drm/i915: Return immediately if trylock fails for direct-reclaim

Ignore trying to shrink from i915 if we fail to acquire the struct_mutex
in the shrinker while performing direct-reclaim. The trade-off being
(much) lower latency for non-i915 clients at an increased risk of being
unable to obtain a page from direct-reclaim without hitting the
oom-notifier. The proviso being that we still keep trying to hard
obtain the lock for kswapd so that we can reap under heavy memory
pressure.

v2: Taint all mutexes taken within the shrinker with the struct_mutex
subclass as an early warning system, and drop I915_SHRINK_ACTIVE from
vmap to reduce the number of dangerous paths. We also have to drop
I915_SHRINK_ACTIVE from oom-notifier to be able to make the same claim
that ACTIVE is only used from outside context, which fits in with a
longer strategy of avoiding stalls due to scanning active during
shrinking.

The danger in using the subclass struct_mutex is that we declare
ourselves more knowledgable than lockdep and deprive ourselves of
automatic coverage. Instead, we require ourselves to mark up any mutex
taken inside the shrinker in order to detect lock-inversion, and if we
miss any we are doomed to a deadlock at the worst possible moment.

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/20190107115509.12523-1-chris@chris-wilson.co.uk

drivers/gpu/drm/i915/i915_drv.h

@@ -2899,9 +2899,9 @@ i915_gem_object_unpin_pages(struct drm_i915_gem_object *obj)
 	__i915_gem_object_unpin_pages(obj);
 }
 
-enum i915_mm_subclass { /* lockdep subclass for obj->mm.lock */
+enum i915_mm_subclass { /* lockdep subclass for obj->mm.lock/struct_mutex */
 	I915_MM_NORMAL = 0,
-	I915_MM_SHRINKER
+	I915_MM_SHRINKER /* called "recursively" from direct-reclaim-esque */
 };
 
 void __i915_gem_object_put_pages(struct drm_i915_gem_object *obj,
@@ -3187,7 +3187,8 @@ unsigned long i915_gem_shrink(struct drm_i915_private *i915,
 unsigned long i915_gem_shrink_all(struct drm_i915_private *i915);
 void i915_gem_shrinker_register(struct drm_i915_private *i915);
 void i915_gem_shrinker_unregister(struct drm_i915_private *i915);
-void i915_gem_shrinker_taints_mutex(struct mutex *mutex);
+void i915_gem_shrinker_taints_mutex(struct drm_i915_private *i915,
+				    struct mutex *mutex);
 
 /* i915_gem_tiling.c */
 static inline bool i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)

drivers/gpu/drm/i915/i915_gem_gtt.c

@@ -483,7 +483,7 @@ static void i915_address_space_init(struct i915_address_space *vm,
 	 * attempt holding the lock is immediately reported by lockdep.
 	 */
 	mutex_init(&vm->mutex);
-	i915_gem_shrinker_taints_mutex(&vm->mutex);
+	i915_gem_shrinker_taints_mutex(vm->i915, &vm->mutex);
 
 	GEM_BUG_ON(!vm->total);
 	drm_mm_init(&vm->mm, 0, vm->total);
@@ -2245,7 +2245,8 @@ int i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj,
 				     DMA_ATTR_NO_WARN))
 			return 0;
 
-		/* If the DMA remap fails, one cause can be that we have
+		/*
+		 * If the DMA remap fails, one cause can be that we have
 		 * too many objects pinned in a small remapping table,
 		 * such as swiotlb. Incrementally purge all other objects and
 		 * try again - if there are no more pages to remove from
@@ -2255,8 +2256,7 @@ int i915_gem_gtt_prepare_pages(struct drm_i915_gem_object *obj,
 	} while (i915_gem_shrink(to_i915(obj->base.dev),
 				 obj->base.size >> PAGE_SHIFT, NULL,
 				 I915_SHRINK_BOUND |
-				 I915_SHRINK_UNBOUND |
-				 I915_SHRINK_ACTIVE));
+				 I915_SHRINK_UNBOUND));
 
 	return -ENOSPC;
 }

drivers/gpu/drm/i915/i915_gem_shrinker.c

@@ -36,7 +36,9 @@
 #include "i915_drv.h"
 #include "i915_trace.h"
 
-static bool shrinker_lock(struct drm_i915_private *i915, bool *unlock)
+static bool shrinker_lock(struct drm_i915_private *i915,
+			  unsigned int flags,
+			  bool *unlock)
 {
 	switch (mutex_trylock_recursive(&i915->drm.struct_mutex)) {
 	case MUTEX_TRYLOCK_RECURSIVE:
@@ -45,15 +47,11 @@ static bool shrinker_lock(struct drm_i915_private *i915, bool *unlock)
 
 	case MUTEX_TRYLOCK_FAILED:
 		*unlock = false;
-		preempt_disable();
-		do {
-			cpu_relax();
-			if (mutex_trylock(&i915->drm.struct_mutex)) {
-				*unlock = true;
-				break;
-			}
-		} while (!need_resched());
-		preempt_enable();
+		if (flags & I915_SHRINK_ACTIVE) {
+			mutex_lock_nested(&i915->drm.struct_mutex,
+					  I915_MM_SHRINKER);
+			*unlock = true;
+		}
 		return *unlock;
 
 	case MUTEX_TRYLOCK_SUCCESS:
@@ -160,7 +158,7 @@ i915_gem_shrink(struct drm_i915_private *i915,
 	unsigned long scanned = 0;
 	bool unlock;
 
