sub directory gt/ is at linux 5.6.19.

Author: dettus

sys/external/bsd/drm2/dist/drm/i915/gt/gen8_ppgtt.c

@@ -30,6 +30,30 @@ static u64 gen8_pde_encode(const dma_addr_t addr,
 	return pde;
 }
 
+static u64 gen8_pte_encode(dma_addr_t addr,
+			   enum i915_cache_level level,
+			   u32 flags)
+{
+	gen8_pte_t pte = addr | _PAGE_PRESENT | _PAGE_RW;
+
+	if (unlikely(flags & PTE_READ_ONLY))
+		pte &= ~_PAGE_RW;
+
+	switch (level) {
+	case I915_CACHE_NONE:
+		pte |= PPAT_UNCACHED;
+		break;
+	case I915_CACHE_WT:
+		pte |= PPAT_DISPLAY_ELLC;
+		break;
+	default:
+		pte |= PPAT_CACHED;
+		break;
+	}
+
+	return pte;
+}
+
 static void gen8_ppgtt_notify_vgt(struct i915_ppgtt *ppgtt, bool create)
 {
 	struct drm_i915_private *i915 = ppgtt->vm.i915;
@@ -805,6 +829,8 @@ struct i915_ppgtt *gen8_ppgtt_create(struct intel_gt *gt)
 	ppgtt->vm.allocate_va_range = gen8_ppgtt_alloc;
 	ppgtt->vm.clear_range = gen8_ppgtt_clear;
 
+	ppgtt->vm.pte_encode = gen8_pte_encode;
+
 	if (intel_vgpu_active(gt->i915))
 		gen8_ppgtt_notify_vgt(ppgtt, true);
 

sys/external/bsd/drm2/dist/drm/i915/gt/intel_engine.h

@@ -357,13 +357,4 @@ intel_engine_has_preempt_reset(const struct intel_engine_cs *engine)
 	return intel_engine_has_preemption(engine);
 }
 
-static inline bool
-intel_engine_has_timeslices(const struct intel_engine_cs *engine)
-{
-	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
-		return false;
-
-	return intel_engine_has_semaphores(engine);
-}
-
 #endif /* _INTEL_RINGBUFFER_H_ */

sys/external/bsd/drm2/dist/drm/i915/gt/intel_engine_cs.c

@@ -305,11 +305,11 @@ static int intel_engine_setup(struct intel_gt *gt, enum intel_engine_id id)
 	engine->id = id;
 	engine->legacy_idx = INVALID_ENGINE;
 	engine->mask = BIT(id);
-	engine->i915 = gt->i915;
+	engine->i915 = i915;
 	engine->gt = gt;
 	engine->uncore = gt->uncore;
 	engine->hw_id = engine->guc_id = info->hw_id;
-	engine->mmio_base = __engine_mmio_base(gt->i915, info->mmio_bases);
+	engine->mmio_base = __engine_mmio_base(i915, info->mmio_bases);
 
 	engine->class = info->class;
 	engine->instance = info->instance;
@@ -324,11 +324,15 @@ static int intel_engine_setup(struct intel_gt *gt, enum intel_engine_id id)
 	engine->props.timeslice_duration_ms =
 		CONFIG_DRM_I915_TIMESLICE_DURATION;
 
+	/* Override to uninterruptible for OpenCL workloads. */
+	if (INTEL_GEN(i915) == 12 && engine->class == RENDER_CLASS)
+		engine->props.preempt_timeout_ms = 0;
+
 	engine->context_size = intel_engine_context_size(gt, engine->class);
 	if (WARN_ON(engine->context_size > BIT(20)))
 		engine->context_size = 0;
 	if (engine->context_size)
-		DRIVER_CAPS(gt->i915)->has_logical_contexts = true;
+		DRIVER_CAPS(i915)->has_logical_contexts = true;
 
