[PATCH 11/46] drm/i915: Keep timeline HWSP allocated until idle across the system

[Chris Wilson <chris@chris-wilson.co.uk>]

In preparation for enabling HW semaphores, we need to keep in flight
timeline HWSP alive until its use across entire system has completed,
as any other timeline active on the GPU may still refer back to the
already retired timeline. We both have to delay recycling available
cachelines and unpinning old HWSP until the next idle point.

An easy option would be to simply keep all used HWSP until the system as
a whole was idle, i.e. we could release them all at once on parking.
However, on a busy system, we may never see a global idle point,
essentially meaning the resource will be leaked until we are forced to
do a GC pass. We already employ a fine-grained idle detection mechanism
for vma, which we can reuse here so that each cacheline can be freed
immediately after the last request using it is retired.

v3: Keep track of the activity of each cacheline.
v4: cacheline_free() on canceling the seqno tracking
v5: Finally with a testcase to exercise wraparound

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
---
 drivers/gpu/drm/i915/i915_request.c           |  30 +-
 drivers/gpu/drm/i915/i915_timeline.c          | 264 ++++++++++++++++--
 drivers/gpu/drm/i915/i915_timeline.h          |   9 +-
 .../gpu/drm/i915/selftests/i915_timeline.c    | 113 ++++++++
 4 files changed, 377 insertions(+), 39 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 174d15c9dd00..13e0388791b6 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -331,11 +331,6 @@ void i915_request_retire_upto(struct i915_request *rq)
 	} while (tmp != rq);
 }
 
-static u32 timeline_get_seqno(struct i915_timeline *tl)
-{
-	return tl->seqno += 1 + tl->has_initial_breadcrumb;
-}
-
 static void move_to_timeline(struct i915_request *request,
 			     struct i915_timeline *timeline)
 {
@@ -556,8 +551,10 @@ 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);
@@ -632,24 +629,26 @@ 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);
+
+	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->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);
@@ -710,6 +709,7 @@ i915_request_alloc(struct intel_engine_cs *engine, struct i915_gem_context *ctx)
 	GEM_BUG_ON(!list_empty(&rq->sched.signalers_list));
 	GEM_BUG_ON(!list_empty(&rq->sched.waiters_list));
 
+err_free:
 	kmem_cache_free(global.slab_requests, rq);
 err_unreserve:
 	unreserve_gt(i915);
diff --git a/drivers/gpu/drm/i915/i915_timeline.c b/drivers/gpu/drm/i915/i915_timeline.c
index b2202d2e58a2..3608e544012f 100644
--- a/drivers/gpu/drm/i915/i915_timeline.c
+++ b/drivers/gpu/drm/i915/i915_timeline.c
@@ -6,19 +6,29 @@
 
 #include "i915_drv.h"
 
-#include "i915_timeline.h"
+#include "i915_active.h"
 #include "i915_syncmap.h"
+#include "i915_timeline.h"
 
 struct i915_timeline_hwsp {
-	struct i915_vma *vma;
+	struct i915_gt_timelines *gt;
 	struct list_head free_link;
+	struct i915_vma *vma;
 	u64 free_bitmap;
 };
 
-static inline struct i915_timeline_hwsp *
-i915_timeline_hwsp(const struct i915_timeline *tl)
+struct i915_timeline_cacheline {
+	struct i915_active active;
+	struct i915_timeline_hwsp *hwsp;
+	void *vaddr;
+	unsigned int cacheline : 6;
+	unsigned int free : 1;
+};
+
+static inline struct drm_i915_private *
+hwsp_to_i915(struct i915_timeline_hwsp *hwsp)
 {
-	return tl->hwsp_ggtt->private;
+	return container_of(hwsp->gt, struct drm_i915_private, gt.timelines);
 }
 
 static struct i915_vma *__hwsp_alloc(struct drm_i915_private *i915)
@@ -71,6 +81,7 @@ hwsp_alloc(struct i915_timeline *timeline, unsigned int *cacheline)
 		vma->private = hwsp;
 		hwsp->vma = vma;
 		hwsp->free_bitmap = ~0ull;
+		hwsp->gt = gt;
 
 		spin_lock(&gt->hwsp_lock);
 		list_add(&hwsp->free_link, &gt->hwsp_free_list);
@@ -88,14 +99,9 @@ hwsp_alloc(struct i915_timeline *timeline, unsigned int *cacheline)
 	return hwsp->vma;
 }
 
