1 files changed, 167 insertions, 126 deletions

diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c
index c7560f66dca8..4b237074f453 100644
--- a/drivers/md/bcache/writeback.c
+++ b/drivers/md/bcache/writeback.c
@@ -157,6 +157,53 @@ static void __update_writeback_rate(struct cached_dev *dc)
 	dc->writeback_rate_target = target;
 }
 
+static bool idle_counter_exceeded(struct cache_set *c)
+{
+	int counter, dev_nr;
+
+	/*
+	 * If c->idle_counter is overflow (idel for really long time),
+	 * reset as 0 and not set maximum rate this time for code
+	 * simplicity.
+	 */
+	counter = atomic_inc_return(&c->idle_counter);
+	if (counter <= 0) {
+		atomic_set(&c->idle_counter, 0);
+		return false;
+	}
+
+	dev_nr = atomic_read(&c->attached_dev_nr);
+	if (dev_nr == 0)
+		return false;
+
+	/*
+	 * c->idle_counter is increased by writeback thread of all
+	 * attached backing devices, in order to represent a rough
+	 * time period, counter should be divided by dev_nr.
+	 * Otherwise the idle time cannot be larger with more backing
+	 * device attached.
+	 * The following calculation equals to checking
+	 *	(counter / dev_nr) < (dev_nr * 6)
+	 */
+	if (counter < (dev_nr * dev_nr * 6))
+		return false;
+
+	return true;
+}
+
+/*
+ * Idle_counter is increased every time when update_writeback_rate() is
+ * called. If all backing devices attached to the same cache set have
+ * identical dc->writeback_rate_update_seconds values, it is about 6
+ * rounds of update_writeback_rate() on each backing device before
+ * c->at_max_writeback_rate is set to 1, and then max wrteback rate set
+ * to each dc->writeback_rate.rate.
+ * In order to avoid extra locking cost for counting exact dirty cached
+ * devices number, c->attached_dev_nr is used to calculate the idle
+ * throushold. It might be bigger if not all cached device are in write-
+ * back mode, but it still works well with limited extra rounds of
+ * update_writeback_rate().
+ */
 static bool set_at_max_writeback_rate(struct cache_set *c,
 				       struct cached_dev *dc)
 {
@@ -167,21 +214,8 @@ static bool set_at_max_writeback_rate(struct cache_set *c,
 	/* Don't set max writeback rate if gc is running */
 	if (!c->gc_mark_valid)
 		return false;
-	/*
-	 * Idle_counter is increased everytime when update_writeback_rate() is
-	 * called. If all backing devices attached to the same cache set have
-	 * identical dc->writeback_rate_update_seconds values, it is about 6
-	 * rounds of update_writeback_rate() on each backing device before
-	 * c->at_max_writeback_rate is set to 1, and then max wrteback rate set
-	 * to each dc->writeback_rate.rate.
-	 * In order to avoid extra locking cost for counting exact dirty cached
-	 * devices number, c->attached_dev_nr is used to calculate the idle
-	 * throushold. It might be bigger if not all cached device are in write-
-	 * back mode, but it still works well with limited extra rounds of
-	 * update_writeback_rate().
-	 */
-	if (atomic_inc_return(&c->idle_counter) <
-	    atomic_read(&c->attached_dev_nr) * 6)
+
+	if (!idle_counter_exceeded(c))
 		return false;
 
 	if (atomic_read(&c->at_max_writeback_rate) != 1)
@@ -195,13 +229,10 @@ static bool set_at_max_writeback_rate(struct cache_set *c,
 	dc->writeback_rate_change = 0;
 
 	/*
-	 * Check c->idle_counter and c->at_max_writeback_rate agagain in case
-	 * new I/O arrives during before set_at_max_writeback_rate() returns.
-	 * Then the writeback rate is set to 1, and its new value should be
-	 * decided via __update_writeback_rate().
+	 * In case new I/O arrives during before
+	 * set_at_max_writeback_rate() returns.
 	 */
-	if ((atomic_read(&c->idle_counter) <
-	     atomic_read(&c->attached_dev_nr) * 6) ||
+	if (!idle_counter_exceeded(c) ||
 	    !atomic_read(&c->at_max_writeback_rate))
 		return false;
 