 	/* Nothing to do here, execute in order of dependencies */
 	engine->schedule = NULL;
@@ -344,7 +348,7 @@ static int intel_engine_setup(struct intel_gt *gt, enum intel_engine_id id)
 	gt->engine_class[info->class][info->instance] = engine;
 	gt->engine[id] = engine;
 
-	gt->i915->engine[id] = engine;
+	i915->engine[id] = engine;
 
 	return 0;
 }

sys/external/bsd/drm2/dist/drm/i915/gt/intel_engine_types.h

@@ -488,10 +488,11 @@ struct intel_engine_cs {
 #define I915_ENGINE_SUPPORTS_STATS   BIT(1)
 #define I915_ENGINE_HAS_PREEMPTION   BIT(2)
 #define I915_ENGINE_HAS_SEMAPHORES   BIT(3)
-#define I915_ENGINE_NEEDS_BREADCRUMB_TASKLET BIT(4)
-#define I915_ENGINE_IS_VIRTUAL       BIT(5)
-#define I915_ENGINE_HAS_RELATIVE_MMIO BIT(6)
-#define I915_ENGINE_REQUIRES_CMD_PARSER BIT(7)
+#define I915_ENGINE_HAS_TIMESLICES   BIT(4)
+#define I915_ENGINE_NEEDS_BREADCRUMB_TASKLET BIT(5)
+#define I915_ENGINE_IS_VIRTUAL       BIT(6)
+#define I915_ENGINE_HAS_RELATIVE_MMIO BIT(7)
+#define I915_ENGINE_REQUIRES_CMD_PARSER BIT(8)
 	unsigned int flags;
 
 	/*
@@ -588,6 +589,15 @@ intel_engine_has_semaphores(const struct intel_engine_cs *engine)
 	return engine->flags & I915_ENGINE_HAS_SEMAPHORES;
 }
 
+static inline bool
+intel_engine_has_timeslices(const struct intel_engine_cs *engine)
+{
+	if (!IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
+		return false;
+
+	return engine->flags & I915_ENGINE_HAS_TIMESLICES;
+}
+
 static inline bool
 intel_engine_needs_breadcrumb_tasklet(const struct intel_engine_cs *engine)
 {

sys/external/bsd/drm2/dist/drm/i915/gt/intel_ggtt.c

@@ -193,6 +193,13 @@ static void gmch_ggtt_invalidate(struct i915_ggtt *ggtt)
 	intel_gtt_chipset_flush();
 }
 
+static u64 gen8_ggtt_pte_encode(dma_addr_t addr,
+				enum i915_cache_level level,
+				u32 flags)
+{
+	return addr | _PAGE_PRESENT;
+}
+
 #ifdef __NetBSD__
 static inline void
 gen8_set_pte(bus_space_tag_t bst, bus_space_handle_t bsh, unsigned i,
@@ -228,9 +235,9 @@ static void gen8_ggtt_insert_page(struct i915_address_space *vm,
 
 #ifdef __NetBSD__
 	gen8_set_pte(ggtt->gsmt, ggtt->gsmh, offset / I915_GTT_PAGE_SIZE,
-	    gen8_pte_encode(addr, level, 0));
+	    gen8_ggtt_pte_encode(addr, level, 0));
 #else
-	gen8_set_pte(pte, gen8_pte_encode(addr, level, 0));
+	gen8_set_pte(pte, gen8_ggtt_pte_encode(addr, level, 0));
 #endif
 
 	ggtt->invalidate(ggtt);
@@ -250,7 +257,8 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
 	struct sgt_iter sgt_iter;
 	gen8_pte_t __iomem *gtt_entries;
 #endif
-	const gen8_pte_t pte_encode = gen8_pte_encode(0, level, 0);
+	gen8_pte_t __iomem *gte;
+	gen8_pte_t __iomem *end;
 	dma_addr_t addr;
 