-static void hwsp_free(struct i915_timeline *timeline)
+static void __idle_hwsp_free(struct i915_timeline_hwsp *hwsp, int cacheline)
 {
-	struct i915_gt_timelines *gt = &timeline->i915->gt.timelines;
-	struct i915_timeline_hwsp *hwsp;
-
-	hwsp = i915_timeline_hwsp(timeline);
-	if (!hwsp) /* leave global HWSP alone! */
-		return;
+	struct i915_gt_timelines *gt = hwsp->gt;
 
 	spin_lock(&gt->hwsp_lock);
 
@@ -103,7 +109,8 @@ static void hwsp_free(struct i915_timeline *timeline)
 	if (!hwsp->free_bitmap)
 		list_add_tail(&hwsp->free_link, &gt->hwsp_free_list);
 
-	hwsp->free_bitmap |= BIT_ULL(timeline->hwsp_offset / CACHELINE_BYTES);
+	GEM_BUG_ON(cacheline >= BITS_PER_TYPE(hwsp->free_bitmap));
+	hwsp->free_bitmap |= BIT_ULL(cacheline);
 
 	/* And if no one is left using it, give the page back to the system */
 	if (hwsp->free_bitmap == ~0ull) {
@@ -115,6 +122,78 @@ static void hwsp_free(struct i915_timeline *timeline)
 	spin_unlock(&gt->hwsp_lock);
 }
 
+static void __idle_cacheline_free(struct i915_timeline_cacheline *cl)
+{
+	GEM_BUG_ON(!i915_active_is_idle(&cl->active));
+
+	i915_gem_object_unpin_map(cl->hwsp->vma->obj);
+	i915_vma_put(cl->hwsp->vma);
+	__idle_hwsp_free(cl->hwsp, cl->cacheline);
+
+	i915_active_fini(&cl->active);
+	kfree(cl);
+}
+
+static void __cacheline_retire(struct i915_active *active)
+{
+	struct i915_timeline_cacheline *cl =
+		container_of(active, typeof(*cl), active);
+
+	i915_vma_unpin(cl->hwsp->vma);
+	if (cl->free)
+		__idle_cacheline_free(cl);
+}
+
+static struct i915_timeline_cacheline *
+cacheline_alloc(struct i915_timeline_hwsp *hwsp, unsigned int cacheline)
+{
+	struct i915_timeline_cacheline *cl;
+	void *vaddr;
+
+	GEM_BUG_ON(cacheline >= 64);
+
+	cl = kmalloc(sizeof(*cl), GFP_KERNEL);
+	if (!cl)
+		return ERR_PTR(-ENOMEM);
+
+	vaddr = i915_gem_object_pin_map(hwsp->vma->obj, I915_MAP_WB);
+	if (IS_ERR(vaddr)) {
+		kfree(cl);
+		return ERR_CAST(vaddr);
+	}
+
+	i915_vma_get(hwsp->vma);
+	cl->hwsp = hwsp;
+	cl->vaddr = vaddr;
+	cl->cacheline = cacheline;
+	cl->free = false;
+
+	i915_active_init(hwsp_to_i915(hwsp), &cl->active, __cacheline_retire);
+
+	return cl;
+}
+
+static void cacheline_acquire(struct i915_timeline_cacheline *cl)
+{
+	if (cl && i915_active_acquire(&cl->active))
+		__i915_vma_pin(cl->hwsp->vma);
+}
+
+static void cacheline_release(struct i915_timeline_cacheline *cl)
+{
+	if (cl)
+		i915_active_release(&cl->active);
+}
+
+static void cacheline_free(struct i915_timeline_cacheline *cl)
+{
+	GEM_BUG_ON(cl->free);
+	cl->free = true;
+
+	if (i915_active_is_idle(&cl->active))
+		__idle_cacheline_free(cl);
+}
+
 int i915_timeline_init(struct drm_i915_private *i915,
 		       struct i915_timeline *timeline,
 		       const char *name,
@@ -136,29 +215,40 @@ int i915_timeline_init(struct drm_i915_private *i915,
 	timeline->name = name;
 	timeline->pin_count = 0;
 	timeline->has_initial_breadcrumb = !hwsp;
+	timeline->hwsp_cacheline = NULL;
 
-	timeline->hwsp_offset = I915_GEM_HWS_SEQNO_ADDR;
 	if (!hwsp) {
+		struct i915_timeline_cacheline *cl;
 		unsigned int cacheline;
 
 		hwsp = hwsp_alloc(timeline, &cacheline);
 		if (IS_ERR(hwsp))
 			return PTR_ERR(hwsp);
 
