1 files changed, 38 insertions, 18 deletions

diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index ba434d9ac720..b12186c87f52 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -34,6 +34,7 @@
 #include <linux/random.h>
 #include <linux/rcupdate.h>
 #include <linux/sched/clock.h>
+#include <linux/sched/signal.h>
 #include <linux/rculist.h>
 #include <linux/delay.h>
 #include <trace/events/bcache.h>
@@ -543,6 +544,11 @@ static void bch_btree_leaf_dirty(struct btree *b, atomic_t *journal_ref)
 
 	set_btree_node_dirty(b);
 
+	/*
+	 * w->journal is always the oldest journal pin of all bkeys
+	 * in the leaf node, to make sure the oldest jset seq won't
+	 * be increased before this btree node is flushed.
+	 */
 	if (journal_ref) {
 		if (w->journal &&
 		    journal_pin_cmp(b->c, w->journal, journal_ref)) {
@@ -723,38 +729,38 @@ static unsigned long bch_mca_scan(struct shrinker *shrink,
 	 * IO can always make forward progress:
 	 */
 	nr /= c->btree_pages;
+	if (nr == 0)
+		nr = 1;
 	nr = min_t(unsigned long, nr, mca_can_free(c));
 
 	i = 0;
 	btree_cache_used = c->btree_cache_used;
-	list_for_each_entry_safe(b, t, &c->btree_cache_freeable, list) {
+	list_for_each_entry_safe_reverse(b, t, &c->btree_cache_freeable, list) {
 		if (nr <= 0)
 			goto out;
 
-		if (++i > 3 &&
-		    !mca_reap(b, 0, false)) {
+		if (!mca_reap(b, 0, false)) {
 			mca_data_free(b);
 			rw_unlock(true, b);
 			freed++;
 		}
 		nr--;
+		i++;
 	}
 
-	for (;  (nr--) && i < btree_cache_used; i++) {
-		if (list_empty(&c->btree_cache))
+	list_for_each_entry_safe_reverse(b, t, &c->btree_cache, list) {
+		if (nr <= 0 || i >= btree_cache_used)
 			goto out;
 
-		b = list_first_entry(&c->btree_cache, struct btree, list);
-		list_rotate_left(&c->btree_cache);
-
-		if (!b->accessed &&
-		    !mca_reap(b, 0, false)) {
+		if (!mca_reap(b, 0, false)) {
 			mca_bucket_free(b);
 			mca_data_free(b);
 			rw_unlock(true, b);
 			freed++;
-		} else
-			b->accessed = 0;
+		}
+
+		nr--;
+		i++;
 	}
 out:
 	mutex_unlock(&c->bucket_lock);
@@ -884,15 +890,17 @@ out:
 
 static int mca_cannibalize_lock(struct cache_set *c, struct btree_op *op)
 {
-	struct task_struct *old;
-
-	old = cmpxchg(&c->btree_cache_alloc_lock, NULL, current);
-	if (old && old != current) {
+	spin_lock(&c->btree_cannibalize_lock);
+	if (likely(c->btree_cache_alloc_lock == NULL)) {
+		c->btree_cache_alloc_lock = current;
+	} else if (c->btree_cache_alloc_lock != current) {
 		if (op)
 			prepare_to_wait(&c->btree_cache_wait, &op->wait,
 					TASK_UNINTERRUPTIBLE);
+		spin_unlock(&c->btree_cannibalize_lock);
 		return -EINTR;
 	}
+	spin_unlock(&c->btree_cannibalize_lock);
 
 	return 0;
 }
@@ -927,10 +935,12 @@ static struct btree *mca_cannibalize(struct cache_set *c, struct btree_op *op,
  */
 static void bch_cannibalize_unlock(struct cache_set *c)
 {
+	spin_lock(&c->btree_cannibalize_lock);
 	if (c->btree_cache_alloc_lock == current) {
 		c->btree_cache_alloc_lock = NULL;
 		wake_up(&c->btree_cache_wait);
 	}
+	spin_unlock(&c->btree_cannibalize_lock);
 }
 
 static struct btree *mca_alloc(struct cache_set *c, struct btree_op *op,
@@ -1058,7 +1068,6 @@ retry:
 	BUG_ON(!b->written);
 
 	b->parent = parent;
-	b->accessed = 1;
 
 	for (; i <= b->keys.nsets && b->keys.set[i].size; i++) {
 		prefetch(b->keys.set[i].tree);
@@ -1149,7 +1158,6 @@ retry:
 		goto retry;
 	}
 
-	b->accessed = 1;
 	b->parent = parent;
 	bch_bset_init_next(&b->keys, b->keys.set->data, bset_magic(&b->c->sb));
 
@@ -1906,6 +1914,18 @@ static int bch_gc_thread(void *arg)
 
 int bch_gc_thread_start(struct cache_set *c)
 {
+	/*
+	 * In case previous btree check operation occupies too many
+	 * system memory for bcache btree node cache, and the
+	 * registering process is selected by OOM killer. Here just
+	 * ignore the SIGKILL sent by OOM killer if there is, to
+	 * avoid kthread_run() being failed by pending signals. The
+	 * bcache registering process will exit after the registration
+	 * done.
+	 */
+	if (signal_pending(current))
+		flush_signals(current);
+
 	c->gc_thread = kthread_run(bch_gc_thread, c, "bcache_gc");
 	return PTR_ERR_OR_ZERO(c->gc_thread);
 }