +#include <linux/irq_work.h> +#include <linux/prefetch.h> #include <linux/sched.h> #include <linux/sched/clock.h> #include <linux/sched/signal.h> #include "i915_drv.h" #include "i915_active.h" +#include "i915_globals.h" #include "i915_reset.h" +struct execute_cb { + struct list_head link; + struct irq_work work; + struct i915_sw_fence *fence; +}; + +static struct i915_global_request { + struct i915_global base; + struct kmem_cache *slab_requests; + struct kmem_cache *slab_dependencies; + struct kmem_cache *slab_execute_cbs; +} global; + static const char *i915_fence_get_driver_name(struct dma_fence *fence) { return "i915"; @@ -68,7 +83,9 @@ static signed long i915_fence_wait(struct dma_fence *fence, bool interruptible, signed long timeout) { - return i915_request_wait(to_request(fence), interruptible, timeout); + return i915_request_wait(to_request(fence), + interruptible | I915_WAIT_PRIORITY, + timeout); } static void i915_fence_release(struct dma_fence *fence) @@ -84,7 +101,7 @@ static void i915_fence_release(struct dma_fence *fence) */ i915_sw_fence_fini(&rq->submit); - kmem_cache_free(rq->i915->requests, rq); + kmem_cache_free(global.slab_requests, rq); } const struct dma_fence_ops i915_fence_ops = { @@ -177,12 +194,10 @@ static void free_capture_list(struct i915_request *request) static void __retire_engine_request(struct intel_engine_cs *engine, struct i915_request *rq) { - GEM_TRACE("%s(%s) fence %llx:%lld, global=%d, current %d:%d\n", + GEM_TRACE("%s(%s) fence %llx:%lld, current %d\n", __func__, engine->name, rq->fence.context, rq->fence.seqno, - rq->global_seqno, - hwsp_seqno(rq), - intel_engine_get_seqno(engine)); + hwsp_seqno(rq)); GEM_BUG_ON(!i915_request_completed(rq)); @@ -241,12 +256,10 @@ static void i915_request_retire(struct i915_request *request) { struct i915_active_request *active, *next; - GEM_TRACE("%s fence %llx:%lld, global=%d, current %d:%d\n", + GEM_TRACE("%s fence %llx:%lld, current %d\n", request->engine->name, request->fence.context, request->fence.seqno, - request->global_seqno, - hwsp_seqno(request), - intel_engine_get_seqno(request->engine)); + hwsp_seqno(request)); lockdep_assert_held(&request->i915->drm.struct_mutex); GEM_BUG_ON(!i915_sw_fence_signaled(&request->submit)); @@ -288,15 +301,13 @@ static void i915_request_retire(struct i915_request *request) i915_request_remove_from_client(request); - /* Retirement decays the ban score as it is a sign of ctx progress */ - atomic_dec_if_positive(&request->gem_context->ban_score); intel_context_unpin(request->hw_context); __retire_engine_upto(request->engine, request); unreserve_gt(request->i915); - i915_sched_node_fini(request->i915, &request->sched); + i915_sched_node_fini(&request->sched); i915_request_put(request); } @@ -305,12 +316,10 @@ void i915_request_retire_upto(struct i915_request *rq) struct intel_ring *ring = rq->ring; struct i915_request *tmp; - GEM_TRACE("%s fence %llx:%lld, global=%d, current %d:%d\n", + GEM_TRACE("%s fence %llx:%lld, current %d\n", rq->engine->name, rq->fence.context, rq->fence.seqno, - rq->global_seqno, - hwsp_seqno(rq), - intel_engine_get_seqno(rq->engine)); + hwsp_seqno(rq)); lockdep_assert_held(&rq->i915->drm.struct_mutex); GEM_BUG_ON(!i915_request_completed(rq)); @@ -326,9 +335,67 @@ void i915_request_retire_upto(struct i915_request *rq) } while (tmp != rq); } -static u32 timeline_get_seqno(struct i915_timeline *tl) +static void irq_execute_cb(struct irq_work *wrk) +{ + struct execute_cb *cb = container_of(wrk, typeof(*cb), work); + + i915_sw_fence_complete(cb->fence); + kmem_cache_free(global.slab_execute_cbs, cb); +} + +static void __notify_execute_cb(struct i915_request *rq) +{ + struct execute_cb *cb; + + lockdep_assert_held(&rq->lock); + + if (list_empty(&rq->execute_cb)) + return; + + list_for_each_entry(cb, &rq->execute_cb, link) + irq_work_queue(&cb->work); + + /* + * XXX Rollback on __i915_request_unsubmit() + * + * In the future, perhaps when we have an active time-slicing scheduler, + * it will be interesting to unsubmit parallel execution and remove + * busywaits from the GPU until their master is restarted. This is + * quite hairy, we have to carefully rollback the fence and do a + * preempt-to-idle cycle on the target engine, all the while the + * master execute_cb may refire. + */ + INIT_LIST_HEAD(&rq->execute_cb); +} + +static int +i915_request_await_execution(struct i915_request *rq, + struct i915_request *signal, + gfp_t gfp) { - return tl->seqno += 1 + tl->has_initial_breadcrumb; + struct execute_cb *cb; + + if (i915_request_is_active(signal)) + return 0; + + cb = kmem_cache_alloc(global.slab_execute_cbs, gfp); + if (!cb) + return -ENOMEM; + + cb->fence = &rq->submit; + i915_sw_fence_await(cb->fence); + init_irq_work(&cb->work, irq_execute_cb); + + spin_lock_irq(&signal->lock); + if (i915_request_is_active(signal)) { + i915_sw_fence_complete(cb->fence); + kmem_cache_free(global.slab_execute_cbs, cb); + } else { + list_add_tail(&cb->link, &signal->execute_cb); + } + spin_unlock_irq(&signal->lock); + + return 0; } static void move_to_timeline(struct i915_request *request, @@ -342,42 +409,33 @@ static void move_to_timeline(struct i915_request *request, spin_unlock(&request->timeline->lock); } -static u32 next_global_seqno(struct i915_timeline *tl) -{ - if (!++tl->seqno) - ++tl->seqno; - return tl->seqno; -} - void __i915_request_submit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; - u32 seqno; - GEM_TRACE("%s fence %llx:%lld -> global=%d, current %d:%d\n", + GEM_TRACE("%s fence %llx:%lld -> current %d\n", engine->name, request->fence.context, request->fence.seqno, - engine->timeline.seqno + 1, - hwsp_seqno(request), - intel_engine_get_seqno(engine)); + hwsp_seqno(request)); GEM_BUG_ON(!irqs_disabled()); lockdep_assert_held(&engine->timeline.lock); - GEM_BUG_ON(request->global_seqno); - - seqno = next_global_seqno(&engine->timeline); - GEM_BUG_ON(!seqno); - GEM_BUG_ON(intel_engine_signaled(engine, seqno)); + if (i915_gem_context_is_banned(request->gem_context)) + i915_request_skip(request, -EIO); /* We may be recursing from the signal callback of another i915 fence */ spin_lock_nested(&request->lock, SINGLE_DEPTH_NESTING); + GEM_BUG_ON(test_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags)); set_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags); - request->global_seqno = seqno; + if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags) && !i915_request_enable_breadcrumb(request)) intel_engine_queue_breadcrumbs(engine); + + __notify_execute_cb(request); + spin_unlock(&request->lock); engine->emit_fini_breadcrumb(request, @@ -406,12 +464,10 @@ void __i915_request_unsubmit(struct i915_request *request) { struct intel_engine_cs *engine = request->engine; - GEM_TRACE("%s fence %llx:%lld <- global=%d, current %d:%d\n", + GEM_TRACE("%s fence %llx:%lld, current %d\n", engine->name, request->fence.context, request->fence.seqno, - request->global_seqno, - hwsp_seqno(request), - intel_engine_get_seqno(engine)); + hwsp_seqno(request)); GEM_BUG_ON(!irqs_disabled()); lockdep_assert_held(&engine->timeline.lock); @@ -420,18 +476,25 @@ void __i915_request_unsubmit(struct i915_request *request) * Only unwind in reverse order, required so that the per-context list * is kept in seqno/ring order. */ - GEM_BUG_ON(!request->global_seqno); - GEM_BUG_ON(request->global_seqno != engine->timeline.seqno); - GEM_BUG_ON(intel_engine_has_completed(engine, request->global_seqno)); - engine->timeline.seqno--; /* We may be recursing from the signal callback of another i915 fence */ spin_lock_nested(&request->lock, SINGLE_DEPTH_NESTING); - request->global_seqno = 0; + + /* + * As we do not allow WAIT to preempt inflight requests, + * once we have executed a request, along with triggering + * any execution callbacks, we must preserve its ordering + * within the non-preemptible FIFO. + */ + BUILD_BUG_ON(__NO_PREEMPTION & ~I915_PRIORITY_MASK); /* only internal */ + request->sched.attr.priority |= __NO_PREEMPTION; + if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &request->fence.flags)) i915_request_cancel_breadcrumb(request); + GEM_BUG_ON(!test_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags)); clear_bit(I915_FENCE_FLAG_ACTIVE, &request->fence.flags); + spin_unlock(&request->lock); /* Transfer back from the global per-engine timeline to per-context */ @@ -518,7 +581,7 @@ i915_request_alloc_slow(struct intel_context *ce) ring_retire_requests(ring); out: - return kmem_cache_alloc(ce->gem_context->i915->requests, GFP_KERNEL); + return kmem_cache_alloc(global.slab_requests, GFP_KERNEL); } static int add_timeline_barrier(struct i915_request *rq) @@ -539,8 +602,10 @@ struct i915_request * i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) { struct drm_i915_private *i915 = engine->i915; - struct i915_request *rq; struct intel_context *ce; + struct i915_timeline *tl; + struct i915_request *rq; + u32 seqno; int ret; lockdep_assert_held(&i915->drm.struct_mutex); @@ -556,8 +621,9 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) * ABI: Before userspace accesses the GPU (e.g. execbuffer), report * EIO if the GPU is already wedged. */ - if (i915_terminally_wedged(&i915->gpu_error)) - return ERR_PTR(-EIO); + ret = i915_terminally_wedged(i915); + if (ret) + return ERR_PTR(ret); /* * Pinning the contexts may generate requests in order to acquire @@ -569,6 +635,7 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) return ERR_CAST(ce); reserve_gt(i915); + mutex_lock(&ce->ring->timeline->mutex); /* Move our oldest request to the slab-cache (if not in use!) */ rq = list_first_entry(&ce->ring->request_list, typeof(*rq), ring_link); @@ -605,7 +672,7 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) * * Do not use kmem_cache_zalloc() here! */ - rq = kmem_cache_alloc(i915->requests, + rq = kmem_cache_alloc(global.slab_requests, GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN); if (unlikely(!rq)) { rq = i915_request_alloc_slow(ce); @@ -615,24 +682,28 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) } } - rq->rcustate = get_state_synchronize_rcu(); - INIT_LIST_HEAD(&rq->active_list); + INIT_LIST_HEAD(&rq->execute_cb); + + tl = ce->ring->timeline; + ret = i915_timeline_get_seqno(tl, rq, &seqno); + if (ret) + goto err_free; + rq->i915 = i915; rq->engine = engine; rq->gem_context = ctx; rq->hw_context = ce; rq->ring = ce->ring; - rq->timeline = ce->ring->timeline; + rq->timeline = tl; GEM_BUG_ON(rq->timeline == &engine->timeline); - rq->hwsp_seqno = rq->timeline->hwsp_seqno; + rq->hwsp_seqno = tl->hwsp_seqno; + rq->hwsp_cacheline = tl->hwsp_cacheline; + rq->rcustate = get_state_synchronize_rcu(); /* acts as smp_mb() */ spin_lock_init(&rq->lock); - dma_fence_init(&rq->fence, - &i915_fence_ops, - &rq->lock, - rq->timeline->fence_context, - timeline_get_seqno(rq->timeline)); + dma_fence_init(&rq->fence, &i915_fence_ops, &rq->lock, + tl->fence_context, seqno); /* We bump the ref for the fence chain */ i915_sw_fence_init(&i915_request_get(rq)->submit, submit_notify); @@ -640,7 +711,6 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx) i915_sched_node_init(&rq->sched); /* No zalloc, must clear what we need by hand */ - rq->global_seqno = 0; rq->file_priv = NULL; rq->batch = NULL; rq->capture_list = NULL; @@ -693,14 +763,63 @@ err_unwind: GEM_BUG_ON(!list_empty(&rq->sched.signalers_list)); GEM_BUG_ON(!list_empty(&rq->sched.waiters_list)); - kmem_cache_free(i915->requests, rq); +err_free: + kmem_cache_free(global.slab_requests, rq); err_unreserve: + mutex_unlock(&ce->ring->timeline->mutex); unreserve_gt(i915); intel_context_unpin(ce); return ERR_PTR(ret); } static int +emit_semaphore_wait(struct i915_request *to, + struct i915_request *from, + gfp_t gfp) +{ + u32 hwsp_offset; + u32 *cs; + int err; + + GEM_BUG_ON(!from->timeline->has_initial_breadcrumb); + GEM_BUG_ON(INTEL_GEN(to->i915) < 8); + + /* We need to pin the signaler's HWSP until we are finished reading. */ + err = i915_timeline_read_hwsp(from, to, &hwsp_offset); + if (err) + return err; + + /* Only submit our spinner after the signaler is running! */ + err = i915_request_await_execution(to, from, gfp); + if (err) + return err; + + cs = intel_ring_begin(to, 4); + if (IS_ERR(cs)) + return PTR_ERR(cs); + + /* + * Using greater-than-or-equal here means we have to worry + * about seqno wraparound. To side step that issue, we swap + * the timeline HWSP upon wrapping, so that everyone listening + * for the old (pre-wrap) values do not see the much smaller + * (post-wrap) values than they were expecting (and so wait + * forever). + */ + *cs++ = MI_SEMAPHORE_WAIT | + MI_SEMAPHORE_GLOBAL_GTT | + MI_SEMAPHORE_POLL | + MI_SEMAPHORE_SAD_GTE_SDD; + *cs++ = from->fence.seqno; + *cs++ = hwsp_offset; + *cs++ = 0; + + intel_ring_advance(to, cs); + to->sched.flags |= I915_SCHED_HAS_SEMAPHORE; + return 0; +} + +static int i915_request_await_request(struct i915_request *to, struct i915_request *from) { int ret; @@ -712,9 +831,7 @@ i915_request_await_request(struct i915_request *to, struct i915_request *from) return 0; if (to->engine->schedule) { - ret = i915_sched_node_add_dependency(to->i915, - &to->sched, - &from->sched); + ret = i915_sched_node_add_dependency(&to->sched, &from->sched); if (ret < 0) return ret; } @@ -723,6 +840,9 @@ i915_request_await_request(struct i915_request *to, struct i915_request *from) ret = i915_sw_fence_await_sw_fence_gfp(&to->submit, &from->submit, I915_FENCE_GFP); + } else if (intel_engine_has_semaphores(to->engine) && + to->gem_context->sched.priority >= I915_PRIORITY_NORMAL) { + ret = emit_semaphore_wait(to, from, I915_FENCE_GFP); } else { ret = i915_sw_fence_await_dma_fence(&to->submit, &from->fence, 0, @@ -889,7 +1009,7 @@ void i915_request_add(struct i915_request *request) GEM_TRACE("%s fence %llx:%lld\n", engine->name, request->fence.context, request->fence.seqno); - lockdep_assert_held(&request->i915->drm.struct_mutex); + lockdep_assert_held(&request->timeline->mutex); trace_i915_request_add(request); /* @@ -948,6 +1068,7 @@ void i915_request_add(struct i915_request *request) GEM_TRACE("marking %s as active\n", ring->timeline->name); list_add(&ring->active_link, &request->i915->gt.active_rings); } + request->i915->gt.active_engines |= request->engine->mask; request->emitted_jiffies = jiffies; /* @@ -967,6 +1088,21 @@ void i915_request_add(struct i915_request *request) struct i915_sched_attr attr = request->gem_context->sched; /* + * Boost actual workloads past semaphores! + * + * With semaphores we spin on one engine waiting for another, + * simply to reduce the latency of starting our work when + * the signaler completes. However, if there is any other + * work that we could be doing on this engine instead, that + * is better utilisation and will reduce the overall duration + * of the current work. To avoid PI boosting a semaphore + * far in the distance past over useful work, we keep a history + * of any semaphore use along our dependency chain. + */ + if (!