@@ -235,19 +266,27 @@ static void update_writeback_rate(struct work_struct *work)
 		return;
 	}
 
-	if (atomic_read(&dc->has_dirty) && dc->writeback_percent) {
-		/*
-		 * If the whole cache set is idle, set_at_max_writeback_rate()
-		 * will set writeback rate to a max number. Then it is
-		 * unncessary to update writeback rate for an idle cache set
-		 * in maximum writeback rate number(s).
-		 */
-		if (!set_at_max_writeback_rate(c, dc)) {
-			down_read(&dc->writeback_lock);
+	/*
+	 * If the whole cache set is idle, set_at_max_writeback_rate()
+	 * will set writeback rate to a max number. Then it is
+	 * unncessary to update writeback rate for an idle cache set
+	 * in maximum writeback rate number(s).
+	 */
+	if (atomic_read(&dc->has_dirty) && dc->writeback_percent &&
+	    !set_at_max_writeback_rate(c, dc)) {
+		do {
+			if (!down_read_trylock((&dc->writeback_lock))) {
+				dc->rate_update_retry++;
+				if (dc->rate_update_retry <=
+				    BCH_WBRATE_UPDATE_MAX_SKIPS)
+					break;
+				down_read(&dc->writeback_lock);
+				dc->rate_update_retry = 0;
+			}
 			__update_writeback_rate(dc);
 			update_gc_after_writeback(c);
 			up_read(&dc->writeback_lock);
-		}
+		} while (0);
 	}
 
 
@@ -292,26 +331,26 @@ static void dirty_init(struct keybuf_key *w)
 	struct dirty_io *io = w->private;
 	struct bio *bio = &io->bio;
 
-	bio_init(bio, bio->bi_inline_vecs,
-		 DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS));
+	bio_init_inline(bio, NULL,
+		 DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS), 0);
 	if (!io->dc->writeback_percent)
-		bio_set_prio(bio, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0));
+		bio->bi_ioprio = IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0);
 
 	bio->bi_iter.bi_size	= KEY_SIZE(&w->key) << 9;
 	bio->bi_private		= w;
 	bch_bio_map(bio, NULL);
 }
 
-static void dirty_io_destructor(struct closure *cl)
+static CLOSURE_CALLBACK(dirty_io_destructor)
 {
-	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
+	closure_type(io, struct dirty_io, cl);
 
 	kfree(io);
 }
 
-static void write_dirty_finish(struct closure *cl)
+static CLOSURE_CALLBACK(write_dirty_finish)
 {
-	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
+	closure_type(io, struct dirty_io, cl);
 	struct keybuf_key *w = io->bio.bi_private;
 	struct cached_dev *dc = io->dc;
 
@@ -361,9 +400,9 @@ static void dirty_endio(struct bio *bio)
 	closure_put(&io->cl);
 }
 
-static void write_dirty(struct closure *cl)
+static CLOSURE_CALLBACK(write_dirty)
 {
-	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
+	closure_type(io, struct dirty_io, cl);
 	struct keybuf_key *w = io->bio.bi_private;
 	struct cached_dev *dc = io->dc;
 
@@ -395,7 +434,7 @@ static void write_dirty(struct closure *cl)
 	 */
 	if (KEY_DIRTY(&w->key)) {
 		dirty_init(w);
-		bio_set_op_attrs(&io->bio, REQ_OP_WRITE, 0);
+		io->bio.bi_opf = REQ_OP_WRITE;
 		io->bio.bi_iter.bi_sector = KEY_START(&w->key);
 		bio_set_dev(&io->bio, io->dc->bdev);
 		io->bio.bi_end_io	= dirty_endio;
@@ -423,9 +462,9 @@ static void read_dirty_endio(struct bio *bio)
 	dirty_endio(bio);
 }
 
-static void read_dirty_submit(struct closure *cl)
+static CLOSURE_CALLBACK(read_dirty_submit)
 {
-	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
+	closure_type(io, struct dirty_io, cl);
 
 	closure_bio_submit(io->dc->disk.c, &io->bio, cl);
 