+		cl = cacheline_alloc(hwsp->private, cacheline);
+		if (IS_ERR(cl)) {
+			__idle_hwsp_free(hwsp->private, cacheline);
+			return PTR_ERR(cl);
+		}
+
+		timeline->hwsp_cacheline = cl;
 		timeline->hwsp_offset = cacheline * CACHELINE_BYTES;
-	}
-	timeline->hwsp_ggtt = i915_vma_get(hwsp);
 
-	vaddr = i915_gem_object_pin_map(hwsp->obj, I915_MAP_WB);
-	if (IS_ERR(vaddr)) {
-		hwsp_free(timeline);
-		i915_vma_put(hwsp);
-		return PTR_ERR(vaddr);
+		vaddr = cl->vaddr;
+	} else {
+		timeline->hwsp_offset = I915_GEM_HWS_SEQNO_ADDR;
+
+		vaddr = i915_gem_object_pin_map(hwsp->obj, I915_MAP_WB);
+		if (IS_ERR(vaddr))
+			return PTR_ERR(vaddr);
 	}
 
 	timeline->hwsp_seqno =
 		memset(vaddr + timeline->hwsp_offset, 0, CACHELINE_BYTES);
 
+	timeline->hwsp_ggtt = i915_vma_get(hwsp);
+	GEM_BUG_ON(timeline->hwsp_offset >= hwsp->size);
+
 	timeline->fence_context = dma_fence_context_alloc(1);
 
 	spin_lock_init(&timeline->lock);
@@ -239,9 +329,12 @@ void i915_timeline_fini(struct i915_timeline *timeline)
 	GEM_BUG_ON(i915_active_request_isset(&timeline->barrier));
 
 	i915_syncmap_free(&timeline->sync);
-	hwsp_free(timeline);
 
-	i915_gem_object_unpin_map(timeline->hwsp_ggtt->obj);
+	if (timeline->hwsp_cacheline)
+		cacheline_free(timeline->hwsp_cacheline);
+	else
+		i915_gem_object_unpin_map(timeline->hwsp_ggtt->obj);
+
 	i915_vma_put(timeline->hwsp_ggtt);
 }
 
@@ -284,6 +377,7 @@ int i915_timeline_pin(struct i915_timeline *tl)
 		i915_ggtt_offset(tl->hwsp_ggtt) +
 		offset_in_page(tl->hwsp_offset);
 
+	cacheline_acquire(tl->hwsp_cacheline);
 	timeline_add_to_active(tl);
 
 	return 0;
@@ -293,6 +387,129 @@ int i915_timeline_pin(struct i915_timeline *tl)
 	return err;
 }
 