 	/*
@@ -274,11 +282,17 @@ static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
 		KASSERT(len == 0);
 	}
 #else
-	gtt_entries = (gen8_pte_t __iomem *)ggtt->gsm;
-	gtt_entries += vma->node.start / I915_GTT_PAGE_SIZE;
-	for_each_sgt_daddr(addr, sgt_iter, vma->pages)
-		gen8_set_pte(gtt_entries++, pte_encode | addr);
+	gte = (gen8_pte_t __iomem *)ggtt->gsm;
+	gte += vma->node.start / I915_GTT_PAGE_SIZE;
+	end = gte + vma->node.size / I915_GTT_PAGE_SIZE;
+
+	for_each_sgt_daddr(addr, iter, vma->pages)
+		gen8_set_pte(gte++, pte_encode | addr);
+	GEM_BUG_ON(gte > end);
 #endif
+	/* Fill the allocated but "unused" space beyond the end of the buffer */
+	while (gte < end)
+		gen8_set_pte(gte++, vm->scratch[0].encode);
 
 	/*
 	 * We want to flush the TLBs only after we're certain all the PTE
@@ -329,8 +343,8 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
 	unsigned seg;
 	unsigned pgno;
 #else
-	gen6_pte_t __iomem *entries = (gen6_pte_t __iomem *)ggtt->gsm;
-	unsigned int i = vma->node.start / I915_GTT_PAGE_SIZE;
+	gen6_pte_t __iomem *gte;
+	gen6_pte_t __iomem *end;
 	struct sgt_iter iter;
 #endif
 	dma_addr_t addr;
@@ -355,8 +369,17 @@ static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
 		/* XXX KASSERT(pgno <= ...)?  */
 	}
 #else
+	gte = (gen6_pte_t __iomem *)ggtt->gsm;
+	gte += vma->node.start / I915_GTT_PAGE_SIZE;
+	end = gte + vma->node.size / I915_GTT_PAGE_SIZE;
+
 	for_each_sgt_daddr(addr, iter, vma->pages)
-		iowrite32(vm->pte_encode(addr, level, flags), &entries[i++]);
+		iowrite32(vm->pte_encode(addr, level, flags), gte++);
+	GEM_BUG_ON(gte > end);
+
+	/* Fill the allocated but "unused" space beyond the end of the buffer */
+	while (gte < end)
+		iowrite32(vm->scratch[0].encode, gte++);
 #endif
 
 	/*
@@ -1064,7 +1087,7 @@ static int gen8_gmch_probe(struct i915_ggtt *ggtt)
 	ggtt->vm.vma_ops.set_pages   = ggtt_set_pages;
 	ggtt->vm.vma_ops.clear_pages = clear_pages;
 
-	ggtt->vm.pte_encode = gen8_pte_encode;
+	ggtt->vm.pte_encode = gen8_ggtt_pte_encode;
 
 	setup_private_pat(ggtt->vm.gt->uncore);
 

sys/external/bsd/drm2/dist/drm/i915/gt/intel_gtt.c

@@ -575,30 +575,6 @@ void gtt_write_workarounds(struct intel_gt *gt)
 	}
 }
 
-u64 gen8_pte_encode(dma_addr_t addr,
-		    enum i915_cache_level level,
-		    u32 flags)
-{
-	gen8_pte_t pte = addr | _PAGE_PRESENT | _PAGE_RW;
-
-	if (unlikely(flags & PTE_READ_ONLY))
-		pte &= ~_PAGE_RW;
-
-	switch (level) {
-	case I915_CACHE_NONE:
-		pte |= PPAT_UNCACHED;
-		break;
-	case I915_CACHE_WT:
-		pte |= PPAT_DISPLAY_ELLC;
-		break;
-	default:
-		pte |= PPAT_CACHED;
-		break;
-	}
-
-	return pte;
-}
-
 static void tgl_setup_private_ppat(struct intel_uncore *uncore)
 {
 	/* TGL doesn't support LLC or AGE settings */

sys/external/bsd/drm2/dist/drm/i915/gt/intel_lrc.c

@@ -1699,6 +1699,9 @@ static void defer_request(struct i915_request *rq, struct list_head * const pl)
 			struct i915_request *w =
 				container_of(p->waiter, typeof(*w), sched);
 