@@ -497,9 +536,9 @@ static void read_dirty(struct cached_dev *dc)
 		for (i = 0; i < nk; i++) {
 			w = keys[i];
 
-			io = kzalloc(struct_size(io, bio.bi_inline_vecs,
-						DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS)),
-				     GFP_KERNEL);
+			io = kzalloc(sizeof(*io) + sizeof(struct bio_vec) *
+				DIV_ROUND_UP(KEY_SIZE(&w->key), PAGE_SECTORS),
+				GFP_KERNEL);
 			if (!io)
 				goto err;
 
@@ -508,7 +547,7 @@ static void read_dirty(struct cached_dev *dc)
 			io->sequence    = sequence++;
 
 			dirty_init(w);
-			bio_set_op_attrs(&io->bio, REQ_OP_READ, 0);
+			io->bio.bi_opf = REQ_OP_READ;
 			io->bio.bi_iter.bi_sector = PTR_OFFSET(&w->key, 0);
 			bio_set_dev(&io->bio, dc->disk.c->cache->bdev);
 			io->bio.bi_end_io	= read_dirty_endio;
@@ -585,10 +624,13 @@ void bcache_dev_sectors_dirty_add(struct cache_set *c, unsigned int inode,
 
 		sectors_dirty = atomic_add_return(s,
 					d->stripe_sectors_dirty + stripe);
-		if (sectors_dirty == d->stripe_size)
-			set_bit(stripe, d->full_dirty_stripes);
-		else
-			clear_bit(stripe, d->full_dirty_stripes);
+		if (sectors_dirty == d->stripe_size) {
+			if (!test_bit(stripe, d->full_dirty_stripes))
+				set_bit(stripe, d->full_dirty_stripes);
+		} else {
+			if (test_bit(stripe, d->full_dirty_stripes))
+				clear_bit(stripe, d->full_dirty_stripes);
+		}
 
 		nr_sectors -= s;
 		stripe_offset = 0;
@@ -763,8 +805,7 @@ static int bch_writeback_thread(void *arg)
 			 * may set BCH_ENABLE_AUTO_GC via sysfs, then when
 			 * BCH_DO_AUTO_GC is set, garbage collection thread
 			 * will be wake up here. After moving gc, the shrunk
-			 * btree and discarded free buckets SSD space may be
-			 * helpful for following write requests.
+			 * btree may be helpful for following write requests.
 			 */
 			if (c->gc_after_writeback ==
 			    (BCH_ENABLE_AUTO_GC|BCH_DO_AUTO_GC)) {
@@ -790,10 +831,9 @@ static int bch_writeback_thread(void *arg)
 		}
 	}
 
-	if (dc->writeback_write_wq) {
-		flush_workqueue(dc->writeback_write_wq);
+	if (dc->writeback_write_wq)
 		destroy_workqueue(dc->writeback_write_wq);
-	}
+
 	cached_dev_put(dc);
 	wait_for_kthread_stop();
 
@@ -802,13 +842,11 @@ static int bch_writeback_thread(void *arg)
 
 /* Init */
 #define INIT_KEYS_EACH_TIME	500000
-#define INIT_KEYS_SLEEP_MS	100
 
 struct sectors_dirty_init {
 	struct btree_op	op;
 	unsigned int	inode;
 	size_t		count;
-	struct bkey	start;
 };
 
 static int sectors_dirty_init_fn(struct btree_op *_op, struct btree *b,
@@ -824,11 +862,8 @@ static int sectors_dirty_init_fn(struct btree_op *_op, struct btree *b,
 					     KEY_START(k), KEY_SIZE(k));
 
 	op->count++;
-	if (atomic_read(&b->c->search_inflight) &&
-	    !(op->count % INIT_KEYS_EACH_TIME)) {
-		bkey_copy_key(&op->start, k);
-		return -EAGAIN;
-	}
+	if (!(op->count % INIT_KEYS_EACH_TIME))
+		cond_resched();
 