(request->sched.flags & I915_SCHED_HAS_SEMAPHORE)) + attr.priority |= I915_PRIORITY_NOSEMAPHORE; + + /* * Boost priorities to new clients (new request flows). * * Allow interactive/synchronous clients to jump ahead of @@ -1000,6 +1136,8 @@ void i915_request_add(struct i915_request *request) */ if (prev && i915_request_completed(prev)) i915_request_retire_upto(prev); + + mutex_unlock(&request->timeline->mutex); } static unsigned long local_clock_us(unsigned int *cpu) @@ -1136,8 +1274,23 @@ long i915_request_wait(struct i915_request *rq, if (__i915_spin_request(rq, state, 5)) goto out; - if (flags & I915_WAIT_PRIORITY) + /* + * This client is about to stall waiting for the GPU. In many cases + * this is undesirable and limits the throughput of the system, as + * many clients cannot continue processing user input/output whilst + * blocked. RPS autotuning may take tens of milliseconds to respond + * to the GPU load and thus incurs additional latency for the client. + * We can circumvent that by promoting the GPU frequency to maximum + * before we sleep. This makes the GPU throttle up much more quickly + * (good for benchmarks and user experience, e.g. window animations), + * but at a cost of spending more power processing the workload + * (bad for battery). + */ + if (flags & I915_WAIT_PRIORITY) { + if (!i915_request_started(rq) && INTEL_GEN(rq->i915) >= 6) + gen6_rps_boost(rq); i915_schedule_bump_priority(rq, I915_PRIORITY_WAIT); + } wait.tsk = current; if (dma_fence_add_callback(&rq->fence, &wait.cb, request_wait_wake)) @@ -1179,11 +1332,66 @@ void i915_retire_requests(struct drm_i915_private *i915) if (!i915->gt.active_requests) return; - list_for_each_entry_safe(ring, tmp, &i915->gt.active_rings, active_link) + list_for_each_entry_safe(ring, tmp, + &i915->gt.active_rings, active_link) { + intel_ring_get(ring); /* last rq holds reference! */ ring_retire_requests(ring); + intel_ring_put(ring); + } } #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST) #include "selftests/mock_request.c" #include "selftests/i915_request.c" #endif + +static void i915_global_request_shrink(void) +{ + kmem_cache_shrink(global.slab_dependencies); + kmem_cache_shrink(global.slab_execute_cbs); + kmem_cache_shrink(global.slab_requests); +} + +static void i915_global_request_exit(void) +{ + kmem_cache_destroy(global.slab_dependencies); + kmem_cache_destroy(global.slab_execute_cbs); + kmem_cache_destroy(global.slab_requests); +} + +static struct i915_global_request global = { { + .shrink = i915_global_request_shrink, + .exit = i915_global_request_exit, +} }; + +int __init i915_global_request_init(void) +{ + global.slab_requests = KMEM_CACHE(i915_request, + SLAB_HWCACHE_ALIGN | + SLAB_RECLAIM_ACCOUNT | + SLAB_TYPESAFE_BY_RCU); + if (!global.slab_requests) + return -ENOMEM; + + global.slab_execute_cbs = KMEM_CACHE(execute_cb, + SLAB_HWCACHE_ALIGN | + SLAB_RECLAIM_ACCOUNT | + SLAB_TYPESAFE_BY_RCU); + if (!global.slab_execute_cbs) + goto err_requests; + + global.slab_dependencies = KMEM_CACHE(i915_dependency, + SLAB_HWCACHE_ALIGN | + SLAB_RECLAIM_ACCOUNT); + if (!global.slab_dependencies) + goto err_execute_cbs; + + i915_global_register(&global.base); + return 0; + +err_execute_cbs: + kmem_cache_destroy(global.slab_execute_cbs); +err_requests: + kmem_cache_destroy(global.slab_requests); + return -ENOMEM; +} |