-	if (!shrinker_lock(i915, &unlock))
+	if (!shrinker_lock(i915, flags, &unlock))
 		return 0;
 
 	/*
@@ -357,7 +355,7 @@ i915_gem_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc)
 
 	sc->nr_scanned = 0;
 
-	if (!shrinker_lock(i915, &unlock))
+	if (!shrinker_lock(i915, 0, &unlock))
 		return SHRINK_STOP;
 
 	freed = i915_gem_shrink(i915,
@@ -397,7 +395,7 @@ shrinker_lock_uninterruptible(struct drm_i915_private *i915, bool *unlock,
 	do {
 		if (i915_gem_wait_for_idle(i915,
 					   0, MAX_SCHEDULE_TIMEOUT) == 0 &&
-		    shrinker_lock(i915, unlock))
+		    shrinker_lock(i915, 0, unlock))
 			break;
 
 		schedule_timeout_killable(1);
@@ -421,7 +419,11 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	struct drm_i915_gem_object *obj;
 	unsigned long unevictable, bound, unbound, freed_pages;
 
-	freed_pages = i915_gem_shrink_all(i915);
+	intel_runtime_pm_get(i915);
+	freed_pages = i915_gem_shrink(i915, -1UL, NULL,
+				      I915_SHRINK_BOUND |
+				      I915_SHRINK_UNBOUND);
+	intel_runtime_pm_put(i915);
 
 	/* Because we may be allocating inside our own driver, we cannot
 	 * assert that there are no objects with pinned pages that are not
@@ -447,10 +449,6 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 		pr_info("Purging GPU memory, %lu pages freed, "
 			"%lu pages still pinned.\n",
 			freed_pages, unevictable);
-	if (unbound || bound)
-		pr_err("%lu and %lu pages still available in the "
-		       "bound and unbound GPU page lists.\n",
-		       bound, unbound);
 
 	*(unsigned long *)ptr += freed_pages;
 	return NOTIFY_DONE;
@@ -480,7 +478,6 @@ i915_gem_shrinker_vmap(struct notifier_block *nb, unsigned long event, void *ptr
 	freed_pages += i915_gem_shrink(i915, -1UL, NULL,
 				       I915_SHRINK_BOUND |
 				       I915_SHRINK_UNBOUND |
-				       I915_SHRINK_ACTIVE |
 				       I915_SHRINK_VMAPS);
 	intel_runtime_pm_put(i915);
 
@@ -533,13 +530,40 @@ void i915_gem_shrinker_unregister(struct drm_i915_private *i915)
 	unregister_shrinker(&i915->mm.shrinker);
 }
 
-void i915_gem_shrinker_taints_mutex(struct mutex *mutex)
+void i915_gem_shrinker_taints_mutex(struct drm_i915_private *i915,
+				    struct mutex *mutex)
 {
+	bool unlock = false;
+
 	if (!IS_ENABLED(CONFIG_LOCKDEP))
 		return;
 
+	if (!lockdep_is_held_type(&i915->drm.struct_mutex, -1)) {
+		mutex_acquire(&i915->drm.struct_mutex.dep_map,
+			      I915_MM_NORMAL, 0, _RET_IP_);
+		unlock = true;
+	}
+
 	fs_reclaim_acquire(GFP_KERNEL);
-	mutex_lock(mutex);
-	mutex_unlock(mutex);
+
+	/*
+	 * As we invariably rely on the struct_mutex within the shrinker,
+	 * but have a complicated recursion dance, taint all the mutexes used
+	 * within the shrinker with the struct_mutex. For completeness, we
+	 * taint with all subclass of struct_mutex, even though we should
+	 * only need tainting by I915_MM_NORMAL to catch possible ABBA
+	 * deadlocks from using struct_mutex inside @mutex.
+	 */
+	mutex_acquire(&i915->drm.struct_mutex.dep_map,
+		      I915_MM_SHRINKER, 0, _RET_IP_);
+
+	mutex_acquire(&mutex->dep_map, 0, 0, _RET_IP_);
+	mutex_release(&mutex->dep_map, 0, _RET_IP_);
+
+	mutex_release(&i915->drm.struct_mutex.dep_map, 0, _RET_IP_);
+
 	fs_reclaim_release(GFP_KERNEL);
+
+	if (unlock)
+		mutex_release(&i915->drm.struct_mutex.dep_map, 0, _RET_IP_);
 }