 	return MAP_CONTINUE;
 }
@@ -843,24 +878,26 @@ static int bch_root_node_dirty_init(struct cache_set *c,
 	bch_btree_op_init(&op.op, -1);
 	op.inode = d->id;
 	op.count = 0;
-	op.start = KEY(op.inode, 0, 0);
-
-	do {
-		ret = bcache_btree(map_keys_recurse,
-				   k,
-				   c->root,
-				   &op.op,
-				   &op.start,
-				   sectors_dirty_init_fn,
-				   0);
-		if (ret == -EAGAIN)
-			schedule_timeout_interruptible(
-				msecs_to_jiffies(INIT_KEYS_SLEEP_MS));
-		else if (ret < 0) {
-			pr_warn("sectors dirty init failed, ret=%d!\n", ret);
-			break;
-		}
-	} while (ret == -EAGAIN);
+
+	ret = bcache_btree(map_keys_recurse,
+			   k,
+			   c->root,
+			   &op.op,
+			   &KEY(op.inode, 0, 0),
+			   sectors_dirty_init_fn,
+			   0);
+	if (ret < 0)
+		pr_warn("sectors dirty init failed, ret=%d!\n", ret);
+
+	/*
+	 * The op may be added to cache_set's btree_cache_wait
+	 * in mca_cannibalize(), must ensure it is removed from
+	 * the list and release btree_cache_alloc_lock before
+	 * free op memory.
+	 * Otherwise, the btree_cache_wait will be damaged.
+	 */
+	bch_cannibalize_unlock(c);
+	finish_wait(&c->btree_cache_wait, &(&op.op)->wait);
 
 	return ret;
 }
@@ -870,15 +907,15 @@ static int bch_dirty_init_thread(void *arg)
 	struct dirty_init_thrd_info *info = arg;
 	struct bch_dirty_init_state *state = info->state;
 	struct cache_set *c = state->c;
-	struct btree_iter iter;
+	struct btree_iter_stack iter;
 	struct bkey *k, *p;
 	int cur_idx, prev_idx, skip_nr;
 
 	k = p = NULL;
-	cur_idx = prev_idx = 0;
+	prev_idx = 0;
 
-	bch_btree_iter_init(&c->root->keys, &iter, NULL);
-	k = bch_btree_iter_next_filter(&iter, &c->root->keys, bch_ptr_bad);
+	bch_btree_iter_stack_init(&c->root->keys, &iter, NULL);
+	k = bch_btree_iter_next_filter(&iter.iter, &c->root->keys, bch_ptr_bad);
 	BUG_ON(!k);
 
 	p = k;
@@ -892,7 +929,7 @@ static int bch_dirty_init_thread(void *arg)
 		skip_nr = cur_idx - prev_idx;
 
 		while (skip_nr) {
-			k = bch_btree_iter_next_filter(&iter,
+			k = bch_btree_iter_next_filter(&iter.iter,
 						       &c->root->keys,
 						       bch_ptr_bad);
 			if (k)
@@ -904,7 +941,6 @@ static int bch_dirty_init_thread(void *arg)
 				goto out;
 			}
 			skip_nr--;
-			cond_resched();
 		}
 
 		if (p) {
@@ -914,7 +950,6 @@ static int bch_dirty_init_thread(void *arg)
 
 		p = NULL;
 		prev_idx = cur_idx;
-		cond_resched();
 	}
 
 out:
@@ -941,69 +976,72 @@ static int bch_btre_dirty_init_thread_nr(void)
 void bch_sectors_dirty_init(struct bcache_device *d)
 {
 	int i;
+	struct btree *b = NULL;
 	struct bkey *k = NULL;
-	struct btree_iter iter;
+	struct btree_iter_stack iter;
 	struct sectors_dirty_init op;
 	struct cache_set *c = d->c;
-	struct bch_dirty_init_state *state;
-	char name[32];
+	struct bch_dirty_init_state state;
+
+retry_lock:
+	b = c->root;
+	rw_lock(0, b, b->level);
+	if (b != c->root) {
+		rw_unlock(0, b);
+		goto retry_lock;
+	}
 
 	/* Just count root keys if no leaf node */
 	if (c->root->level == 0) {
 		bch_btree_op_init(&op.op, -1);
 		op.inode = d->id;
 		op.count = 0;
-		op.start = KEY(op.inode, 0, 0);
 
 		for_each_key_filter(&c->root->keys,
-				    k, &iter, bch_ptr_invalid)
+				    k, &iter, bch_ptr_invalid) {
+			if (KEY_INODE(k) != op.inode)
+				continue;
 			sectors_dirty_init_fn(&op.op, c->root, k);
-		return;
-	}
+		}
 
-	state = kzalloc(sizeof(struct bch_dirty_init_state), GFP_KERNEL);
-	if (!state) {
-		pr_warn("sectors dirty init failed: cannot allocate memory\n");
+		rw_unlock(0, b);
 		return;
 	}
 
-	state->c = c;
-	state->d = d;
-	state->total_threads = bch_btre_dirty_init_thread_nr();
-	state->key_idx = 0;
-	spin_lock_init(&state->idx_lock);
-	atomic_set(&state->started, 0);
-	atomic_set(&state->enough, 0);
-	init_waitqueue_head(&state->wait);
-
-	for (i = 0; i < state->total_threads; i++) {
-		/* Fetch latest state->enough earlier */
+	memset(&state, 0, sizeof(struct bch_dirty_init_state));
+	state.c = c;
+	state.d = d;
+	state.total_threads = bch_btre_dirty_init_thread_nr();
+	state.key_idx = 0;
+	spin_lock_init(&state.idx_lock);
+	atomic_set(&state.started, 0);
+	atomic_set(&state.enough, 0);
+	init_waitqueue_head(&state.wait);
+
+	for (i = 0; i < state.total_threads; i++) {
+		/* Fetch latest state.enough earlier */
 		smp_mb__before_atomic();
-		if (atomic_read(&state->enough))
+		if (atomic_read(&state.enough))
 			break;
 
-		state->infos[i].state = state;
-		atomic_inc(&state->started);
-		snprintf(name, sizeof(name), "bch_dirty_init[%d]", i);
-
-		state->infos[i].thread =
-			kthread_run(bch_dirty_init_thread,
-				    &state->infos[i],
-				    name);
-		if (IS_ERR(state->infos[i].thread)) {
+		atomic_inc(&state.started);
+		state.infos[i].state = &state;
+		state.infos[i].thread =
+			kthread_run(bch_dirty_init_thread, &state.infos[i],
+				    "bch_dirtcnt[%d]", i);
+		if (IS_ERR(state.infos[i].thread)) {
 			pr_err("fails to run thread bch_dirty_init[%d]\n", i);
+			atomic_dec(&state.started);
 			for (--i; i >= 0; i--)
-				kthread_stop(state->infos[i].thread);
+				kthread_stop(state.infos[i].thread);
 			goto out;
 		}
 	}
 
-	wait_event_interruptible(state->wait,
-		 atomic_read(&state->started) == 0 ||
-		 test_bit(CACHE_SET_IO_DISABLE, &c->flags));
-
 out:
-	kfree(state);
+	/* Must wait for all threads to stop. */
+	wait_event(state.wait, atomic_read(&state.started) == 0);
+	rw_unlock(0, b);
 }
 
 void bch_cached_dev_writeback_init(struct cached_dev *dc)
@@ -1027,6 +1065,9 @@ void bch_cached_dev_writeback_init(struct cached_dev *dc)
 	dc->writeback_rate_fp_term_high = 1000;
 	dc->writeback_rate_i_term_inverse = 10000;
 
+	/* For dc->writeback_lock contention in update_writeback_rate() */
+	dc->rate_update_retry = 0;
+
 	WARN_ON(test_and_clear_bit(BCACHE_DEV_WB_RUNNING, &dc->disk.flags));
 	INIT_DELAYED_WORK(&dc->writeback_rate_update, update_writeback_rate);
 }
@@ -1034,7 +1075,7 @@ void bch_cached_dev_writeback_init(struct cached_dev *dc)
 int bch_cached_dev_writeback_start(struct cached_dev *dc)
 {
 	dc->writeback_write_wq = alloc_workqueue("bcache_writeback_wq",
-						WQ_MEM_RECLAIM, 0);
+						WQ_MEM_RECLAIM | WQ_PERCPU, 0);
 	if (!dc->writeback_write_wq)
 		return -ENOMEM;