+static u32 timeline_advance(struct i915_timeline *tl)
+{
+	GEM_BUG_ON(!tl->pin_count);
+	GEM_BUG_ON(tl->seqno & tl->has_initial_breadcrumb);
+
+	return tl->seqno += 1 + tl->has_initial_breadcrumb;
+}
+
+static void timeline_rollback(struct i915_timeline *tl)
+{
+	tl->seqno -= 1 + tl->has_initial_breadcrumb;
+}
+
+static noinline int
+__i915_timeline_get_seqno(struct i915_timeline *tl,
+			  struct i915_request *rq,
+			  u32 *seqno)
+{
+	struct i915_timeline_cacheline *cl;
+	struct i915_vma *vma;
+	unsigned int cacheline;
+	int err;
+
+	/*
+	 * If there is an outstanding GPU reference to this cacheline,
+	 * such as it being sampled by a HW semaphore on another timeline,
+	 * we cannot wraparound our seqno value (the HW semaphore does
+	 * a strict greater-than-or-equals compare, not i915_seqno_passed).
+	 * So if the cacheline is still busy, we must detach ourselves
+	 * from it and leave it inflight alongside its users.
+	 *
+	 * However, if nobody is watching and we can guarantee that nobody
+	 * will, we could simply reuse the same cacheline.
+	 *
+	 * if (i915_active_request_is_signaled(&tl->last_request) &&
+	 *     i915_active_is_signaled(&tl->hwsp_cacheline->active))
+	 *	return 0;
+	 *
+	 * That seems unlikely for a busy timeline that needed to wrap in
+	 * the first place, so just replace the cacheline.
+	 */
+
+	vma = hwsp_alloc(tl, &cacheline);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_rollback;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL | PIN_HIGH);
+	if (err) {
+		__idle_hwsp_free(vma->private, cacheline);
+		goto err_rollback;
+	}
+
+	cl = cacheline_alloc(vma->private, cacheline);
+	if (IS_ERR(cl)) {
+		err = PTR_ERR(cl);
+		__idle_hwsp_free(vma->private, cacheline);
+		goto err_unpin;
+	}
+	GEM_BUG_ON(cl->hwsp->vma != vma);
+
+	/*
+	 * Attach the old cacheline to the current request, so that we only
+	 * free it after the current request is retired, which ensures that
+	 * all writes into the cacheline from previous requests are complete.
+	 */
+	err = i915_active_ref(&tl->hwsp_cacheline->active,
+			      tl->fence_context, rq);
+	if (err)
+		goto err_cacheline;
+
+	cacheline_release(tl->hwsp_cacheline); /* ownership now xfered to rq */
+	cacheline_free(tl->hwsp_cacheline);
+
+	i915_vma_unpin(tl->hwsp_ggtt); /* binding kept alive by old cacheline */
+	i915_vma_put(tl->hwsp_ggtt);
+
+	tl->hwsp_ggtt = i915_vma_get(vma);
+
+	tl->hwsp_offset = cacheline * CACHELINE_BYTES;
+	tl->hwsp_seqno =
+		memset(cl->vaddr + tl->hwsp_offset, 0, CACHELINE_BYTES);
+
+	tl->hwsp_offset += i915_ggtt_offset(vma);
+
+	cacheline_acquire(cl);
+	tl->hwsp_cacheline = cl;
+
+	*seqno = timeline_advance(tl);
+	GEM_BUG_ON(i915_seqno_passed(*tl->hwsp_seqno, *seqno));
+	return 0;
+
+err_cacheline:
+	cacheline_free(cl);
+err_unpin:
+	i915_vma_unpin(vma);
+err_rollback:
+	timeline_rollback(tl);
+	return err;
+}
+
+int i915_timeline_get_seqno(struct i915_timeline *tl,
+			    struct i915_request *rq,
+			    u32 *seqno)
+{
+	*seqno = timeline_advance(tl);
+
+	/* Replace the HWSP on wraparound for HW semaphores */
+	if (unlikely(!*seqno && tl->hwsp_cacheline))
+		return __i915_timeline_get_seqno(tl, rq, seqno);
+
+	return 0;
+}
+
+int i915_timeline_read_lock(struct i915_timeline *tl, struct i915_request *rq)
+{
+	GEM_BUG_ON(!tl->pin_count);
+	GEM_BUG_ON(!tl->hwsp_cacheline);
+	return i915_active_ref(&tl->hwsp_cacheline->active,
+			       rq->fence.context, rq);
+}
+
 void i915_timeline_unpin(struct i915_timeline *tl)
 {
 	GEM_BUG_ON(!tl->pin_count);
@@ -300,6 +517,7 @@ void i915_timeline_unpin(struct i915_timeline *tl)
 		return;
 
 	timeline_remove_from_active(tl);
+	cacheline_release(tl->hwsp_cacheline);
 
 	/*
 	 * Since this timeline is idle, all bariers upon which we were waiting
diff --git a/drivers/gpu/drm/i915/i915_timeline.h b/drivers/gpu/drm/i915/i915_timeline.h
index 7bec7d2e45bf..d78ec6fbc000 100644
--- a/drivers/gpu/drm/i915/i915_timeline.h
+++ b/drivers/gpu/drm/i915/i915_timeline.h
@@ -34,7 +34,7 @@
 #include "i915_utils.h"
 
 struct i915_vma;
-struct i915_timeline_hwsp;
+struct i915_timeline_cacheline;
 
 struct i915_timeline {
 	u64 fence_context;
@@ -49,6 +49,8 @@ struct i915_timeline {
 	struct i915_vma *hwsp_ggtt;
 	u32 hwsp_offset;
 
+	struct i915_timeline_cacheline *hwsp_cacheline;
+
 	bool has_initial_breadcrumb;
 
 	/**
@@ -160,6 +162,11 @@ static inline bool i915_timeline_sync_is_later(struct i915_timeline *tl,
 }
 
 int i915_timeline_pin(struct i915_timeline *tl);
+int i915_timeline_get_seqno(struct i915_timeline *tl,
+			    struct i915_request *rq,
+			    u32 *seqno);
+int i915_timeline_read_lock(struct i915_timeline *tl,
+			    struct i915_request *rq);
 void i915_timeline_unpin(struct i915_timeline *tl);
 
 void i915_timelines_init(struct drm_i915_private *i915);
diff --git a/drivers/gpu/drm/i915/selftests/i915_timeline.c b/drivers/gpu/drm/i915/selftests/i915_timeline.c
index 12ea69b1a1e5..844701759ffc 100644
--- a/drivers/gpu/drm/i915/selftests/i915_timeline.c
+++ b/drivers/gpu/drm/i915/selftests/i915_timeline.c
@@ -641,6 +641,118 @@ static int live_hwsp_alternate(void *arg)
 #undef NUM_TIMELINES
 }
 