+			if (p->flags & I915_DEPENDENCY_WEAK)
+				continue;
+
 			/* Leave semaphores spinning on the other engines */
 			if (w->engine != rq->engine)
 				continue;
@@ -4286,8 +4289,11 @@ void intel_execlists_set_default_submission(struct intel_engine_cs *engine)
 	engine->flags |= I915_ENGINE_SUPPORTS_STATS;
 	if (!intel_vgpu_active(engine->i915)) {
 		engine->flags |= I915_ENGINE_HAS_SEMAPHORES;
-		if (HAS_LOGICAL_RING_PREEMPTION(engine->i915))
+		if (HAS_LOGICAL_RING_PREEMPTION(engine->i915)) {
 			engine->flags |= I915_ENGINE_HAS_PREEMPTION;
+			if (IS_ACTIVE(CONFIG_DRM_I915_TIMESLICE_DURATION))
+				engine->flags |= I915_ENGINE_HAS_TIMESLICES;
+		}
 	}
 
 	if (INTEL_GEN(engine->i915) >= 12)

sys/external/bsd/drm2/dist/drm/i915/gt/intel_rps.c

@@ -94,7 +94,8 @@ static void rps_enable_interrupts(struct intel_rps *rps)
 	gen6_gt_pm_enable_irq(gt, rps->pm_events);
 	spin_unlock_irq(&gt->irq_lock);
 
-	set(gt->uncore, GEN6_PMINTRMSK, rps_pm_mask(rps, rps->cur_freq));
+	intel_uncore_write(gt->uncore,
+			   GEN6_PMINTRMSK, rps_pm_mask(rps, rps->last_freq));
 }
 
 static void gen6_rps_reset_interrupts(struct intel_rps *rps)
@@ -128,7 +129,8 @@ static void rps_disable_interrupts(struct intel_rps *rps)
 
 	rps->pm_events = 0;
 
-	set(gt->uncore, GEN6_PMINTRMSK, rps_pm_sanitize_mask(rps, ~0u));
+	intel_uncore_write(gt->uncore,
+			   GEN6_PMINTRMSK, rps_pm_sanitize_mask(rps, ~0u));
 
 	spin_lock_irq(&gt->irq_lock);
 	gen6_gt_pm_disable_irq(gt, GEN6_PM_RPS_EVENTS);
@@ -774,6 +776,19 @@ void intel_rps_park(struct intel_rps *rps)
 	intel_uncore_forcewake_get(rps_to_uncore(rps), FORCEWAKE_MEDIA);
 	rps_set(rps, rps->idle_freq, false);
 	intel_uncore_forcewake_put(rps_to_uncore(rps), FORCEWAKE_MEDIA);
+
+	/*
+	 * Since we will try and restart from the previously requested
+	 * frequency on unparking, treat this idle point as a downclock
+	 * interrupt and reduce the frequency for resume. If we park/unpark
+	 * more frequently than the rps worker can run, we will not respond
+	 * to any EI and never see a change in frequency.
+	 *
+	 * (Note we accommodate Cherryview's limitation of only using an
+	 * even bin by applying it to all.)
+	 */
+	rps->cur_freq =
+		max_t(int, round_down(rps->cur_freq - 1, 2), rps->min_freq);
 }
 
 void intel_rps_boost(struct i915_request *rq)

sys/external/bsd/drm2/dist/drm/i915/gt/intel_timeline.c

@@ -524,6 +524,8 @@ int intel_timeline_read_hwsp(struct i915_request *from,
 
 	rcu_read_lock();
 	cl = rcu_dereference(from->hwsp_cacheline);
+	if (i915_request_completed(from)) /* confirm cacheline is valid */
+		goto unlock;
 	if (unlikely(!i915_active_acquire_if_busy(&cl->active)))
 		goto unlock; /* seqno wrapped and completed! */
 	if (unlikely(i915_request_completed(from)))