+static int live_hwsp_wrap(void *arg)
+{
+	struct drm_i915_private *i915 = arg;
+	struct intel_engine_cs *engine;
+	struct i915_timeline *tl;
+	enum intel_engine_id id;
+	intel_wakeref_t wakeref;
+	int err = 0;
+
+	/*
+	 * Across a seqno wrap, we need to keep the old cacheline alive for
+	 * foreign GPU references.
+	 */
+
+	mutex_lock(&i915->drm.struct_mutex);
+	wakeref = intel_runtime_pm_get(i915);
+
+	tl = i915_timeline_create(i915, __func__, NULL);
+	if (IS_ERR(tl)) {
+		err = PTR_ERR(tl);
+		goto out_rpm;
+	}
+	if (!tl->has_initial_breadcrumb || !tl->hwsp_cacheline)
+		goto out_free;
+
+	err = i915_timeline_pin(tl);
+	if (err)
+		goto out_free;
+
+	for_each_engine(engine, i915, id) {
+		const u32 *hwsp_seqno[2];
+		struct i915_request *rq;
+		u32 seqno[2];
+
+		if (!intel_engine_can_store_dword(engine))
+			continue;
+
+		rq = i915_request_alloc(engine, i915->kernel_context);
+		if (IS_ERR(rq)) {
+			err = PTR_ERR(rq);
+			goto out;
+		}
+
+		tl->seqno = -4u;
+
+		err = i915_timeline_get_seqno(tl, rq, &seqno[0]);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+		pr_debug("seqno[0]:%08x, hwsp_offset:%08x\n",
+			 seqno[0], tl->hwsp_offset);
+
+		err = emit_ggtt_store_dw(rq, tl->hwsp_offset, seqno[0]);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+		hwsp_seqno[0] = tl->hwsp_seqno;
+
+		err = i915_timeline_get_seqno(tl, rq, &seqno[1]);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+		pr_debug("seqno[1]:%08x, hwsp_offset:%08x\n",
+			 seqno[1], tl->hwsp_offset);
+
+		err = emit_ggtt_store_dw(rq, tl->hwsp_offset, seqno[1]);
+		if (err) {
+			i915_request_add(rq);
+			goto out;
+		}
+		hwsp_seqno[1] = tl->hwsp_seqno;
+
+		/* With wrap should come a new hwsp */
+		GEM_BUG_ON(seqno[1] >= seqno[0]);
+		GEM_BUG_ON(hwsp_seqno[0] == hwsp_seqno[1]);
+
+		i915_request_add(rq);
+
+		if (i915_request_wait(rq, I915_WAIT_LOCKED, HZ / 5) < 0) {
+			pr_err("Wait for timeline writes timed out!\n");
+			err = -EIO;
+			goto out;
+		}
+
+		if (*hwsp_seqno[0] != seqno[0] || *hwsp_seqno[1] != seqno[1]) {
+			pr_err("Bad timeline values: found (%x, %x), expected (%x, %x)\n",
+			       *hwsp_seqno[0], *hwsp_seqno[1],
+			       seqno[0], seqno[1]);
+			err = -EINVAL;
+			goto out;
+		}
+
+		i915_retire_requests(i915); /* recycle HWSP */
+	}
+
+out:
+	if (igt_flush_test(i915, I915_WAIT_LOCKED))
+		err = -EIO;
+
+	i915_timeline_unpin(tl);
+out_free:
+	i915_timeline_put(tl);
+out_rpm:
+	intel_runtime_pm_put(i915, wakeref);
+	mutex_unlock(&i915->drm.struct_mutex);
+
+	return err;
+}
+
 static int live_hwsp_recycle(void *arg)
 {
 	struct drm_i915_private *i915 = arg;
@@ -723,6 +835,7 @@ int i915_timeline_live_selftests(struct drm_i915_private *i915)
 		SUBTEST(live_hwsp_recycle),
 		SUBTEST(live_hwsp_engine),
 		SUBTEST(live_hwsp_alternate),
+		SUBTEST(live_hwsp_wrap),
 	};
 
 	return i915_subtests(tests, i915);
-- 
2.20.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx