Merge branch 'for-3.11/drivers' of git://git.kernel.dk/linux-block

Pull block IO driver bits from Jens Axboe:
 "As I mentioned in the core block pull request, due to real life
  circumstances the driver pull request would be late.  Now it looks
  like -rc2 late...  On the plus side, apart form the rsxx update, these
  are all things that I could argue could go in later in the cycle as
  they are fixes and not features.  So even though things are late, it's
  not ALL bad.

  The pull request contains:

   - Updates to bcache, all bug fixes, from Kent.

   - A pile of drbd bug fixes (no big features this time!).

   - xen blk front/back fixes.

   - rsxx driver updates, some of them deferred form 3.10.  So should be
     well cooked by now"

* 'for-3.11/drivers' of git://git.kernel.dk/linux-block: (63 commits)
  bcache: Allocation kthread fixes
  bcache: Fix GC_SECTORS_USED() calculation
  bcache: Journal replay fix
  bcache: Shutdown fix
  bcache: Fix a sysfs splat on shutdown
  bcache: Advertise that flushes are supported
  bcache: check for allocation failures
  bcache: Fix a dumb race
  bcache: Use standard utility code
  bcache: Update email address
  bcache: Delete fuzz tester
  bcache: Document shrinker reserve better
  bcache: FUA fixes
  drbd: Allow online change of al-stripes and al-stripe-size
  drbd: Constants should be UPPERCASE
  drbd: Ignore the exit code of a fence-peer handler if it returns too late
  drbd: Fix rcu_read_lock balance on error path
  drbd: fix error return code in drbd_init()
  drbd: Do not sleep inside rcu
  bcache: Refresh usage docs
  ...

21 files changed, 869 insertions, 801 deletions

diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c
index 048f2947e08b..e45f5575fd4d 100644
--- a/drivers/md/bcache/alloc.c
+++ b/drivers/md/bcache/alloc.c
@@ -63,7 +63,10 @@
 #include "bcache.h"
 #include "btree.h"
 
+#include <linux/freezer.h>
+#include <linux/kthread.h>
 #include <linux/random.h>
+#include <trace/events/bcache.h>
 
 #define MAX_IN_FLIGHT_DISCARDS		8U
 
@@ -151,7 +154,7 @@ static void discard_finish(struct work_struct *w)
 	mutex_unlock(&ca->set->bucket_lock);
 
 	closure_wake_up(&ca->set->bucket_wait);
-	wake_up(&ca->set->alloc_wait);
+	wake_up_process(ca->alloc_thread);
 
 	closure_put(&ca->set->cl);
 }
@@ -350,38 +353,30 @@ static void invalidate_buckets(struct cache *ca)
 		break;
 	}
 
-	pr_debug("free %zu/%zu free_inc %zu/%zu unused %zu/%zu",
-		 fifo_used(&ca->free), ca->free.size,
-		 fifo_used(&ca->free_inc), ca->free_inc.size,
-		 fifo_used(&ca->unused), ca->unused.size);
+	trace_bcache_alloc_invalidate(ca);
 }
 
 #define allocator_wait(ca, cond)					\
 do {									\
-	DEFINE_WAIT(__wait);						\
-									\
 	while (1) {							\
-		prepare_to_wait(&ca->set->alloc_wait,			\
-				&__wait, TASK_INTERRUPTIBLE);		\
+		set_current_state(TASK_INTERRUPTIBLE);			\
 		if (cond)						\
 			break;						\
 									\
 		mutex_unlock(&(ca)->set->bucket_lock);			\
-		if (test_bit(CACHE_SET_STOPPING_2, &ca->set->flags)) {	\
-			finish_wait(&ca->set->alloc_wait, &__wait);	\
-			closure_return(cl);				\
-		}							\
+		if (kthread_should_stop())				\
+			return 0;					\
 									\
+		try_to_freeze();					\
 		schedule();						\
 		mutex_lock(&(ca)->set->bucket_lock);			\
 	}								\
-									\
-	finish_wait(&ca->set->alloc_wait, &__wait);			\
+	__set_current_state(TASK_RUNNING);				\
 } while (0)
 
-void bch_allocator_thread(struct closure *cl)
+static int bch_allocator_thread(void *arg)
 {
-	struct cache *ca = container_of(cl, struct cache, alloc);
+	struct cache *ca = arg;
 
 	mutex_lock(&ca->set->bucket_lock);
 
@@ -442,7 +437,7 @@ long bch_bucket_alloc(struct cache *ca, unsigned watermark, struct closure *cl)
 {
 	long r = -1;
 again:
-	wake_up(&ca->set->alloc_wait);
+	wake_up_process(ca->alloc_thread);
 
 	if (fifo_used(&ca->free) > ca->watermark[watermark] &&
 	    fifo_pop(&ca->free, r)) {
@@ -476,9 +471,7 @@ again:
 		return r;
 	}
 
-	pr_debug("alloc failure: blocked %i free %zu free_inc %zu unused %zu",
-		 atomic_read(&ca->set->prio_blocked), fifo_used(&ca->free),
-		 fifo_used(&ca->free_inc), fifo_used(&ca->unused));
+	trace_bcache_alloc_fail(ca);
 
 	if (cl) {
 		closure_wait(&ca->set->bucket_wait, cl);
@@ -552,6 +545,17 @@ int bch_bucket_alloc_set(struct cache_set *c, unsigned watermark,
 
 /* Init */
 
+int bch_cache_allocator_start(struct cache *ca)
+{
+	struct task_struct *k = kthread_run(bch_allocator_thread,
+					    ca, "bcache_allocator");
+	if (IS_ERR(k))
+		return PTR_ERR(k);
+
+	ca->alloc_thread = k;
+	return 0;
+}
+
 void bch_cache_allocator_exit(struct cache *ca)
 {
 	struct discard *d;
diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index d3e15b42a4ab..b39f6f0b45f2 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -178,7 +178,6 @@
 #define pr_fmt(fmt) "bcache: %s() " fmt "\n", __func__
 
 #include <linux/bio.h>
-#include <linux/blktrace_api.h>
 #include <linux/kobject.h>
 #include <linux/list.h>
 #include <linux/mutex.h>
@@ -388,8 +387,6 @@ struct keybuf_key {
 typedef bool (keybuf_pred_fn)(struct keybuf *, struct bkey *);
 
 struct keybuf {
-	keybuf_pred_fn		*key_predicate;
-
 	struct bkey		last_scanned;
 	spinlock_t		lock;
 
@@ -437,9 +434,12 @@ struct bcache_device {
 
 	/* If nonzero, we're detaching/unregistering from cache set */
 	atomic_t		detaching;
+	int			flush_done;
+
+	uint64_t		nr_stripes;
+	unsigned		stripe_size_bits;
+	atomic_t		*stripe_sectors_dirty;
 
-	atomic_long_t		sectors_dirty;
-	unsigned long		sectors_dirty_gc;
 	unsigned long		sectors_dirty_last;
 	long			sectors_dirty_derivative;
 
@@ -531,6 +531,7 @@ struct cached_dev {
 	unsigned		sequential_merge:1;
 	unsigned		verify:1;
 
+	unsigned		partial_stripes_expensive:1;
 	unsigned		writeback_metadata:1;
 	unsigned		writeback_running:1;
 	unsigned char		writeback_percent;
@@ -565,8 +566,7 @@ struct cache {
 
 	unsigned		watermark[WATERMARK_MAX];
 
-	struct closure		alloc;
-	struct workqueue_struct	*alloc_workqueue;
+	struct task_struct	*alloc_thread;
 
 	struct closure		prio;
 	struct prio_set		*disk_buckets;
@@ -664,13 +664,9 @@ struct gc_stat {
  * CACHE_SET_STOPPING always gets set first when we're closing down a cache set;
  * we'll continue to run normally for awhile with CACHE_SET_STOPPING set (i.e.
  * flushing dirty data).
- *
- * CACHE_SET_STOPPING_2 gets set at the last phase, when it's time to shut down
- * the allocation thread.
  */
 #define CACHE_SET_UNREGISTERING		0
 #define	CACHE_SET_STOPPING		1
-#define	CACHE_SET_STOPPING_2		2
 
 struct cache_set {
 	struct closure		cl;
@@ -703,9 +699,6 @@ struct cache_set {
 	/* For the btree cache */
 	struct shrinker		shrink;
 
-	/* For the allocator itself */
-	wait_queue_head_t	alloc_wait;
-
 	/* For the btree cache and anything allocation related */
 	struct mutex		bucket_lock;
 
@@ -823,10 +816,9 @@ struct cache_set {
 
 	/*
 	 * A btree node on disk could have too many bsets for an iterator to fit
-	 * on the stack - this is a single element mempool for btree_read_work()
+	 * on the stack - have to dynamically allocate them
 	 */
-	struct mutex		fill_lock;
-	struct btree_iter	*fill_iter;
+	mempool_t		*fill_iter;
 
 	/*
 	 * btree_sort() is a merge sort and requires temporary space - single
@@ -834,6 +826,7 @@ struct cache_set {
 	 */
 	struct mutex		sort_lock;
 	struct bset		*sort;
+	unsigned		sort_crit_factor;
 
 	/* List of buckets we're currently writing data to */
 	struct list_head	data_buckets;
@@ -906,8 +899,6 @@ static inline unsigned local_clock_us(void)
 	return local_clock() >> 10;
 }
 
-#define MAX_BSETS		4U
-
 #define BTREE_PRIO		USHRT_MAX
 #define INITIAL_PRIO		32768
 
@@ -1112,23 +1103,6 @@ static inline void __bkey_put(struct cache_set *c, struct bkey *k)
 		atomic_dec_bug(&PTR_BUCKET(c, k, i)->pin);
 }
 
-/* Blktrace macros */
-
-#define blktrace_msg(c, fmt, ...)					\
-do {									\
-	struct request_queue *q = bdev_get_queue(c->bdev);		\
-	if (q)								\
-		blk_add_trace_msg(q, fmt, ##__VA_ARGS__);		\
-} while (0)
-
-#define blktrace_msg_all(s, fmt, ...)					\
-do {									\
-	struct cache *_c;						\
-	unsigned i;							\
-	for_each_cache(_c, (s), i)					\
-		blktrace_msg(_c, fmt, ##__VA_ARGS__);			\
-} while (0)
-
 static inline void cached_dev_put(struct cached_dev *dc)
 {
 	if (atomic_dec_and_test(&dc->count))
@@ -1173,10 +1147,16 @@ static inline uint8_t bucket_disk_gen(struct bucket *b)
 	static struct kobj_attribute ksysfs_##n =			\
 		__ATTR(n, S_IWUSR|S_IRUSR, show, store)
 
-/* Forward declarations */
+static inline void wake_up_allocators(struct cache_set *c)
+{
+	struct cache *ca;
+	unsigned i;
+
+	for_each_cache(ca, c, i)
+		wake_up_process(ca->alloc_thread);
+}
 
-void bch_writeback_queue(struct cached_dev *);
-void bch_writeback_add(struct cached_dev *, unsigned);
+/* Forward declarations */
 
 void bch_count_io_errors(struct cache *, int, const char *);
 void bch_bbio_count_io_errors(struct cache_set *, struct bio *,
@@ -1193,7 +1173,6 @@ void bch_submit_bbio(struct bio *, struct cache_set *, struct bkey *, unsigned);
 uint8_t bch_inc_gen(struct cache *, struct bucket *);
 void bch_rescale_priorities(struct cache_set *, int);
 bool bch_bucket_add_unused(struct cache *, struct bucket *);
-void bch_allocator_thread(struct closure *);
 
 long bch_bucket_alloc(struct cache *, unsigned, struct closure *);
 void bch_bucket_free(struct cache_set *, struct bkey *);
@@ -1241,9 +1220,9 @@ void bch_cache_set_stop(struct cache_set *);
 struct cache_set *bch_cache_set_alloc(struct cache_sb *);
 void bch_btree_cache_free(struct cache_set *);
 int bch_btree_cache_alloc(struct cache_set *);
-void bch_cached_dev_writeback_init(struct cached_dev *);
 void bch_moving_init_cache_set(struct cache_set *);
 
+int bch_cache_allocator_start(struct cache *ca);
 void bch_cache_allocator_exit(struct cache *ca);
 int bch_cache_allocator_init(struct cache *ca);
 
diff --git a/drivers/md/bcache/bset.c b/drivers/md/bcache/bset.c
index 1d27d3af3251..8010eed06a51 100644
--- a/drivers/md/bcache/bset.c
+++ b/drivers/md/bcache/bset.c
@@ -78,6 +78,7 @@ struct bkey *bch_keylist_pop(struct keylist *l)
 bool __bch_ptr_invalid(struct cache_set *c, int level, const struct bkey *k)
 {
 	unsigned i;
+	char buf[80];
 
 	if (level && (!KEY_PTRS(k) || !KEY_SIZE(k) || KEY_DIRTY(k)))
 		goto bad;
@@ -102,7 +103,8 @@ bool __bch_ptr_invalid(struct cache_set *c, int level, const struct bkey *k)
 
 	return false;
 bad:
-	cache_bug(c, "spotted bad key %s: %s", pkey(k), bch_ptr_status(c, k));
+	bch_bkey_to_text(buf, sizeof(buf), k);
+	cache_bug(c, "spotted bad key %s: %s", buf, bch_ptr_status(c, k));
 	return true;
 }
 
@@ -162,10 +164,16 @@ bool bch_ptr_bad(struct btree *b, const struct bkey *k)
 #ifdef CONFIG_BCACHE_EDEBUG
 bug:
 	mutex_unlock(&b->c->bucket_lock);
-	btree_bug(b,
+
+	{
+		char buf[80];
+
+		bch_bkey_to_text(buf, sizeof(buf), k);
+		btree_bug(b,
 "inconsistent pointer %s: bucket %zu pin %i prio %i gen %i last_gc %i mark %llu gc_gen %i",
-		  pkey(k), PTR_BUCKET_NR(b->c, k, i), atomic_read(&g->pin),
-		  g->prio, g->gen, g->last_gc, GC_MARK(g), g->gc_gen);
+			  buf, PTR_BUCKET_NR(b->c, k, i), atomic_read(&g->pin),
+			  g->prio, g->gen, g->last_gc, GC_MARK(g), g->gc_gen);
+	}
 	return true;
 #endif
 }
@@ -1084,33 +1092,39 @@ void bch_btree_sort_into(struct btree *b, struct btree *new)
 	new->sets->size = 0;
 }
 
+#define SORT_CRIT	(4096 / sizeof(uint64_t))
+
 void bch_btree_sort_lazy(struct btree *b)
 {
-	if (b->nsets) {
-		unsigned i, j, keys = 0, total;
+	unsigned crit = SORT_CRIT;
+	int i;
 
-		for (i = 0; i <= b->nsets; i++)
-			keys += b->sets[i].data->keys;
-
-		total = keys;
+	/* Don't sort if nothing to do */
+	if (!b->nsets)
+		goto out;
 
-		for (j = 0; j < b->nsets; j++) {
-			if (keys * 2 < total ||
-			    keys < 1000) {
-				bch_btree_sort_partial(b, j);
-				return;
-			}
+	/* If not a leaf node, always sort */
+	if (b->level) {
+		bch_btree_sort(b);
+		return;
+	}
 
-			keys -= b->sets[j].data->keys;
-		}
+	for (i = b->nsets - 1; i >= 0; --i) {
+		crit *= b->c->sort_crit_factor;
 
-		/* Must sort if b->nsets == 3 or we'll overflow */
-		if (b->nsets >= (MAX_BSETS - 1) - b->level) {
-			bch_btree_sort(b);
+		if (b->sets[i].data->keys < crit) {
+			bch_btree_sort_partial(b, i);
 			return;
 		}
 	}
 
+	/* Sort if we'd overflow */
+	if (b->nsets + 1 == MAX_BSETS) {
+		bch_btree_sort(b);
+		return;
+	}
+
+out:
 	bset_build_written_tree(b);
 }
 
diff --git a/drivers/md/bcache/bset.h b/drivers/md/bcache/bset.h
index 57a9cff41546..ae115a253d73 100644
--- a/drivers/md/bcache/bset.h
+++ b/drivers/md/bcache/bset.h
@@ -1,6 +1,8 @@
 #ifndef _BCACHE_BSET_H
 #define _BCACHE_BSET_H
 
+#include <linux/slab.h>
+
 /*
  * BKEYS:
  *
@@ -142,6 +144,8 @@
 
 /* Btree key comparison/iteration */
 
+#define MAX_BSETS		4U
+
 struct btree_iter {
 	size_t size, used;
 	struct btree_iter_set {
diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 7a5658f04e62..ee372884c405 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -24,6 +24,7 @@
 #include "btree.h"
 #include "debug.h"
 #include "request.h"
+#include "writeback.h"
 
 #include <linux/slab.h>
 #include <linux/bitops.h>
@@ -134,44 +135,17 @@ static uint64_t btree_csum_set(struct btree *b, struct bset *i)
 	return crc ^ 0xffffffffffffffffULL;
 }
 
-static void btree_bio_endio(struct bio *bio, int error)
+static void bch_btree_node_read_done(struct btree *b)
 {
-	struct closure *cl = bio->bi_private;
-	struct btree *b = container_of(cl, struct btree, io.cl);
-
-	if (error)
-		set_btree_node_io_error(b);
-
-	bch_bbio_count_io_errors(b->c, bio, error, (bio->bi_rw & WRITE)
-				 ? "writing btree" : "reading btree");
-	closure_put(cl);
-}
-
-static void btree_bio_init(struct btree *b)
-{
-	BUG_ON(b->bio);
-	b->bio = bch_bbio_alloc(b->c);
-
-	b->bio->bi_end_io	= btree_bio_endio;
-	b->bio->bi_private	= &b->io.cl;
-}
-
-void bch_btree_read_done(struct closure *cl)
-{
-	struct btree *b = container_of(cl, struct btree, io.cl);
-	struct bset *i = b->sets[0].data;
-	struct btree_iter *iter = b->c->fill_iter;
 	const char *err = "bad btree header";
-	BUG_ON(b->nsets || b->written);
-
-	bch_bbio_free(b->bio, b->c);
-	b->bio = NULL;
+	struct bset *i = b->sets[0].data;
+	struct btree_iter *iter;
 
-	mutex_lock(&b->c->fill_lock);
+	iter = mempool_alloc(b->c->fill_iter, GFP_NOWAIT);
+	iter->size = b->c->sb.bucket_size / b->c->sb.block_size;
 	iter->used = 0;
 
-	if (btree_node_io_error(b) ||
-	    !i->seq)
+	if (!i->seq)
 		goto err;
 
 	for (;
@@ -228,17 +202,8 @@ void bch_btree_read_done(struct closure *cl)
 	if (b->written < btree_blocks(b))
 		bch_bset_init_next(b);
 out:
-
-	mutex_unlock(&b->c->fill_lock);
-
-	spin_lock(&b->c->btree_read_time_lock);
-	bch_time_stats_update(&b->c->btree_read_time, b->io_start_time);
-	spin_unlock(&b->c->btree_read_time_lock);
-
-	smp_wmb(); /* read_done is our write lock */
-	set_btree_node_read_done(b);
-
-	closure_return(cl);
+	mempool_free(iter, b->c->fill_iter);
+	return;
 err:
 	set_btree_node_io_error(b);
 	bch_cache_set_error(b->c, "%s at bucket %zu, block %zu, %u keys",
@@ -247,48 +212,69 @@ err:
 	goto out;
 }
 
-void bch_btree_read(struct btree *b)
+static void btree_node_read_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+	closure_put(cl);
+}
+
+void bch_btree_node_read(struct btree *b)
 {
-	BUG_ON(b->nsets || b->written);
+	uint64_t start_time = local_clock();
+	struct closure cl;
+	struct bio *bio;
+
+	trace_bcache_btree_read(b);
+
+	closure_init_stack(&cl);
+
+	bio = bch_bbio_alloc(b->c);
+	bio->bi_rw	= REQ_META|READ_SYNC;
+	bio->bi_size	= KEY_SIZE(&b->key) << 9;
+	bio->bi_end_io	= btree_node_read_endio;
+	bio->bi_private	= &cl;
+
+	bch_bio_map(bio, b->sets[0].data);
+
+	bch_submit_bbio(bio, b->c, &b->key, 0);
+	closure_sync(&cl);
 
-	if (!closure_trylock(&b->io.cl, &b->c->cl))
-		BUG();
+	if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
+		set_btree_node_io_error(b);
 
-	b->io_start_time = local_clock();
+	bch_bbio_free(bio, b->c);
 
-	btree_bio_init(b);
-	b->bio->bi_rw	= REQ_META|READ_SYNC;
-	b->bio->bi_size	= KEY_SIZE(&b->key) << 9;
+	if (btree_node_io_error(b))
+		goto err;
 
-	bch_bio_map(b->bio, b->sets[0].data);
+	bch_btree_node_read_done(b);
 
-	pr_debug("%s", pbtree(b));
-	trace_bcache_btree_read(b->bio);
-	bch_submit_bbio(b->bio, b->c, &b->key, 0);
+	spin_lock(&b->c->btree_read_time_lock);
+	bch_time_stats_update(&b->c->btree_read_time, start_time);
+	spin_unlock(&b->c->btree_read_time_lock);
 
-	continue_at(&b->io.cl, bch_btree_read_done, system_wq);
+	return;
+err:
+	bch_cache_set_error(b->c, "io error reading bucket %lu",
+			    PTR_BUCKET_NR(b->c, &b->key, 0));
 }
 
 static void btree_complete_write(struct btree *b, struct btree_write *w)
 {
 	if (w->prio_blocked &&
 	    !atomic_sub_return(w->prio_blocked, &b->c->prio_blocked))
-		wake_up(&b->c->alloc_wait);
+		wake_up_allocators(b->c);
 
 	if (w->journal) {
 		atomic_dec_bug(w->journal);
 		__closure_wake_up(&b->c->journal.wait);
 	}
 
-	if (w->owner)
-		closure_put(w->owner);
-
 	w->prio_blocked	= 0;
 	w->journal	= NULL;
-	w->owner	= NULL;
 }
 
-static void __btree_write_done(struct closure *cl)
+static void __btree_node_write_done(struct closure *cl)
 {
 	struct btree *b = container_of(cl, struct btree, io.cl);
 	struct btree_write *w = btree_prev_write(b);
@@ -304,7 +290,7 @@ static void __btree_write_done(struct closure *cl)
 	closure_return(cl);
 }
 
-static void btree_write_done(struct closure *cl)
+static void btree_node_write_done(struct closure *cl)
 {
 	struct btree *b = container_of(cl, struct btree, io.cl);
 	struct bio_vec *bv;
@@ -313,10 +299,22 @@ static void btree_write_done(struct closure *cl)
 	__bio_for_each_segment(bv, b->bio, n, 0)
 		__free_page(bv->bv_page);
 
-	__btree_write_done(cl);
+	__btree_node_write_done(cl);
 }
 
-static void do_btree_write(struct btree *b)
+static void btree_node_write_endio(struct bio *bio, int error)
+{
+	struct closure *cl = bio->bi_private;
+	struct btree *b = container_of(cl, struct btree, io.cl);
+
+	if (error)
+		set_btree_node_io_error(b);
+
+	bch_bbio_count_io_errors(b->c, bio, error, "writing btree");
+	closure_put(cl);
+}
+
+static void do_btree_node_write(struct btree *b)
 {
 	struct closure *cl = &b->io.cl;
 	struct bset *i = b->sets[b->nsets].data;
@@ -325,15 +323,34 @@ static void do_btree_write(struct btree *b)
 	i->version	= BCACHE_BSET_VERSION;
 	i->csum		= btree_csum_set(b, i);
 
-	btree_bio_init(b);
-	b->bio->bi_rw	= REQ_META|WRITE_SYNC;
-	b->bio->bi_size	= set_blocks(i, b->c) * block_bytes(b->c);
+	BUG_ON(b->bio);
+	b->bio = bch_bbio_alloc(b->c);
+
+	b->bio->bi_end_io	= btree_node_write_endio;
+	b->bio->bi_private	= &b->io.cl;
+	b->bio->bi_rw		= REQ_META|WRITE_SYNC|REQ_FUA;
+	b->bio->bi_size		= set_blocks(i, b->c) * block_bytes(b->c);
 	bch_bio_map(b->bio, i);
 
+	/*
+	 * If we're appending to a leaf node, we don't technically need FUA -
+	 * this write just needs to be persisted before the next journal write,
+	 * which will be marked FLUSH|FUA.
+	 *
+	 * Similarly if we're writing a new btree root - the pointer is going to
+	 * be in the next journal entry.
+	 *
+	 * But if we're writing a new btree node (that isn't a root) or
+	 * appending to a non leaf btree node, we need either FUA or a flush
+	 * when we write the parent with the new pointer. FUA is cheaper than a
+	 * flush, and writes appending to leaf nodes aren't blocking anything so
+	 * just make all btree node writes FUA to keep things sane.
+	 */
+
 	bkey_copy(&k.key, &b->key);
 	SET_PTR_OFFSET(&k.key, 0, PTR_OFFSET(&k.key, 0) + bset_offset(b, i));
 
-	if (!bch_bio_alloc_pages(b->bio, GFP_NOIO)) {
+	if (!bio_alloc_pages(b->bio, GFP_NOIO)) {
 		int j;
 		struct bio_vec *bv;
 		void *base = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1));
@@ -342,40 +359,41 @@ static void do_btree_write(struct btree *b)
 			memcpy(page_address(bv->bv_page),
 			       base + j * PAGE_SIZE, PAGE_SIZE);
 
-		trace_bcache_btree_write(b->bio);
 		bch_submit_bbio(b->bio, b->c, &k.key, 0);
 
-		continue_at(cl, btree_write_done, NULL);
+		continue_at(cl, btree_node_write_done, NULL);
 	} else {
 		b->bio->bi_vcnt = 0;
 		bch_bio_map(b->bio, i);
 
-		trace_bcache_btree_write(b->bio);
 		bch_submit_bbio(b->bio, b->c, &k.key, 0);
 
 		closure_sync(cl);
-		__btree_write_done(cl);
+		__btree_node_write_done(cl);
 	}
 }
 
-static void __btree_write(struct btree *b)
+void bch_btree_node_write(struct btree *b, struct closure *parent)
 {
 	struct bset *i = b->sets[b->nsets].data;
 
+	trace_bcache_btree_write(b);
+
 	BUG_ON(current->bio_list);
+	BUG_ON(b->written >= btree_blocks(b));
+	BUG_ON(b->written && !i->keys);
+	BUG_ON(b->sets->data->seq != i->seq);
+	bch_check_key_order(b, i);
 
-	closure_lock(&b->io, &b->c->cl);
 	cancel_delayed_work(&b->work);
 
+	/* If caller isn't waiting for write, parent refcount is cache set */
+	closure_lock(&b->io, parent ?: &b->c->cl);
+
 	clear_bit(BTREE_NODE_dirty,	 &b->flags);
 	change_bit(BTREE_NODE_write_idx, &b->flags);
 
-	bch_check_key_order(b, i);
-	BUG_ON(b->written && !i->keys);
-
-	do_btree_write(b);
-
-	pr_debug("%s block %i keys %i", pbtree(b), b->written, i->keys);
+	do_btree_node_write(b);
 
 	b->written += set_blocks(i, b->c);
 	atomic_long_add(set_blocks(i, b->c) * b->c->sb.block_size,
@@ -387,37 +405,31 @@ static void __btree_write(struct btree *b)
 		bch_bset_init_next(b);
 }
 
-static void btree_write_work(struct work_struct *w)
+static void btree_node_write_work(struct work_struct *w)
 {
 	struct btree *b = container_of(to_delayed_work(w), struct btree, work);
 
-	down_write(&b->lock);
+	rw_lock(true, b, b->level);
 
 	if (btree_node_dirty(b))
-		__btree_write(b);
-	up_write(&b->lock);
+		bch_btree_node_write(b, NULL);
+	rw_unlock(true, b);
 }
 
-void bch_btree_write(struct btree *b, bool now, struct btree_op *op)
+static void bch_btree_leaf_dirty(struct btree *b, struct btree_op *op)
 {
 	struct bset *i = b->sets[b->nsets].data;
 	struct btree_write *w = btree_current_write(b);
 
-	BUG_ON(b->written &&
-	       (b->written >= btree_blocks(b) ||
-		i->seq != b->sets[0].data->seq ||
-		!i->keys));
+	BUG_ON(!b->written);
+	BUG_ON(!i->keys);
 
-	if (!btree_node_dirty(b)) {
-		set_btree_node_dirty(b);
-		queue_delayed_work(btree_io_wq, &b->work,
-				   msecs_to_jiffies(30000));
-	}
+	if (!btree_node_dirty(b))
+		queue_delayed_work(btree_io_wq, &b->work, 30 * HZ);
 
-	w->prio_blocked += b->prio_blocked;
-	b->prio_blocked = 0;
+	set_btree_node_dirty(b);
 
-	if (op && op->journal && !b->level) {
+	if (op && op->journal) {
 		if (w->journal &&
 		    journal_pin_cmp(b->c, w, op)) {
 			atomic_dec_bug(w->journal);
@@ -430,23 +442,10 @@ void bch_btree_write(struct btree *b, bool now, struct btree_op *op)
 		}
 	}
 
-	if (current->bio_list)
-		return;
-
 	/* Force write if set is too big */
-	if (now ||
-	    b->level ||
-	    set_bytes(i) > PAGE_SIZE - 48) {
-		if (op && now) {
-			/* Must wait on multiple writes */
-			BUG_ON(w->owner);
-			w->owner = &op->cl;
-			closure_get(&op->cl);
-		}
-
-		__btree_write(b);
-	}
-	BUG_ON(!b->written);
+	if (set_bytes(i) > PAGE_SIZE - 48 &&
+	    !current->bio_list)
+		bch_btree_node_write(b, NULL);
 }
 
 /*
@@ -559,7 +558,7 @@ static struct btree *mca_bucket_alloc(struct cache_set *c,
 	init_rwsem(&b->lock);
 	lockdep_set_novalidate_class(&b->lock);
 	INIT_LIST_HEAD(&b->list);
-	INIT_DELAYED_WORK(&b->work, btree_write_work);
+	INIT_DELAYED_WORK(&b->work, btree_node_write_work);
 	b->c = c;
 	closure_init_unlocked(&b->io);
 
@@ -582,7 +581,7 @@ static int mca_reap(struct btree *b, struct closure *cl, unsigned min_order)
 	BUG_ON(btree_node_dirty(b) && !b->sets[0].data);
 
 	if (cl && btree_node_dirty(b))
-		bch_btree_write(b, true, NULL);
+		bch_btree_node_write(b, NULL);
 
 	if (cl)
 		closure_wait_event_async(&b->io.wait, cl,
@@ -623,6 +622,13 @@ static int bch_mca_shrink(struct shrinker *shrink, struct shrink_control *sc)
 	else if (!mutex_trylock(&c->bucket_lock))
 		return -1;
 
+	/*
+	 * It's _really_ critical that we don't free too many btree nodes - we
+	 * have to always leave ourselves a reserve. The reserve is how we
+	 * guarantee that allocating memory for a new btree node can always
+	 * succeed, so that inserting keys into the btree can always succeed and
+	 * IO can always make forward progress:
+	 */
 	nr /= c->btree_pages;
 	nr = min_t(unsigned long, nr, mca_can_free(c));
 
@@ -766,6 +772,8 @@ static struct btree *mca_cannibalize(struct cache_set *c, struct bkey *k,
 	int ret = -ENOMEM;
 	struct btree *i;
 
+	trace_bcache_btree_cache_cannibalize(c);
+
 	if (!cl)
 		return ERR_PTR(-ENOMEM);
 
@@ -784,7 +792,6 @@ static struct btree *mca_cannibalize(struct cache_set *c, struct bkey *k,
 		return ERR_PTR(-EAGAIN);
 	}
 
-	/* XXX: tracepoint */
 	c->try_harder = cl;
 	c->try_harder_start = local_clock();
 retry:
@@ -905,6 +912,9 @@ retry:
 	b = mca_find(c, k);
 
 	if (!b) {
+		if (current->bio_list)
+			return ERR_PTR(-EAGAIN);
+
 		mutex_lock(&c->bucket_lock);
 		b = mca_alloc(c, k, level, &op->cl);
 		mutex_unlock(&c->bucket_lock);
@@ -914,7 +924,7 @@ retry:
 		if (IS_ERR(b))
 			return b;
 
-		bch_btree_read(b);
+		bch_btree_node_read(b);
 
 		if (!write)
 			downgrade_write(&b->lock);
@@ -937,15 +947,12 @@ retry:
 	for (; i <= b->nsets; i++)
 		prefetch(b->sets[i].data);
 
-	if (!closure_wait_event(&b->io.wait, &op->cl,
-				btree_node_read_done(b))) {
-		rw_unlock(write, b);
-		b = ERR_PTR(-EAGAIN);
-	} else if (btree_node_io_error(b)) {
+	if (btree_node_io_error(b)) {
 		rw_unlock(write, b);
-		b = ERR_PTR(-EIO);
-	} else
-		BUG_ON(!b->written);
+		return ERR_PTR(-EIO);
+	}
+
+	BUG_ON(!b->written);
 
 	return b;
 }
@@ -959,7 +966,7 @@ static void btree_node_prefetch(struct cache_set *c, struct bkey *k, int level)
 	mutex_unlock(&c->bucket_lock);
 
 	if (!IS_ERR_OR_NULL(b)) {
-		bch_btree_read(b);
+		bch_btree_node_read(b);
 		rw_unlock(true, b);
 	}
 }
@@ -970,24 +977,19 @@ static void btree_node_free(struct btree *b, struct btree_op *op)
 {
 	unsigned i;
 
+	trace_bcache_btree_node_free(b);
+
 	/*
 	 * The BUG_ON() in btree_node_get() implies that we must have a write
 	 * lock on parent to free or even invalidate a node
 	 */
 	BUG_ON(op->lock <= b->level);
 	BUG_ON(b == b->c->root);
-	pr_debug("bucket %s", pbtree(b));
 
 	if (btree_node_dirty(b))
 		btree_complete_write(b, btree_current_write(b));
 	clear_bit(BTREE_NODE_dirty, &b->flags);
 
-	if (b->prio_blocked &&
-	    !atomic_sub_return(b->prio_blocked, &b->c->prio_blocked))
-		wake_up(&b->c->alloc_wait);
-
-	b->prio_blocked = 0;
-
 	cancel_delayed_work(&b->work);
 
 	mutex_lock(&b->c->bucket_lock);
@@ -1028,17 +1030,20 @@ retry:
 		goto retry;
 	}
 
-	set_btree_node_read_done(b);
 	b->accessed = 1;
 	bch_bset_init_next(b);
 
 	mutex_unlock(&c->bucket_lock);
+
+	trace_bcache_btree_node_alloc(b);
 	return b;
 err_free:
 	bch_bucket_free(c, &k.key);
 	__bkey_put(c, &k.key);
 err:
 	mutex_unlock(&c->bucket_lock);
+
+	trace_bcache_btree_node_alloc_fail(b);
 	return b;
 }
 
@@ -1137,11 +1142,8 @@ static int btree_gc_mark_node(struct btree *b, unsigned *keys,
 		gc->nkeys++;
 
 		gc->data += KEY_SIZE(k);
-		if (KEY_DIRTY(k)) {
+		if (KEY_DIRTY(k))
 			gc->dirty += KEY_SIZE(k);
-			if (d)
-				d->sectors_dirty_gc += KEY_SIZE(k);
-		}
 	}
 
 	for (t = b->sets; t <= &b->sets[b->nsets]; t++)
@@ -1166,14 +1168,11 @@ static struct btree *btree_gc_alloc(struct btree *b, struct bkey *k,
 
 	if (!IS_ERR_OR_NULL(n)) {
 		swap(b, n);
+		__bkey_put(b->c, &b->key);
 
 		memcpy(k->ptr, b->key.ptr,
 		       sizeof(uint64_t) * KEY_PTRS(&b->key));
 
-		__bkey_put(b->c, &b->key);
-		atomic_inc(&b->c->prio_blocked);
-		b->prio_blocked++;
-
 		btree_node_free(n, op);
 		up_write(&n->lock);
 	}
@@ -1278,7 +1277,7 @@ static void btree_gc_coalesce(struct btree *b, struct btree_op *op,
 	btree_node_free(r->b, op);
 	up_write(&r->b->lock);
 
-	pr_debug("coalesced %u nodes", nodes);
+	trace_bcache_btree_gc_coalesce(nodes);
 
 	gc->nodes--;
 	nodes--;
@@ -1293,14 +1292,9 @@ static int btree_gc_recurse(struct btree *b, struct btree_op *op,
 	void write(struct btree *r)
 	{
 		if (!r->written)
-			bch_btree_write(r, true, op);
-		else if (btree_node_dirty(r)) {
-			BUG_ON(btree_current_write(r)->owner);
-			btree_current_write(r)->owner = writes;
-			closure_get(writes);
-
-			bch_btree_write(r, true, NULL);
-		}
+			bch_btree_node_write(r, &op->cl);
+		else if (btree_node_dirty(r))
+			bch_btree_node_write(r, writes);
 
 		up_write(&r->lock);
 	}
@@ -1386,9 +1380,7 @@ static int bch_btree_gc_root(struct btree *b, struct btree_op *op,
 		ret = btree_gc_recurse(b, op, writes, gc);
 
 	if (!b->written || btree_node_dirty(b)) {
-		atomic_inc(&b->c->prio_blocked);
-		b->prio_blocked++;
-		bch_btree_write(b, true, n ? op : NULL);
+		bch_btree_node_write(b, n ? &op->cl : NULL);
 	}
 
 	if (!IS_ERR_OR_NULL(n)) {
@@ -1405,7 +1397,6 @@ static void btree_gc_start(struct cache_set *c)
 {
 	struct cache *ca;
 	struct bucket *b;
-	struct bcache_device **d;
 	unsigned i;
 
 	if (!c->gc_mark_valid)
@@ -1419,16 +1410,12 @@ static void btree_gc_start(struct cache_set *c)
 	for_each_cache(ca, c, i)
 		for_each_bucket(b, ca) {
 			b->gc_gen = b->gen;
-			if (!atomic_read(&b->pin))
+			if (!atomic_read(&b->pin)) {
 				SET_GC_MARK(b, GC_MARK_RECLAIMABLE);
+				SET_GC_SECTORS_USED(b, 0);
+			}
 		}
 
-	for (d = c->devices;
-	     d < c->devices + c->nr_uuids;
-	     d++)
-		if (*d)
-			(*d)->sectors_dirty_gc = 0;
-
 	mutex_unlock(&c->bucket_lock);
 }
 
@@ -1437,7 +1424,6 @@ size_t bch_btree_gc_finish(struct cache_set *c)
 	size_t available = 0;
 	struct bucket *b;
 	struct cache *ca;
-	struct bcache_device **d;
 	unsigned i;
 
 	mutex_lock(&c->bucket_lock);
@@ -1480,22 +1466,6 @@ size_t bch_btree_gc_finish(struct cache_set *c)
 		}
 	}
 
-	for (d = c->devices;
-	     d < c->devices + c->nr_uuids;
-	     d++)
-		if (*d) {
-			unsigned long last =
-				atomic_long_read(&((*d)->sectors_dirty));
-			long difference = (*d)->sectors_dirty_gc - last;
-
-			pr_debug("sectors dirty off by %li", difference);
-
-			(*d)->sectors_dirty_last += difference;
-
-			atomic_long_set(&((*d)->sectors_dirty),
-					(*d)->sectors_dirty_gc);
-		}
-
 	mutex_unlock(&c->bucket_lock);
 	return available;
 }
@@ -1508,10 +1478,9 @@ static void bch_btree_gc(struct closure *cl)
 	struct gc_stat stats;
 	struct closure writes;
 	struct btree_op op;
-
 	uint64_t start_time = local_clock();
-	trace_bcache_gc_start(c->sb.set_uuid);
-	blktrace_msg_all(c, "Starting gc");
+
+	trace_bcache_gc_start(c);
 
 	memset(&stats, 0, sizeof(struct gc_stat));
 	closure_init_stack(&writes);
@@ -1520,14 +1489,14 @@ static void bch_btree_gc(struct closure *cl)
 
 	btree_gc_start(c);
 
+	atomic_inc(&c->prio_blocked);
+
 	ret = btree_root(gc_root, c, &op, &writes, &stats);
 	closure_sync(&op.cl);
 	closure_sync(&writes);
 
 	if (ret) {
-		blktrace_msg_all(c, "Stopped gc");
 		pr_warn("gc failed!");
-
 		continue_at(cl, bch_btree_gc, bch_gc_wq);
 	}
 
@@ -1537,6 +1506,9 @@ static void bch_btree_gc(struct closure *cl)
 
 	available = bch_btree_gc_finish(c);
 
+	atomic_dec(&c->prio_blocked);
+	wake_up_allocators(c);
+
 	bch_time_stats_update(&c->btree_gc_time, start_time);
 
 	stats.key_bytes *= sizeof(uint64_t);
@@ -1544,10 +1516,8 @@ static void bch_btree_gc(struct closure *cl)
 	stats.data	<<= 9;
 	stats.in_use	= (c->nbuckets - available) * 100 / c->nbuckets;
 	memcpy(&c->gc_stats, &stats, sizeof(struct gc_stat));
-	blktrace_msg_all(c, "Finished gc");
 
-	trace_bcache_gc_end(c->sb.set_uuid);
-	wake_up(&c->alloc_wait);
+	trace_bcache_gc_end(c);
 
 	continue_at(cl, bch_moving_gc, bch_gc_wq);
 }
@@ -1654,14 +1624,14 @@ static bool fix_overlapping_extents(struct btree *b,
 				    struct btree_iter *iter,
 				    struct btree_op *op)
 {
-	void subtract_dirty(struct bkey *k, int sectors)
+	void subtract_dirty(struct bkey *k, uint64_t offset, int sectors)
 	{
-		struct bcache_device *d = b->c->devices[KEY_INODE(k)];
-
-		if (KEY_DIRTY(k) && d)
-			atomic_long_sub(sectors, &d->sectors_dirty);
+		if (KEY_DIRTY(k))
+			bcache_dev_sectors_dirty_add(b->c, KEY_INODE(k),
+						     offset, -sectors);
 	}
 
+	uint64_t old_offset;
 	unsigned old_size, sectors_found = 0;
 
 	while (1) {
@@ -1673,6 +1643,7 @@ static bool fix_overlapping_extents(struct btree *b,
 		if (bkey_cmp(k, &START_KEY(insert)) <= 0)
 			continue;
 
+		old_offset = KEY_START(k);
 		old_size = KEY_SIZE(k);
 
 		/*
@@ -1728,7 +1699,7 @@ static bool fix_overlapping_extents(struct btree *b,
 
 			struct bkey *top;
 
-			subtract_dirty(k, KEY_SIZE(insert));
+			subtract_dirty(k, KEY_START(insert), KEY_SIZE(insert));
 
 			if (bkey_written(b, k)) {
 				/*
@@ -1775,7 +1746,7 @@ static bool fix_overlapping_extents(struct btree *b,
 			}
 		}
 
-		subtract_dirty(k, old_size - KEY_SIZE(k));
+		subtract_dirty(k, old_offset, old_size - KEY_SIZE(k));
 	}
 
 check_failed:
@@ -1798,7 +1769,7 @@ static bool btree_insert_key(struct btree *b, struct btree_op *op,
 {
 	struct bset *i = b->sets[b->nsets].data;
 	struct bkey *m, *prev;
-	const char *status = "insert";
+	unsigned status = BTREE_INSERT_STATUS_INSERT;
 
 	BUG_ON(bkey_cmp(k, &b->key) > 0);
 	BUG_ON(b->level && !KEY_PTRS(k));
@@ -1831,17 +1802,17 @@ static bool btree_insert_key(struct btree *b, struct btree_op *op,
 			goto insert;
 
 		/* prev is in the tree, if we merge we're done */
-		status = "back merging";
+		status = BTREE_INSERT_STATUS_BACK_MERGE;
 		if (prev &&
 		    bch_bkey_try_merge(b, prev, k))
 			goto merged;
 
-		status = "overwrote front";
+		status = BTREE_INSERT_STATUS_OVERWROTE;
 		if (m != end(i) &&
 		    KEY_PTRS(m) == KEY_PTRS(k) && !KEY_SIZE(m))
 			goto copy;
 
-		status = "front merge";
+		status = BTREE_INSERT_STATUS_FRONT_MERGE;
 		if (m != end(i) &&
 		    bch_bkey_try_merge(b, k, m))
 			goto copy;
@@ -1851,21 +1822,21 @@ static bool btree_insert_key(struct btree *b, struct btree_op *op,
 insert:	shift_keys(b, m, k);
 copy:	bkey_copy(m, k);
 merged:
-	bch_check_keys(b, "%s for %s at %s: %s", status,
-		       op_type(op), pbtree(b), pkey(k));
-	bch_check_key_order_msg(b, i, "%s for %s at %s: %s", status,
-				op_type(op), pbtree(b), pkey(k));
+	if (KEY_DIRTY(k))
+		bcache_dev_sectors_dirty_add(b->c, KEY_INODE(k),
+					     KEY_START(k), KEY_SIZE(k));
+
+	bch_check_keys(b, "%u for %s", status, op_type(op));
 
 	if (b->level && !KEY_OFFSET(k))
-		b->prio_blocked++;
+		btree_current_write(b)->prio_blocked++;
 
-	pr_debug("%s for %s at %s: %s", status,
-		 op_type(op), pbtree(b), pkey(k));
+	trace_bcache_btree_insert_key(b, k, op->type, status);
 
 	return true;
 }
 
-bool bch_btree_insert_keys(struct btree *b, struct btree_op *op)
+static bool bch_btree_insert_keys(struct btree *b, struct btree_op *op)
 {
 	bool ret = false;
 	struct bkey *k;
@@ -1896,7 +1867,7 @@ bool bch_btree_insert_check_key(struct btree *b, struct btree_op *op,
 	    should_split(b))
 		goto out;
 
-	op->replace = KEY(op->inode, bio_end(bio), bio_sectors(bio));
+	op->replace = KEY(op->inode, bio_end_sector(bio), bio_sectors(bio));
 
 	SET_KEY_PTRS(&op->replace, 1);
 	get_random_bytes(&op->replace.ptr[0], sizeof(uint64_t));
@@ -1907,7 +1878,6 @@ bool bch_btree_insert_check_key(struct btree *b, struct btree_op *op,
 
 	BUG_ON(op->type != BTREE_INSERT);
 	BUG_ON(!btree_insert_key(b, op, &tmp.k));
-	bch_btree_write(b, false, NULL);
 	ret = true;
 out:
 	downgrade_write(&b->lock);
@@ -1929,12 +1899,11 @@ static int btree_split(struct btree *b, struct btree_op *op)
 
 	split = set_blocks(n1->sets[0].data, n1->c) > (btree_blocks(b) * 4) / 5;
 
-	pr_debug("%ssplitting at %s keys %i", split ? "" : "not ",
-		 pbtree(b), n1->sets[0].data->keys);
-
 	if (split) {
 		unsigned keys = 0;
 
+		trace_bcache_btree_node_split(b, n1->sets[0].data->keys);
+
 		n2 = bch_btree_node_alloc(b->c, b->level, &op->cl);
 		if (IS_ERR(n2))
 			goto err_free1;
@@ -1967,18 +1936,21 @@ static int btree_split(struct btree *b, struct btree_op *op)
 		bkey_copy_key(&n2->key, &b->key);
 
 		bch_keylist_add(&op->keys, &n2->key);
-		bch_btree_write(n2, true, op);
+		bch_btree_node_write(n2, &op->cl);
 		rw_unlock(true, n2);
-	} else
+	} else {
+		trace_bcache_btree_node_compact(b, n1->sets[0].data->keys);
+
 		bch_btree_insert_keys(n1, op);
+	}
 
 	bch_keylist_add(&op->keys, &n1->key);
-	bch_btree_write(n1, true, op);
+	bch_btree_node_write(n1, &op->cl);
 
 	if (n3) {
 		bkey_copy_key(&n3->key, &MAX_KEY);
 		bch_btree_insert_keys(n3, op);
-		bch_btree_write(n3, true, op);
+		bch_btree_node_write(n3, &op->cl);
 
 		closure_sync(&op->cl);
 		bch_btree_set_root(n3);
@@ -2082,8 +2054,12 @@ static int bch_btree_insert_recurse(struct btree *b, struct btree_op *op,
 
 		BUG_ON(write_block(b) != b->sets[b->nsets].data);
 
-		if (bch_btree_insert_keys(b, op))
-			bch_btree_write(b, false, op);
+		if (bch_btree_insert_keys(b, op)) {
+			if (!b->level)
+				bch_btree_leaf_dirty(b, op);
+			else
+				bch_btree_node_write(b, &op->cl);
+		}
 	}
 
 	return 0;
@@ -2140,6 +2116,11 @@ int bch_btree_insert(struct btree_op *op, struct cache_set *c)
 void bch_btree_set_root(struct btree *b)
 {
 	unsigned i;
+	struct closure cl;
+
+	closure_init_stack(&cl);
+
+	trace_bcache_btree_set_root(b);
 
 	BUG_ON(!b->written);
 
@@ -2153,8 +2134,8 @@ void bch_btree_set_root(struct btree *b)
 	b->c->root = b;
 	__bkey_put(b->c, &b->key);
 
-	bch_journal_meta(b->c, NULL);
-	pr_debug("%s for %pf", pbtree(b), __builtin_return_address(0));
+	bch_journal_meta(b->c, &cl);
+	closure_sync(&cl);
 }
 
 /* Cache lookup */
@@ -2215,9 +2196,6 @@ static int submit_partial_cache_hit(struct btree *b, struct btree_op *op,
 					 KEY_OFFSET(k) - bio->bi_sector);
 
 		n = bch_bio_split(bio, sectors, GFP_NOIO, s->d->bio_split);
-		if (!n)
-			return -EAGAIN;
-
 		if (n == bio)
 			op->lookup_done = true;
 
@@ -2240,7 +2218,6 @@ static int submit_partial_cache_hit(struct btree *b, struct btree_op *op,
 		n->bi_end_io	= bch_cache_read_endio;
 		n->bi_private	= &s->cl;
 
-		trace_bcache_cache_hit(n);
 		__bch_submit_bbio(n, b->c);
 	}
 
@@ -2257,9 +2234,6 @@ int bch_btree_search_recurse(struct btree *b, struct btree_op *op)
 	struct btree_iter iter;
 	bch_btree_iter_init(b, &iter, &KEY(op->inode, bio->bi_sector, 0));
 
-	pr_debug("at %s searching for %u:%llu", pbtree(b), op->inode,
-		 (uint64_t) bio->bi_sector);
-
 	do {
 		k = bch_btree_iter_next_filter(&iter, b, bch_ptr_bad);
 		if (!k) {
@@ -2303,7 +2277,8 @@ static inline int keybuf_nonoverlapping_cmp(struct keybuf_key *l,
 }
 
 static int bch_btree_refill_keybuf(struct btree *b, struct btree_op *op,
-				   struct keybuf *buf, struct bkey *end)
+				   struct keybuf *buf, struct bkey *end,
+				   keybuf_pred_fn *pred)
 {
 	struct btree_iter iter;
 	bch_btree_iter_init(b, &iter, &buf->last_scanned);
@@ -2322,11 +2297,9 @@ static int bch_btree_refill_keybuf(struct btree *b, struct btree_op *op,
 			if (bkey_cmp(&buf->last_scanned, end) >= 0)
 				break;
 
-			if (buf->key_predicate(buf, k)) {
+			if (pred(buf, k)) {
 				struct keybuf_key *w;
 
-				pr_debug("%s", pkey(k));
-
 				spin_lock(&buf->lock);
 
 				w = array_alloc(&buf->freelist);
@@ -2343,7 +2316,7 @@ static int bch_btree_refill_keybuf(struct btree *b, struct btree_op *op,
 			if (!k)
 				break;
 
-			btree(refill_keybuf, k, b, op, buf, end);
+			btree(refill_keybuf, k, b, op, buf, end, pred);
 			/*
 			 * Might get an error here, but can't really do anything
 			 * and it'll get logged elsewhere. Just read what we
@@ -2361,7 +2334,7 @@ static int bch_btree_refill_keybuf(struct btree *b, struct btree_op *op,
 }
 
 void bch_refill_keybuf(struct cache_set *c, struct keybuf *buf,
-			  struct bkey *end)
+		       struct bkey *end, keybuf_pred_fn *pred)
 {
 	struct bkey start = buf->last_scanned;
 	struct btree_op op;
@@ -2369,7 +2342,7 @@ void bch_refill_keybuf(struct cache_set *c, struct keybuf *buf,
 
 	cond_resched();
 
-	btree_root(refill_keybuf, c, &op, buf, end);
+	btree_root(refill_keybuf, c, &op, buf, end, pred);
 	closure_sync(&op.cl);
 
 	pr_debug("found %s keys from %llu:%llu to %llu:%llu",
@@ -2455,7 +2428,8 @@ struct keybuf_key *bch_keybuf_next(struct keybuf *buf)
 
 struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *c,
 					     struct keybuf *buf,
-					     struct bkey *end)
+					     struct bkey *end,
+					     keybuf_pred_fn *pred)
 {
 	struct keybuf_key *ret;
 
@@ -2469,15 +2443,14 @@ struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *c,
 			break;
 		}
 
-		bch_refill_keybuf(c, buf, end);
+		bch_refill_keybuf(c, buf, end, pred);
 	}
 
 	return ret;
 }
 
-void bch_keybuf_init(struct keybuf *buf, keybuf_pred_fn *fn)
+void bch_keybuf_init(struct keybuf *buf)
 {
-	buf->key_predicate	= fn;
 	buf->last_scanned	= MAX_KEY;
 	buf->keys		= RB_ROOT;
 
diff --git a/drivers/md/bcache/btree.h b/drivers/md/bcache/btree.h
index af4a7092a28c..3333d3723633 100644
--- a/drivers/md/bcache/btree.h
+++ b/drivers/md/bcache/btree.h
@@ -102,7 +102,6 @@
 #include "debug.h"
 
 struct btree_write {
-	struct closure		*owner;
 	atomic_t		*journal;
 
 	/* If btree_split() frees a btree node, it writes a new pointer to that
@@ -142,16 +141,12 @@ struct btree {
 	 */
 	struct bset_tree	sets[MAX_BSETS];
 
-	/* Used to refcount bio splits, also protects b->bio */
+	/* For outstanding btree writes, used as a lock - protects write_idx */
 	struct closure_with_waitlist	io;
 
-	/* Gets transferred to w->prio_blocked - see the comment there */
-	int			prio_blocked;
-
 	struct list_head	list;
 	struct delayed_work	work;
 
-	uint64_t		io_start_time;
 	struct btree_write	writes[2];
 	struct bio		*bio;
 };
@@ -164,13 +159,11 @@ static inline void set_btree_node_ ## flag(struct btree *b)		\
 {	set_bit(BTREE_NODE_ ## flag, &b->flags); }			\
 
 enum btree_flags {
-	BTREE_NODE_read_done,
 	BTREE_NODE_io_error,
 	BTREE_NODE_dirty,
 	BTREE_NODE_write_idx,
 };
 
-BTREE_FLAG(read_done);
 BTREE_FLAG(io_error);
 BTREE_FLAG(dirty);
 BTREE_FLAG(write_idx);
@@ -278,6 +271,13 @@ struct btree_op {
 	BKEY_PADDED(replace);
 };
 
+enum {
+	BTREE_INSERT_STATUS_INSERT,
+	BTREE_INSERT_STATUS_BACK_MERGE,
+	BTREE_INSERT_STATUS_OVERWROTE,
+	BTREE_INSERT_STATUS_FRONT_MERGE,
+};
+
 void bch_btree_op_init_stack(struct btree_op *);
 
 static inline void rw_lock(bool w, struct btree *b, int level)
@@ -293,9 +293,7 @@ static inline void rw_unlock(bool w, struct btree *b)
 #ifdef CONFIG_BCACHE_EDEBUG
 	unsigned i;
 
-	if (w &&
-	    b->key.ptr[0] &&
-	    btree_node_read_done(b))
+	if (w && b->key.ptr[0])
 		for (i = 0; i <= b->nsets; i++)
 			bch_check_key_order(b, b->sets[i].data);
 #endif
@@ -370,9 +368,8 @@ static inline bool should_split(struct btree *b)
 		 > btree_blocks(b));
 }
 
-void bch_btree_read_done(struct closure *);
-void bch_btree_read(struct btree *);
-void bch_btree_write(struct btree *b, bool now, struct btree_op *op);
+void bch_btree_node_read(struct btree *);
+void bch_btree_node_write(struct btree *, struct closure *);
 
 void bch_cannibalize_unlock(struct cache_set *, struct closure *);
 void bch_btree_set_root(struct btree *);
@@ -380,7 +377,6 @@ struct btree *bch_btree_node_alloc(struct cache_set *, int, struct closure *);
 struct btree *bch_btree_node_get(struct cache_set *, struct bkey *,
 				int, struct btree_op *);
 
-bool bch_btree_insert_keys(struct btree *, struct btree_op *);
 bool bch_btree_insert_check_key(struct btree *, struct btree_op *,
 				   struct bio *);
 int bch_btree_insert(struct btree_op *, struct cache_set *);
@@ -393,13 +389,14 @@ void bch_moving_gc(struct closure *);
 int bch_btree_check(struct cache_set *, struct btree_op *);
 uint8_t __bch_btree_mark_key(struct cache_set *, int, struct bkey *);
 
-void bch_keybuf_init(struct keybuf *, keybuf_pred_fn *);
-void bch_refill_keybuf(struct cache_set *, struct keybuf *, struct bkey *);
+void bch_keybuf_init(struct keybuf *);
+void bch_refill_keybuf(struct cache_set *, struct keybuf *, struct bkey *,
+		       keybuf_pred_fn *);
 bool bch_keybuf_check_overlapping(struct keybuf *, struct bkey *,
 				  struct bkey *);
 void bch_keybuf_del(struct keybuf *, struct keybuf_key *);
 struct keybuf_key *bch_keybuf_next(struct keybuf *);
-struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *,
-					  struct keybuf *, struct bkey *);
+struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *, struct keybuf *,
+					  struct bkey *, keybuf_pred_fn *);
 
 #endif
diff --git a/drivers/md/bcache/closure.c b/drivers/md/bcache/closure.c
index bd05a9a8c7cf..9aba2017f0d1 100644
--- a/drivers/md/bcache/closure.c
+++ b/drivers/md/bcache/closure.c
@@ -66,16 +66,18 @@ static inline void closure_put_after_sub(struct closure *cl, int flags)
 		} else {
 			struct closure *parent = cl->parent;
 			struct closure_waitlist *wait = closure_waitlist(cl);
+			closure_fn *destructor = cl->fn;
 
 			closure_debug_destroy(cl);
 
+			smp_mb();
 			atomic_set(&cl->remaining, -1);
 
 			if (wait)
 				closure_wake_up(wait);
 
-			if (cl->fn)
-				cl->fn(cl);
+			if (destructor)
+				destructor(cl);
 
 			if (parent)
 				closure_put(parent);
diff --git a/drivers/md/bcache/debug.c b/drivers/md/bcache/debug.c
index 89fd5204924e..88e6411eab4f 100644
--- a/drivers/md/bcache/debug.c
+++ b/drivers/md/bcache/debug.c
@@ -47,11 +47,10 @@ const char *bch_ptr_status(struct cache_set *c, const struct bkey *k)
 	return "";
 }
 
-struct keyprint_hack bch_pkey(const struct bkey *k)
+int bch_bkey_to_text(char *buf, size_t size, const struct bkey *k)
 {
 	unsigned i = 0;
-	struct keyprint_hack r;
-	char *out = r.s, *end = r.s + KEYHACK_SIZE;
+	char *out = buf, *end = buf + size;
 
 #define p(...)	(out += scnprintf(out, end - out, __VA_ARGS__))
 
@@ -75,16 +74,14 @@ struct keyprint_hack bch_pkey(const struct bkey *k)
 	if (KEY_CSUM(k))
 		p(" cs%llu %llx", KEY_CSUM(k), k->ptr[1]);
 #undef p
-	return r;
+	return out - buf;
 }
 
-struct keyprint_hack bch_pbtree(const struct btree *b)
+int bch_btree_to_text(char *buf, size_t size, const struct btree *b)
 {
-	struct keyprint_hack r;
-
-	snprintf(r.s, 40, "%zu level %i/%i", PTR_BUCKET_NR(b->c, &b->key, 0),
-		 b->level, b->c->root ? b->c->root->level : -1);
-	return r;
+	return scnprintf(buf, size, "%zu level %i/%i",
+			 PTR_BUCKET_NR(b->c, &b->key, 0),
+			 b->level, b->c->root ? b->c->root->level : -1);
 }
 
 #if defined(CONFIG_BCACHE_DEBUG) || defined(CONFIG_BCACHE_EDEBUG)
@@ -100,10 +97,12 @@ static void dump_bset(struct btree *b, struct bset *i)
 {
 	struct bkey *k;
 	unsigned j;
+	char buf[80];
 
 	for (k = i->start; k < end(i); k = bkey_next(k)) {
+		bch_bkey_to_text(buf, sizeof(buf), k);
 		printk(KERN_ERR "block %zu key %zi/%u: %s", index(i, b),
-		       (uint64_t *) k - i->d, i->keys, pkey(k));
+		       (uint64_t *) k - i->d, i->keys, buf);
 
 		for (j = 0; j < KEY_PTRS(k); j++) {
 			size_t n = PTR_BUCKET_NR(b->c, k, j);
@@ -144,7 +143,7 @@ void bch_btree_verify(struct btree *b, struct bset *new)
 	v->written = 0;
 	v->level = b->level;
 
-	bch_btree_read(v);
+	bch_btree_node_read(v);
 	closure_wait_event(&v->io.wait, &cl,
 			   atomic_read(&b->io.cl.remaining) == -1);
 
@@ -200,7 +199,7 @@ void bch_data_verify(struct search *s)
 	if (!check)
 		return;
 
-	if (bch_bio_alloc_pages(check, GFP_NOIO))
+	if (bio_alloc_pages(check, GFP_NOIO))
 		goto out_put;
 
 	check->bi_rw		= READ_SYNC;
@@ -252,6 +251,7 @@ static void vdump_bucket_and_panic(struct btree *b, const char *fmt,
 				   va_list args)
 {
 	unsigned i;
+	char buf[80];
 
 	console_lock();
 
@@ -262,7 +262,8 @@ static void vdump_bucket_and_panic(struct btree *b, const char *fmt,
 
 	console_unlock();
 
-	panic("at %s\n", pbtree(b));
+	bch_btree_to_text(buf, sizeof(buf), b);
+	panic("at %s\n", buf);
 }
 
 void bch_check_key_order_msg(struct btree *b, struct bset *i,
@@ -337,6 +338,7 @@ static ssize_t bch_dump_read(struct file *file, char __user *buf,
 {
 	struct dump_iterator *i = file->private_data;
 	ssize_t ret = 0;
+	char kbuf[80];
 
 	while (size) {
 		struct keybuf_key *w;
@@ -355,11 +357,12 @@ static ssize_t bch_dump_read(struct file *file, char __user *buf,
 		if (i->bytes)
 			break;
 
-		w = bch_keybuf_next_rescan(i->c, &i->keys, &MAX_KEY);
+		w = bch_keybuf_next_rescan(i->c, &i->keys, &MAX_KEY, dump_pred);
 		if (!w)
 			break;
 
-		i->bytes = snprintf(i->buf, PAGE_SIZE, "%s\n", pkey(&w->key));
+		bch_bkey_to_text(kbuf, sizeof(kbuf), &w->key);
+		i->bytes = snprintf(i->buf, PAGE_SIZE, "%s\n", kbuf);
 		bch_keybuf_del(&i->keys, w);
 	}
 
@@ -377,7 +380,7 @@ static int bch_dump_open(struct inode *inode, struct file *file)
 
 	file->private_data = i;
 	i->c = c;
-	bch_keybuf_init(&i->keys, dump_pred);
+	bch_keybuf_init(&i->keys);
 	i->keys.last_scanned = KEY(0, 0, 0);
 
 	return 0;
@@ -409,142 +412,6 @@ void bch_debug_init_cache_set(struct cache_set *c)
 
 #endif
 
-/* Fuzz tester has rotted: */
-#if 0
-
-static ssize_t btree_fuzz(struct kobject *k, struct kobj_attribute *a,
-			  const char *buffer, size_t size)
-{
-	void dump(struct btree *b)
-	{
-		struct bset *i;
-
-		for (i = b->sets[0].data;
-		     index(i, b) < btree_blocks(b) &&
-		     i->seq == b->sets[0].data->seq;
-		     i = ((void *) i) + set_blocks(i, b->c) * block_bytes(b->c))
-			dump_bset(b, i);
-	}
-
-	struct cache_sb *sb;
-	struct cache_set *c;
-	struct btree *all[3], *b, *fill, *orig;
-	int j;
-
-	struct btree_op op;
-	bch_btree_op_init_stack(&op);
-
-	sb = kzalloc(sizeof(struct cache_sb), GFP_KERNEL);
-	if (!sb)
-		return -ENOMEM;
-
-	sb->bucket_size = 128;
-	sb->block_size = 4;
-
-	c = bch_cache_set_alloc(sb);
-	if (!c)
-		return -ENOMEM;
-
-	for (j = 0; j < 3; j++) {
-		BUG_ON(list_empty(&c->btree_cache));
-		all[j] = list_first_entry(&c->btree_cache, struct btree, list);
-		list_del_init(&all[j]->list);
-
-		all[j]->key = KEY(0, 0, c->sb.bucket_size);
-		bkey_copy_key(&all[j]->key, &MAX_KEY);
-	}
-
-	b = all[0];
-	fill = all[1];
-	orig = all[2];
-
-	while (1) {
-		for (j = 0; j < 3; j++)
-			all[j]->written = all[j]->nsets = 0;
-
-		bch_bset_init_next(b);
-
-		while (1) {
-			struct bset *i = write_block(b);
-			struct bkey *k = op.keys.top;
-			unsigned rand;
-
-			bkey_init(k);
-			rand = get_random_int();
-
-			op.type = rand & 1
-				? BTREE_INSERT
-				: BTREE_REPLACE;
-			rand >>= 1;
-
-			SET_KEY_SIZE(k, bucket_remainder(c, rand));
-			rand >>= c->bucket_bits;
-			rand &= 1024 * 512 - 1;
-			rand += c->sb.bucket_size;
-			SET_KEY_OFFSET(k, rand);
-#if 0
-			SET_KEY_PTRS(k, 1);
-#endif
-			bch_keylist_push(&op.keys);
-			bch_btree_insert_keys(b, &op);
-
-			if (should_split(b) ||
-			    set_blocks(i, b->c) !=
-			    __set_blocks(i, i->keys + 15, b->c)) {
-				i->csum = csum_set(i);
-
-				memcpy(write_block(fill),
-				       i, set_bytes(i));
-
-				b->written += set_blocks(i, b->c);
-				fill->written = b->written;
-				if (b->written == btree_blocks(b))
-					break;
-
-				bch_btree_sort_lazy(b);
-				bch_bset_init_next(b);
-			}
-		}
-
-		memcpy(orig->sets[0].data,
-		       fill->sets[0].data,
-		       btree_bytes(c));
-
-		bch_btree_sort(b);
-		fill->written = 0;
-		bch_btree_read_done(&fill->io.cl);
-
-		if (b->sets[0].data->keys != fill->sets[0].data->keys ||
-		    memcmp(b->sets[0].data->start,
-			   fill->sets[0].data->start,
-			   b->sets[0].data->keys * sizeof(uint64_t))) {
-			struct bset *i = b->sets[0].data;
-			struct bkey *k, *l;
-
-			for (k = i->start,
-			     l = fill->sets[0].data->start;
-			     k < end(i);
-			     k = bkey_next(k), l = bkey_next(l))
-				if (bkey_cmp(k, l) ||
-				    KEY_SIZE(k) != KEY_SIZE(l))
-					pr_err("key %zi differs: %s != %s",
-					       (uint64_t *) k - i->d,
-					       pkey(k), pkey(l));
-
-			for (j = 0; j < 3; j++) {
-				pr_err("**** Set %i ****", j);
-				dump(all[j]);
-			}
-			panic("\n");
-		}
-
-		pr_info("fuzz complete: %i keys", b->sets[0].data->keys);
-	}
-}
-
-kobj_attribute_write(fuzz, btree_fuzz);
-#endif
-
 void bch_debug_exit(void)
 {
 	if (!IS_ERR_OR_NULL(debug))
@@ -554,11 +421,6 @@ void bch_debug_exit(void)
 int __init bch_debug_init(struct kobject *kobj)
 {
 	int ret = 0;
-#if 0
-	ret = sysfs_create_file(kobj, &ksysfs_fuzz.attr);
-	if (ret)
-		return ret;
-#endif
 
 	debug = debugfs_create_dir("bcache", NULL);
 	return ret;
diff --git a/drivers/md/bcache/debug.h b/drivers/md/bcache/debug.h
index f9378a218148..1c39b5a2489b 100644
--- a/drivers/md/bcache/debug.h
+++ b/drivers/md/bcache/debug.h
@@ -3,15 +3,8 @@
 
 /* Btree/bkey debug printing */
 
-#define KEYHACK_SIZE 80
-struct keyprint_hack {
-	char s[KEYHACK_SIZE];
-};
-
-struct keyprint_hack bch_pkey(const struct bkey *k);
-struct keyprint_hack bch_pbtree(const struct btree *b);
-#define pkey(k)		(&bch_pkey(k).s[0])
-#define pbtree(b)	(&bch_pbtree(b).s[0])
+int bch_bkey_to_text(char *buf, size_t size, const struct bkey *k);
+int bch_btree_to_text(char *buf, size_t size, const struct btree *b);
 
 #ifdef CONFIG_BCACHE_EDEBUG
 
diff --git a/drivers/md/bcache/io.c b/drivers/md/bcache/io.c
index 48efd4dea645..9056632995b1 100644
--- a/drivers/md/bcache/io.c
+++ b/drivers/md/bcache/io.c
@@ -9,6 +9,8 @@
 #include "bset.h"
 #include "debug.h"
 
+#include <linux/blkdev.h>
+
 static void bch_bi_idx_hack_endio(struct bio *bio, int error)
 {
 	struct bio *p = bio->bi_private;
@@ -66,13 +68,6 @@ static void bch_generic_make_request_hack(struct bio *bio)
  * The newly allocated bio will point to @bio's bi_io_vec, if the split was on a
  * bvec boundry; it is the caller's responsibility to ensure that @bio is not
  * freed before the split.
- *
- * If bch_bio_split() is running under generic_make_request(), it's not safe to
- * allocate more than one bio from the same bio set. Therefore, if it is running
- * under generic_make_request() it masks out __GFP_WAIT when doing the
- * allocation. The caller must check for failure if there's any possibility of
- * it being called from under generic_make_request(); it is then the caller's
- * responsibility to retry from a safe context (by e.g. punting to workqueue).
  */
 struct bio *bch_bio_split(struct bio *bio, int sectors,
 			  gfp_t gfp, struct bio_set *bs)
@@ -83,20 +78,13 @@ struct bio *bch_bio_split(struct bio *bio, int sectors,
 
 	BUG_ON(sectors <= 0);
 
-	/*
-	 * If we're being called from underneath generic_make_request() and we
-	 * already allocated any bios from this bio set, we risk deadlock if we
-	 * use the mempool. So instead, we possibly fail and let the caller punt
-	 * to workqueue or somesuch and retry in a safe context.
-	 */
-	if (current->bio_list)
-		gfp &= ~__GFP_WAIT;
-
 	if (sectors >= bio_sectors(bio))
 		return bio;
 
 	if (bio->bi_rw & REQ_DISCARD) {
 		ret = bio_alloc_bioset(gfp, 1, bs);
+		if (!ret)
+			return NULL;
 		idx = 0;
 		goto out;
 	}
@@ -160,17 +148,18 @@ static unsigned bch_bio_max_sectors(struct bio *bio)
 	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
 	unsigned max_segments = min_t(unsigned, BIO_MAX_PAGES,
 				      queue_max_segments(q));
-	struct bio_vec *bv, *end = bio_iovec(bio) +
-		min_t(int, bio_segments(bio), max_segments);
 
 	if (bio->bi_rw & REQ_DISCARD)
 		return min(ret, q->limits.max_discard_sectors);
 
 	if (bio_segments(bio) > max_segments ||
 	    q->merge_bvec_fn) {
+		struct bio_vec *bv;
+		int i, seg = 0;
+
 		ret = 0;
 
-		for (bv = bio_iovec(bio); bv < end; bv++) {
+		bio_for_each_segment(bv, bio, i) {
 			struct bvec_merge_data bvm = {
 				.bi_bdev	= bio->bi_bdev,
 				.bi_sector	= bio->bi_sector,
@@ -178,10 +167,14 @@ static unsigned bch_bio_max_sectors(struct bio *bio)
 				.bi_rw		= bio->bi_rw,
 			};
 
+			if (seg == max_segments)
+				break;
+
 			if (q->merge_bvec_fn &&
 			    q->merge_bvec_fn(q, &bvm, bv) < (int) bv->bv_len)
 				break;
 
+			seg++;
 			ret += bv->bv_len >> 9;
 		}
 	}
@@ -218,30 +211,10 @@ static void bch_bio_submit_split_endio(struct bio *bio, int error)
 	closure_put(cl);
 }
 
-static void __bch_bio_submit_split(struct closure *cl)
-{
-	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);
-	struct bio *bio = s->bio, *n;
-
-	do {
-		n = bch_bio_split(bio, bch_bio_max_sectors(bio),
-				  GFP_NOIO, s->p->bio_split);
-		if (!n)
-			continue_at(cl, __bch_bio_submit_split, system_wq);
-
-		n->bi_end_io	= bch_bio_submit_split_endio;
-		n->bi_private	= cl;
-
-		closure_get(cl);
-		bch_generic_make_request_hack(n);
-	} while (n != bio);
-
-	continue_at(cl, bch_bio_submit_split_done, NULL);
-}
-
 void bch_generic_make_request(struct bio *bio, struct bio_split_pool *p)
 {
 	struct bio_split_hook *s;
+	struct bio *n;
 
 	if (!bio_has_data(bio) && !(bio->bi_rw & REQ_DISCARD))
 		goto submit;
@@ -250,6 +223,7 @@ void bch_generic_make_request(struct bio *bio, struct bio_split_pool *p)
 		goto submit;
 
 	s = mempool_alloc(p->bio_split_hook, GFP_NOIO);
+	closure_init(&s->cl, NULL);
 
 	s->bio		= bio;
 	s->p		= p;
@@ -257,8 +231,18 @@ void bch_generic_make_request(struct bio *bio, struct bio_split_pool *p)
 	s->bi_private	= bio->bi_private;
 	bio_get(bio);
 
-	closure_call(&s->cl, __bch_bio_submit_split, NULL, NULL);
-	return;
+	do {
+		n = bch_bio_split(bio, bch_bio_max_sectors(bio),
+				  GFP_NOIO, s->p->bio_split);
+
+		n->bi_end_io	= bch_bio_submit_split_endio;
+		n->bi_private	= &s->cl;
+
+		closure_get(&s->cl);
+		bch_generic_make_request_hack(n);
+	} while (n != bio);
+
+	continue_at(&s->cl, bch_bio_submit_split_done, NULL);
 submit:
 	bch_generic_make_request_hack(bio);
 }
diff --git a/drivers/md/bcache/journal.c b/drivers/md/bcache/journal.c
index 8c8dfdcd9d4c..ba95ab84b2be 100644
--- a/drivers/md/bcache/journal.c
+++ b/drivers/md/bcache/journal.c
@@ -9,6 +9,8 @@
 #include "debug.h"
 #include "request.h"
 
+#include <trace/events/bcache.h>
+
 /*
  * Journal replay/recovery:
  *
@@ -182,9 +184,14 @@ bsearch:
 		pr_debug("starting binary search, l %u r %u", l, r);
 
 		while (l + 1 < r) {
+			seq = list_entry(list->prev, struct journal_replay,
+					 list)->j.seq;
+
 			m = (l + r) >> 1;
+			read_bucket(m);
 
-			if (read_bucket(m))
+			if (seq != list_entry(list->prev, struct journal_replay,
+					      list)->j.seq)
 				l = m;
 			else
 				r = m;
@@ -300,7 +307,8 @@ int bch_journal_replay(struct cache_set *s, struct list_head *list,
 		for (k = i->j.start;
 		     k < end(&i->j);
 		     k = bkey_next(k)) {
-			pr_debug("%s", pkey(k));
+			trace_bcache_journal_replay_key(k);
+
 			bkey_copy(op->keys.top, k);
 			bch_keylist_push(&op->keys);
 
@@ -384,7 +392,7 @@ out:
 		return;
 found:
 	if (btree_node_dirty(best))
-		bch_btree_write(best, true, NULL);
+		bch_btree_node_write(best, NULL);
 	rw_unlock(true, best);
 }
 
@@ -617,7 +625,7 @@ static void journal_write_unlocked(struct closure *cl)
 		bio_reset(bio);
 		bio->bi_sector	= PTR_OFFSET(k, i);
 		bio->bi_bdev	= ca->bdev;
-		bio->bi_rw	= REQ_WRITE|REQ_SYNC|REQ_META|REQ_FLUSH;
+		bio->bi_rw	= REQ_WRITE|REQ_SYNC|REQ_META|REQ_FLUSH|REQ_FUA;
 		bio->bi_size	= sectors << 9;
 
 		bio->bi_end_io	= journal_write_endio;
@@ -712,7 +720,8 @@ void bch_journal(struct closure *cl)
 	spin_lock(&c->journal.lock);
 
 	if (journal_full(&c->journal)) {
-		/* XXX: tracepoint */
+		trace_bcache_journal_full(c);
+
 		closure_wait(&c->journal.wait, cl);
 
 		journal_reclaim(c);
@@ -728,13 +737,15 @@ void bch_journal(struct closure *cl)
 
 	if (b * c->sb.block_size > PAGE_SECTORS << JSET_BITS ||
 	    b > c->journal.blocks_free) {
-		/* XXX: If we were inserting so many keys that they won't fit in
+		trace_bcache_journal_entry_full(c);
+
+		/*
+		 * XXX: If we were inserting so many keys that they won't fit in
 		 * an _empty_ journal write, we'll deadlock. For now, handle
 		 * this in bch_keylist_realloc() - but something to think about.
 		 */
 		BUG_ON(!w->data->keys);
 
-		/* XXX: tracepoint */
 		BUG_ON(!closure_wait(&w->wait, cl));
 
 		closure_flush(&c->journal.io);
diff --git a/drivers/md/bcache/movinggc.c b/drivers/md/bcache/movinggc.c
index 8589512c972e..1a3b4f4786c3 100644
--- a/drivers/md/bcache/movinggc.c
+++ b/drivers/md/bcache/movinggc.c
@@ -9,6 +9,8 @@
 #include "debug.h"
 #include "request.h"
 
+#include <trace/events/bcache.h>
+
 struct moving_io {
 	struct keybuf_key	*w;
 	struct search		s;
@@ -44,14 +46,14 @@ static void write_moving_finish(struct closure *cl)
 {
 	struct moving_io *io = container_of(cl, struct moving_io, s.cl);
 	struct bio *bio = &io->bio.bio;
-	struct bio_vec *bv = bio_iovec_idx(bio, bio->bi_vcnt);
+	struct bio_vec *bv;
+	int i;
 
-	while (bv-- != bio->bi_io_vec)
+	bio_for_each_segment_all(bv, bio, i)
 		__free_page(bv->bv_page);
 
-	pr_debug("%s %s", io->s.op.insert_collision
-		 ? "collision moving" : "moved",
-		 pkey(&io->w->key));
+	if (io->s.op.insert_collision)
+		trace_bcache_gc_copy_collision(&io->w->key);
 
 	bch_keybuf_del(&io->s.op.c->moving_gc_keys, io->w);
 
@@ -94,8 +96,6 @@ static void write_moving(struct closure *cl)
 	struct moving_io *io = container_of(s, struct moving_io, s);
 
 	if (!s->error) {
-		trace_bcache_write_moving(&io->bio.bio);
-
 		moving_init(io);
 
 		io->bio.bio.bi_sector	= KEY_START(&io->w->key);
@@ -122,7 +122,6 @@ static void read_moving_submit(struct closure *cl)
 	struct moving_io *io = container_of(s, struct moving_io, s);
 	struct bio *bio = &io->bio.bio;
 
-	trace_bcache_read_moving(bio);
 	bch_submit_bbio(bio, s->op.c, &io->w->key, 0);
 
 	continue_at(cl, write_moving, bch_gc_wq);
@@ -138,7 +137,8 @@ static void read_moving(struct closure *cl)
 	/* XXX: if we error, background writeback could stall indefinitely */
 
 	while (!test_bit(CACHE_SET_STOPPING, &c->flags)) {
-		w = bch_keybuf_next_rescan(c, &c->moving_gc_keys, &MAX_KEY);
+		w = bch_keybuf_next_rescan(c, &c->moving_gc_keys,
+					   &MAX_KEY, moving_pred);
 		if (!w)
 			break;
 
@@ -159,10 +159,10 @@ static void read_moving(struct closure *cl)
 		bio->bi_rw	= READ;
 		bio->bi_end_io	= read_moving_endio;
 
-		if (bch_bio_alloc_pages(bio, GFP_KERNEL))
+		if (bio_alloc_pages(bio, GFP_KERNEL))
 			goto err;
 
-		pr_debug("%s", pkey(&w->key));
+		trace_bcache_gc_copy(&w->key);
 
 		closure_call(&io->s.cl, read_moving_submit, NULL, &c->gc.cl);
 
@@ -250,5 +250,5 @@ void bch_moving_gc(struct closure *cl)
 
 void bch_moving_init_cache_set(struct cache_set *c)
 {
-	bch_keybuf_init(&c->moving_gc_keys, moving_pred);
+	bch_keybuf_init(&c->moving_gc_keys);
 }
diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index e5ff12e52d5b..786a1a4f74d8 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -10,6 +10,7 @@
 #include "btree.h"
 #include "debug.h"
 #include "request.h"
+#include "writeback.h"
 
 #include <linux/cgroup.h>
 #include <linux/module.h>
@@ -21,8 +22,6 @@
 
 #define CUTOFF_CACHE_ADD	95
 #define CUTOFF_CACHE_READA	90
-#define CUTOFF_WRITEBACK	50
-#define CUTOFF_WRITEBACK_SYNC	75
 
 struct kmem_cache *bch_search_cache;
 
@@ -489,6 +488,12 @@ static void bch_insert_data_loop(struct closure *cl)
 		bch_queue_gc(op->c);
 	}
 
+	/*
+	 * Journal writes are marked REQ_FLUSH; if the original write was a
+	 * flush, it'll wait on the journal write.
+	 */
+	bio->bi_rw &= ~(REQ_FLUSH|REQ_FUA);
+
 	do {
 		unsigned i;
 		struct bkey *k;
@@ -510,10 +515,6 @@ static void bch_insert_data_loop(struct closure *cl)
 			goto err;
 
 		n = bch_bio_split(bio, KEY_SIZE(k), GFP_NOIO, split);
-		if (!n) {
-			__bkey_put(op->c, k);
-			continue_at(cl, bch_insert_data_loop, bcache_wq);
-		}
 
 		n->bi_end_io	= bch_insert_data_endio;
 		n->bi_private	= cl;
@@ -530,10 +531,9 @@ static void bch_insert_data_loop(struct closure *cl)
 		if (KEY_CSUM(k))
 			bio_csum(n, k);
 
-		pr_debug("%s", pkey(k));
+		trace_bcache_cache_insert(k);
 		bch_keylist_push(&op->keys);
 
-		trace_bcache_cache_insert(n, n->bi_sector, n->bi_bdev);
 		n->bi_rw |= REQ_WRITE;
 		bch_submit_bbio(n, op->c, k, 0);
 	} while (n != bio);
@@ -716,7 +716,7 @@ static struct search *search_alloc(struct bio *bio, struct bcache_device *d)
 	s->task			= current;
 	s->orig_bio		= bio;
 	s->write		= (bio->bi_rw & REQ_WRITE) != 0;
-	s->op.flush_journal	= (bio->bi_rw & REQ_FLUSH) != 0;
+	s->op.flush_journal	= (bio->bi_rw & (REQ_FLUSH|REQ_FUA)) != 0;
 	s->op.skip		= (bio->bi_rw & REQ_DISCARD) != 0;
 	s->recoverable		= 1;
 	s->start_time		= jiffies;
@@ -784,11 +784,8 @@ static void request_read_error(struct closure *cl)
 	int i;
 
 	if (s->recoverable) {
-		/* The cache read failed, but we can retry from the backing
-		 * device.
-		 */
-		pr_debug("recovering at sector %llu",
-			 (uint64_t) s->orig_bio->bi_sector);
+		/* Retry from the backing device: */
+		trace_bcache_read_retry(s->orig_bio);
 
 		s->error = 0;
 		bv = s->bio.bio.bi_io_vec;
@@ -806,7 +803,6 @@ static void request_read_error(struct closure *cl)
 
 		/* XXX: invalidate cache */
 
-		trace_bcache_read_retry(&s->bio.bio);
 		closure_bio_submit(&s->bio.bio, &s->cl, s->d);
 	}
 
@@ -827,53 +823,13 @@ static void request_read_done(struct closure *cl)
 	 */
 
 	if (s->op.cache_bio) {
-		struct bio_vec *src, *dst;
-		unsigned src_offset, dst_offset, bytes;
-		void *dst_ptr;
-
 		bio_reset(s->op.cache_bio);
 		s->op.cache_bio->bi_sector	= s->cache_miss->bi_sector;
 		s->op.cache_bio->bi_bdev	= s->cache_miss->bi_bdev;
 		s->op.cache_bio->bi_size	= s->cache_bio_sectors << 9;
 		bch_bio_map(s->op.cache_bio, NULL);
 
-		src = bio_iovec(s->op.cache_bio);
-		dst = bio_iovec(s->cache_miss);
-		src_offset = src->bv_offset;
-		dst_offset = dst->bv_offset;
-		dst_ptr = kmap(dst->bv_page);
-
-		while (1) {
-			if (dst_offset == dst->bv_offset + dst->bv_len) {
-				kunmap(dst->bv_page);
-				dst++;
-				if (dst == bio_iovec_idx(s->cache_miss,
-						s->cache_miss->bi_vcnt))
-					break;
-
-				dst_offset = dst->bv_offset;
-				dst_ptr = kmap(dst->bv_page);
-			}
-
-			if (src_offset == src->bv_offset + src->bv_len) {
-				src++;
-				if (src == bio_iovec_idx(s->op.cache_bio,
-						 s->op.cache_bio->bi_vcnt))
-					BUG();
-
-				src_offset = src->bv_offset;
-			}
-
-			bytes = min(dst->bv_offset + dst->bv_len - dst_offset,
-				    src->bv_offset + src->bv_len - src_offset);
-
-			memcpy(dst_ptr + dst_offset,
-			       page_address(src->bv_page) + src_offset,
-			       bytes);
-
-			src_offset	+= bytes;
-			dst_offset	+= bytes;
-		}
+		bio_copy_data(s->cache_miss, s->op.cache_bio);
 
 		bio_put(s->cache_miss);
 		s->cache_miss = NULL;
@@ -899,6 +855,7 @@ static void request_read_done_bh(struct closure *cl)
 	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
 
 	bch_mark_cache_accounting(s, !s->cache_miss, s->op.skip);
+	trace_bcache_read(s->orig_bio, !s->cache_miss, s->op.skip);
 
 	if (s->error)
 		continue_at_nobarrier(cl, request_read_error, bcache_wq);
@@ -917,9 +874,6 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s,
 	struct bio *miss;
 
 	miss = bch_bio_split(bio, sectors, GFP_NOIO, s->d->bio_split);
-	if (!miss)
-		return -EAGAIN;
-
 	if (miss == bio)
 		s->op.lookup_done = true;
 
@@ -938,8 +892,9 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s,
 		reada = min(dc->readahead >> 9,
 			    sectors - bio_sectors(miss));
 
-		if (bio_end(miss) + reada > bdev_sectors(miss->bi_bdev))
-			reada = bdev_sectors(miss->bi_bdev) - bio_end(miss);
+		if (bio_end_sector(miss) + reada > bdev_sectors(miss->bi_bdev))
+			reada = bdev_sectors(miss->bi_bdev) -
+				bio_end_sector(miss);
 	}
 
 	s->cache_bio_sectors = bio_sectors(miss) + reada;
@@ -963,13 +918,12 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s,
 		goto out_put;
 
 	bch_bio_map(s->op.cache_bio, NULL);
-	if (bch_bio_alloc_pages(s->op.cache_bio, __GFP_NOWARN|GFP_NOIO))
+	if (bio_alloc_pages(s->op.cache_bio, __GFP_NOWARN|GFP_NOIO))
 		goto out_put;
 
 	s->cache_miss = miss;
 	bio_get(s->op.cache_bio);
 
-	trace_bcache_cache_miss(s->orig_bio);
 	closure_bio_submit(s->op.cache_bio, &s->cl, s->d);
 
 	return ret;
@@ -1002,24 +956,13 @@ static void cached_dev_write_complete(struct closure *cl)
 	cached_dev_bio_complete(cl);
 }
 
-static bool should_writeback(struct cached_dev *dc, struct bio *bio)
-{
-	unsigned threshold = (bio->bi_rw & REQ_SYNC)
-		? CUTOFF_WRITEBACK_SYNC
-		: CUTOFF_WRITEBACK;
-
-	return !atomic_read(&dc->disk.detaching) &&
-		cache_mode(dc, bio) == CACHE_MODE_WRITEBACK &&
-		dc->disk.c->gc_stats.in_use < threshold;
-}
-
 static void request_write(struct cached_dev *dc, struct search *s)
 {
 	struct closure *cl = &s->cl;
 	struct bio *bio = &s->bio.bio;
 	struct bkey start, end;
 	start = KEY(dc->disk.id, bio->bi_sector, 0);
-	end = KEY(dc->disk.id, bio_end(bio), 0);
+	end = KEY(dc->disk.id, bio_end_sector(bio), 0);
 
 	bch_keybuf_check_overlapping(&s->op.c->moving_gc_keys, &start, &end);
 
@@ -1034,22 +977,37 @@ static void request_write(struct cached_dev *dc, struct search *s)
 	if (bio->bi_rw & REQ_DISCARD)
 		goto skip;
 
+	if (should_writeback(dc, s->orig_bio,
+			     cache_mode(dc, bio),
+			     s->op.skip)) {
+		s->op.skip = false;
+		s->writeback = true;
+	}
+
 	if (s->op.skip)
 		goto skip;
 
-	if (should_writeback(dc, s->orig_bio))
-		s->writeback = true;
+	trace_bcache_write(s->orig_bio, s->writeback, s->op.skip);
 
 	if (!s->writeback) {
 		s->op.cache_bio = bio_clone_bioset(bio, GFP_NOIO,
 						   dc->disk.bio_split);
 
-		trace_bcache_writethrough(s->orig_bio);
 		closure_bio_submit(bio, cl, s->d);
 	} else {
-		s->op.cache_bio = bio;
-		trace_bcache_writeback(s->orig_bio);
-		bch_writeback_add(dc, bio_sectors(bio));
+		bch_writeback_add(dc);
+
+		if (s->op.flush_journal) {
+			/* Also need to send a flush to the backing device */
+			s->op.cache_bio = bio_clone_bioset(bio, GFP_NOIO,
+							   dc->disk.bio_split);
+
+			bio->bi_size = 0;
+			bio->bi_vcnt = 0;
+			closure_bio_submit(bio, cl, s->d);
+		} else {
+			s->op.cache_bio = bio;
+		}
 	}
 out:
 	closure_call(&s->op.cl, bch_insert_data, NULL, cl);
@@ -1058,7 +1016,6 @@ skip:
 	s->op.skip = true;
 	s->op.cache_bio = s->orig_bio;
 	bio_get(s->op.cache_bio);
-	trace_bcache_write_skip(s->orig_bio);
 
 	if ((bio->bi_rw & REQ_DISCARD) &&
 	    !blk_queue_discard(bdev_get_queue(dc->bdev)))
@@ -1088,9 +1045,10 @@ static void request_nodata(struct cached_dev *dc, struct search *s)
 
 /* Cached devices - read & write stuff */
 
-int bch_get_congested(struct cache_set *c)
+unsigned bch_get_congested(struct cache_set *c)
 {
 	int i;
+	long rand;
 
 	if (!c->congested_read_threshold_us &&
 	    !c->congested_write_threshold_us)
@@ -1106,7 +1064,13 @@ int bch_get_congested(struct cache_set *c)
 
 	i += CONGESTED_MAX;
 
-	return i <= 0 ? 1 : fract_exp_two(i, 6);
+	if (i > 0)
+		i = fract_exp_two(i, 6);
+
+	rand = get_random_int();
+	i -= bitmap_weight(&rand, BITS_PER_LONG);
+
+	return i > 0 ? i : 1;
 }
 
 static void add_sequential(struct task_struct *t)
@@ -1126,10 +1090,8 @@ static void check_should_skip(struct cached_dev *dc, struct search *s)
 {
 	struct cache_set *c = s->op.c;
 	struct bio *bio = &s->bio.bio;
-
-	long rand;
-	int cutoff = bch_get_congested(c);
 	unsigned mode = cache_mode(dc, bio);
+	unsigned sectors, congested = bch_get_congested(c);
 
 	if (atomic_read(&dc->disk.detaching) ||
 	    c->gc_stats.in_use > CUTOFF_CACHE_ADD ||
@@ -1147,17 +1109,14 @@ static void check_should_skip(struct cached_dev *dc, struct search *s)
 		goto skip;
 	}
 
-	if (!cutoff) {
-		cutoff = dc->sequential_cutoff >> 9;
+	if (!congested && !dc->sequential_cutoff)
+		goto rescale;
 
-		if (!cutoff)
-			goto rescale;
-
-		if (mode == CACHE_MODE_WRITEBACK &&
-		    (bio->bi_rw & REQ_WRITE) &&
-		    (bio->bi_rw & REQ_SYNC))
-			goto rescale;
-	}
+	if (!congested &&
+	    mode == CACHE_MODE_WRITEBACK &&
+	    (bio->bi_rw & REQ_WRITE) &&
+	    (bio->bi_rw & REQ_SYNC))
+		goto rescale;
 
 	if (dc->sequential_merge) {
 		struct io *i;
@@ -1177,7 +1136,7 @@ found:
 		if (i->sequential + bio->bi_size > i->sequential)
 			i->sequential	+= bio->bi_size;
 
-		i->last			 = bio_end(bio);
+		i->last			 = bio_end_sector(bio);
 		i->jiffies		 = jiffies + msecs_to_jiffies(5000);
 		s->task->sequential_io	 = i->sequential;
 
@@ -1192,12 +1151,19 @@ found:
 		add_sequential(s->task);
 	}
 
-	rand = get_random_int();
-	cutoff -= bitmap_weight(&rand, BITS_PER_LONG);
+	sectors = max(s->task->sequential_io,
+		      s->task->sequential_io_avg) >> 9;
 
-	if (cutoff <= (int) (max(s->task->sequential_io,
-				 s->task->sequential_io_avg) >> 9))
+	if (dc->sequential_cutoff &&
+	    sectors >= dc->sequential_cutoff >> 9) {
+		trace_bcache_bypass_sequential(s->orig_bio);
 		goto skip;
+	}
+
+	if (congested && sectors >= congested) {
+		trace_bcache_bypass_congested(s->orig_bio);
+		goto skip;
+	}
 
 rescale:
 	bch_rescale_priorities(c, bio_sectors(bio));
@@ -1288,30 +1254,25 @@ void bch_cached_dev_request_init(struct cached_dev *dc)
 static int flash_dev_cache_miss(struct btree *b, struct search *s,
 				struct bio *bio, unsigned sectors)
 {
+	struct bio_vec *bv;
+	int i;
+
 	/* Zero fill bio */
 
-	while (bio->bi_idx != bio->bi_vcnt) {
-		struct bio_vec *bv = bio_iovec(bio);
+	bio_for_each_segment(bv, bio, i) {
 		unsigned j = min(bv->bv_len >> 9, sectors);
 
 		void *p = kmap(bv->bv_page);
 		memset(p + bv->bv_offset, 0, j << 9);
 		kunmap(bv->bv_page);
 
-		bv->bv_len	-= j << 9;
-		bv->bv_offset	+= j << 9;
-
-		if (bv->bv_len)
-			return 0;
-
-		bio->bi_sector	+= j;
-		bio->bi_size	-= j << 9;
-
-		bio->bi_idx++;
-		sectors		-= j;
+		sectors	-= j;
 	}
 
-	s->op.lookup_done = true;
+	bio_advance(bio, min(sectors << 9, bio->bi_size));
+
+	if (!bio->bi_size)
+		s->op.lookup_done = true;
 
 	return 0;
 }
@@ -1338,8 +1299,8 @@ static void flash_dev_make_request(struct request_queue *q, struct bio *bio)
 		closure_call(&s->op.cl, btree_read_async, NULL, cl);
 	} else if (bio_has_data(bio) || s->op.skip) {
 		bch_keybuf_check_overlapping(&s->op.c->moving_gc_keys,
-					     &KEY(d->id, bio->bi_sector, 0),
-					     &KEY(d->id, bio_end(bio), 0));
+					&KEY(d->id, bio->bi_sector, 0),
+					&KEY(d->id, bio_end_sector(bio), 0));
 
 		s->writeback	= true;
 		s->op.cache_bio	= bio;
diff --git a/drivers/md/bcache/request.h b/drivers/md/bcache/request.h
index 254d9ab5707c..57dc4784f4f4 100644
--- a/drivers/md/bcache/request.h
+++ b/drivers/md/bcache/request.h
@@ -30,7 +30,7 @@ struct search {
 };
 
 void bch_cache_read_endio(struct bio *, int);
-int bch_get_congested(struct cache_set *);
+unsigned bch_get_congested(struct cache_set *);
 void bch_insert_data(struct closure *cl);
 void bch_btree_insert_async(struct closure *);
 void bch_cache_read_endio(struct bio *, int);
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index f88e2b653a3f..547c4c57b052 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -10,10 +10,13 @@
 #include "btree.h"
 #include "debug.h"
 #include "request.h"
+#include "writeback.h"
 
+#include <linux/blkdev.h>
 #include <linux/buffer_head.h>
 #include <linux/debugfs.h>
 #include <linux/genhd.h>
+#include <linux/kthread.h>
 #include <linux/module.h>
 #include <linux/random.h>
 #include <linux/reboot.h>
@@ -342,6 +345,7 @@ static void uuid_io(struct cache_set *c, unsigned long rw,
 	struct closure *cl = &c->uuid_write.cl;
 	struct uuid_entry *u;
 	unsigned i;
+	char buf[80];
 
 	BUG_ON(!parent);
 	closure_lock(&c->uuid_write, parent);
@@ -362,8 +366,8 @@ static void uuid_io(struct cache_set *c, unsigned long rw,
 			break;
 	}
 
-	pr_debug("%s UUIDs at %s", rw & REQ_WRITE ? "wrote" : "read",
-		 pkey(&c->uuid_bucket));
+	bch_bkey_to_text(buf, sizeof(buf), k);
+	pr_debug("%s UUIDs at %s", rw & REQ_WRITE ? "wrote" : "read", buf);
 
 	for (u = c->uuids; u < c->uuids + c->nr_uuids; u++)
 		if (!bch_is_zero(u->uuid, 16))
@@ -543,7 +547,6 @@ void bch_prio_write(struct cache *ca)
 
 	pr_debug("free %zu, free_inc %zu, unused %zu", fifo_used(&ca->free),
 		 fifo_used(&ca->free_inc), fifo_used(&ca->unused));
-	blktrace_msg(ca, "Starting priorities: " buckets_free(ca));
 
 	for (i = prio_buckets(ca) - 1; i >= 0; --i) {
 		long bucket;
@@ -704,7 +707,8 @@ static void bcache_device_detach(struct bcache_device *d)
 		atomic_set(&d->detaching, 0);
 	}
 
-	bcache_device_unlink(d);
+	if (!d->flush_done)
+		bcache_device_unlink(d);
 
 	d->c->devices[d->id] = NULL;
 	closure_put(&d->c->caching);
@@ -743,13 +747,35 @@ static void bcache_device_free(struct bcache_device *d)
 		mempool_destroy(d->unaligned_bvec);
 	if (d->bio_split)
 		bioset_free(d->bio_split);
+	if (is_vmalloc_addr(d->stripe_sectors_dirty))
+		vfree(d->stripe_sectors_dirty);
+	else
+		kfree(d->stripe_sectors_dirty);
 
 	closure_debug_destroy(&d->cl);
 }
 
-static int bcache_device_init(struct bcache_device *d, unsigned block_size)
+static int bcache_device_init(struct bcache_device *d, unsigned block_size,
+			      sector_t sectors)
 {
 	struct request_queue *q;
+	size_t n;
+
+	if (!d->stripe_size_bits)
+		d->stripe_size_bits = 31;
+
+	d->nr_stripes = round_up(sectors, 1 << d->stripe_size_bits) >>
+		d->stripe_size_bits;
+
+	if (!d->nr_stripes || d->nr_stripes > SIZE_MAX / sizeof(atomic_t))
+		return -ENOMEM;
+
+	n = d->nr_stripes * sizeof(atomic_t);
+	d->stripe_sectors_dirty = n < PAGE_SIZE << 6
+		? kzalloc(n, GFP_KERNEL)
+		: vzalloc(n);
+	if (!d->stripe_sectors_dirty)
+		return -ENOMEM;
 
 	if (!(d->bio_split = bioset_create(4, offsetof(struct bbio, bio))) ||
 	    !(d->unaligned_bvec = mempool_create_kmalloc_pool(1,
@@ -759,6 +785,7 @@ static int bcache_device_init(struct bcache_device *d, unsigned block_size)
 	    !(q = blk_alloc_queue(GFP_KERNEL)))
 		return -ENOMEM;
 
+	set_capacity(d->disk, sectors);
 	snprintf(d->disk->disk_name, DISK_NAME_LEN, "bcache%i", bcache_minor);
 
 	d->disk->major		= bcache_major;
@@ -781,6 +808,8 @@ static int bcache_device_init(struct bcache_device *d, unsigned block_size)
 	set_bit(QUEUE_FLAG_NONROT,	&d->disk->queue->queue_flags);
 	set_bit(QUEUE_FLAG_DISCARD,	&d->disk->queue->queue_flags);
 
+	blk_queue_flush(q, REQ_FLUSH|REQ_FUA);
+
 	return 0;
 }
 
@@ -800,6 +829,17 @@ static void calc_cached_dev_sectors(struct cache_set *c)
 void bch_cached_dev_run(struct cached_dev *dc)
 {
 	struct bcache_device *d = &dc->disk;
+	char buf[SB_LABEL_SIZE + 1];
+	char *env[] = {
+		"DRIVER=bcache",
+		kasprintf(GFP_KERNEL, "CACHED_UUID=%pU", dc->sb.uuid),
+		NULL,
+		NULL,
+	};
+
+	memcpy(buf, dc->sb.label, SB_LABEL_SIZE);
+	buf[SB_LABEL_SIZE] = '\0';
+	env[2] = kasprintf(GFP_KERNEL, "CACHED_LABEL=%s", buf);
 
 	if (atomic_xchg(&dc->running, 1))
 		return;
@@ -816,10 +856,12 @@ void bch_cached_dev_run(struct cached_dev *dc)
 
 	add_disk(d->disk);
 	bd_link_disk_holder(dc->bdev, dc->disk.disk);
-#if 0
-	char *env[] = { "SYMLINK=label" , NULL };
+	/* won't show up in the uevent file, use udevadm monitor -e instead
+	 * only class / kset properties are persistent */
 	kobject_uevent_env(&disk_to_dev(d->disk)->kobj, KOBJ_CHANGE, env);
-#endif
+	kfree(env[1]);
+	kfree(env[2]);
+
 	if (sysfs_create_link(&d->kobj, &disk_to_dev(d->disk)->kobj, "dev") ||
 	    sysfs_create_link(&disk_to_dev(d->disk)->kobj, &d->kobj, "bcache"))
 		pr_debug("error creating sysfs link");
@@ -960,6 +1002,7 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c)
 	atomic_set(&dc->count, 1);
 
 	if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) {
+		bch_sectors_dirty_init(dc);
 		atomic_set(&dc->has_dirty, 1);
 		atomic_inc(&dc->count);
 		bch_writeback_queue(dc);
@@ -1014,6 +1057,14 @@ static void cached_dev_flush(struct closure *cl)
 	struct cached_dev *dc = container_of(cl, struct cached_dev, disk.cl);
 	struct bcache_device *d = &dc->disk;
 
+	mutex_lock(&bch_register_lock);
+	d->flush_done = 1;
+
+	if (d->c)
+		bcache_device_unlink(d);
+
+	mutex_unlock(&bch_register_lock);
+
 	bch_cache_accounting_destroy(&dc->accounting);
 	kobject_del(&d->kobj);
 
@@ -1045,7 +1096,8 @@ static int cached_dev_init(struct cached_dev *dc, unsigned block_size)
 		hlist_add_head(&io->hash, dc->io_hash + RECENT_IO);
 	}
 
-	ret = bcache_device_init(&dc->disk, block_size);
+	ret = bcache_device_init(&dc->disk, block_size,
+			 dc->bdev->bd_part->nr_sects - dc->sb.data_offset);
 	if (ret)
 		return ret;
 
@@ -1144,11 +1196,10 @@ static int flash_dev_run(struct cache_set *c, struct uuid_entry *u)
 
 	kobject_init(&d->kobj, &bch_flash_dev_ktype);
 
-	if (bcache_device_init(d, block_bytes(c)))
+	if (bcache_device_init(d, block_bytes(c), u->sectors))
 		goto err;
 
 	bcache_device_attach(d, c, u - c->uuids);
-	set_capacity(d->disk, u->sectors);
 	bch_flash_dev_request_init(d);
 	add_disk(d->disk);
 
@@ -1255,9 +1306,10 @@ static void cache_set_free(struct closure *cl)
 	free_pages((unsigned long) c->uuids, ilog2(bucket_pages(c)));
 	free_pages((unsigned long) c->sort, ilog2(bucket_pages(c)));
 
-	kfree(c->fill_iter);
 	if (c->bio_split)
 		bioset_free(c->bio_split);
+	if (c->fill_iter)
+		mempool_destroy(c->fill_iter);
 	if (c->bio_meta)
 		mempool_destroy(c->bio_meta);
 	if (c->search)
@@ -1278,11 +1330,9 @@ static void cache_set_free(struct closure *cl)
 static void cache_set_flush(struct closure *cl)
 {
 	struct cache_set *c = container_of(cl, struct cache_set, caching);
+	struct cache *ca;
 	struct btree *b;
-
-	/* Shut down allocator threads */
-	set_bit(CACHE_SET_STOPPING_2, &c->flags);
-	wake_up(&c->alloc_wait);
+	unsigned i;
 
 	bch_cache_accounting_destroy(&c->accounting);
 
@@ -1295,7 +1345,11 @@ static void cache_set_flush(struct closure *cl)
 	/* Should skip this if we're unregistering because of an error */
 	list_for_each_entry(b, &c->btree_cache, list)
 		if (btree_node_dirty(b))
-			bch_btree_write(b, true, NULL);
+			bch_btree_node_write(b, NULL);
+
+	for_each_cache(ca, c, i)
+		if (ca->alloc_thread)
+			kthread_stop(ca->alloc_thread);
 
 	closure_return(cl);
 }
@@ -1303,18 +1357,22 @@ static void cache_set_flush(struct closure *cl)
 static void __cache_set_unregister(struct closure *cl)
 {
 	struct cache_set *c = container_of(cl, struct cache_set, caching);
-	struct cached_dev *dc, *t;
+	struct cached_dev *dc;
 	size_t i;
 
 	mutex_lock(&bch_register_lock);
 
-	if (test_bit(CACHE_SET_UNREGISTERING, &c->flags))
-		list_for_each_entry_safe(dc, t, &c->cached_devs, list)
-			bch_cached_dev_detach(dc);
-
 	for (i = 0; i < c->nr_uuids; i++)
-		if (c->devices[i] && UUID_FLASH_ONLY(&c->uuids[i]))
-			bcache_device_stop(c->devices[i]);
+		if (c->devices[i]) {
+			if (!UUID_FLASH_ONLY(&c->uuids[i]) &&
+			    test_bit(CACHE_SET_UNREGISTERING, &c->flags)) {
+				dc = container_of(c->devices[i],
+						  struct cached_dev, disk);
+				bch_cached_dev_detach(dc);
+			} else {
+				bcache_device_stop(c->devices[i]);
+			}
+		}
 
 	mutex_unlock(&bch_register_lock);
 
@@ -1373,9 +1431,9 @@ struct cache_set *bch_cache_set_alloc(struct cache_sb *sb)
 		c->btree_pages = max_t(int, c->btree_pages / 4,
 				       BTREE_MAX_PAGES);
 
-	init_waitqueue_head(&c->alloc_wait);
+	c->sort_crit_factor = int_sqrt(c->btree_pages);
+
 	mutex_init(&c->bucket_lock);
-	mutex_init(&c->fill_lock);
 	mutex_init(&c->sort_lock);
 	spin_lock_init(&c->sort_time_lock);
 	closure_init_unlocked(&c->sb_write);
@@ -1401,8 +1459,8 @@ struct cache_set *bch_cache_set_alloc(struct cache_sb *sb)
 	    !(c->bio_meta = mempool_create_kmalloc_pool(2,
 				sizeof(struct bbio) + sizeof(struct bio_vec) *
 				bucket_pages(c))) ||
+	    !(c->fill_iter = mempool_create_kmalloc_pool(1, iter_size)) ||
 	    !(c->bio_split = bioset_create(4, offsetof(struct bbio, bio))) ||
-	    !(c->fill_iter = kmalloc(iter_size, GFP_KERNEL)) ||
 	    !(c->sort = alloc_bucket_pages(GFP_KERNEL, c)) ||
 	    !(c->uuids = alloc_bucket_pages(GFP_KERNEL, c)) ||
 	    bch_journal_alloc(c) ||
@@ -1410,8 +1468,6 @@ struct cache_set *bch_cache_set_alloc(struct cache_sb *sb)
 	    bch_open_buckets_alloc(c))
 		goto err;
 
-	c->fill_iter->size = sb->bucket_size / sb->block_size;
-
 	c->congested_read_threshold_us	= 2000;
 	c->congested_write_threshold_us	= 20000;
 	c->error_limit	= 8 << IO_ERROR_SHIFT;
@@ -1496,9 +1552,10 @@ static void run_cache_set(struct cache_set *c)
 		 */
 		bch_journal_next(&c->journal);
 
+		err = "error starting allocator thread";
 		for_each_cache(ca, c, i)
-			closure_call(&ca->alloc, bch_allocator_thread,
-				     system_wq, &c->cl);
+			if (bch_cache_allocator_start(ca))
+				goto err;
 
 		/*
 		 * First place it's safe to allocate: btree_check() and
@@ -1531,17 +1588,16 @@ static void run_cache_set(struct cache_set *c)
 
 		bch_btree_gc_finish(c);
 
+		err = "error starting allocator thread";
 		for_each_cache(ca, c, i)
-			closure_call(&ca->alloc, bch_allocator_thread,
-				     ca->alloc_workqueue, &c->cl);
+			if (bch_cache_allocator_start(ca))
+				goto err;
 
 		mutex_lock(&c->bucket_lock);
 		for_each_cache(ca, c, i)
 			bch_prio_write(ca);
 		mutex_unlock(&c->bucket_lock);
 
-		wake_up(&c->alloc_wait);
-
 		err = "cannot allocate new UUID bucket";
 		if (__uuid_write(c))
 			goto err_unlock_gc;
@@ -1552,7 +1608,7 @@ static void run_cache_set(struct cache_set *c)
 			goto err_unlock_gc;
 
 		bkey_copy_key(&c->root->key, &MAX_KEY);
-		bch_btree_write(c->root, true, &op);
+		bch_btree_node_write(c->root, &op.cl);
 
 		bch_btree_set_root(c->root);
 		rw_unlock(true, c->root);
@@ -1673,9 +1729,6 @@ void bch_cache_release(struct kobject *kobj)
 
 	bio_split_pool_free(&ca->bio_split_hook);
 
-	if (ca->alloc_workqueue)
-		destroy_workqueue(ca->alloc_workqueue);
-
 	free_pages((unsigned long) ca->disk_buckets, ilog2(bucket_pages(ca)));
 	kfree(ca->prio_buckets);
 	vfree(ca->buckets);
@@ -1723,7 +1776,6 @@ static int cache_alloc(struct cache_sb *sb, struct cache *ca)
 	    !(ca->prio_buckets	= kzalloc(sizeof(uint64_t) * prio_buckets(ca) *
 					  2, GFP_KERNEL)) ||
 	    !(ca->disk_buckets	= alloc_bucket_pages(GFP_KERNEL, ca)) ||
-	    !(ca->alloc_workqueue = alloc_workqueue("bch_allocator", 0, 1)) ||
 	    bio_split_pool_init(&ca->bio_split_hook))
 		return -ENOMEM;
 
@@ -1786,6 +1838,36 @@ static ssize_t register_bcache(struct kobject *, struct kobj_attribute *,
 kobj_attribute_write(register,		register_bcache);
 kobj_attribute_write(register_quiet,	register_bcache);
 
+static bool bch_is_open_backing(struct block_device *bdev) {
+	struct cache_set *c, *tc;
+	struct cached_dev *dc, *t;
+
+	list_for_each_entry_safe(c, tc, &bch_cache_sets, list)
+		list_for_each_entry_safe(dc, t, &c->cached_devs, list)
+			if (dc->bdev == bdev)
+				return true;
+	list_for_each_entry_safe(dc, t, &uncached_devices, list)
+		if (dc->bdev == bdev)
+			return true;
+	return false;
+}
+
+static bool bch_is_open_cache(struct block_device *bdev) {
+	struct cache_set *c, *tc;
+	struct cache *ca;
+	unsigned i;
+
+	list_for_each_entry_safe(c, tc, &bch_cache_sets, list)
+		for_each_cache(ca, c, i)
+			if (ca->bdev == bdev)
+				return true;
+	return false;
+}
+
+static bool bch_is_open(struct block_device *bdev) {
+	return bch_is_open_cache(bdev) || bch_is_open_backing(bdev);
+}
+
 static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 			       const char *buffer, size_t size)
 {
@@ -1810,8 +1892,13 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 				  FMODE_READ|FMODE_WRITE|FMODE_EXCL,
 				  sb);
 	if (IS_ERR(bdev)) {
-		if (bdev == ERR_PTR(-EBUSY))
-			err = "device busy";
+		if (bdev == ERR_PTR(-EBUSY)) {
+			bdev = lookup_bdev(strim(path));
+			if (!IS_ERR(bdev) && bch_is_open(bdev))
+				err = "device already registered";
+			else
+				err = "device busy";
+		}
 		goto err;
 	}
 
diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c
index 4d9cca47e4c6..12a2c2846f99 100644
--- a/drivers/md/bcache/sysfs.c
+++ b/drivers/md/bcache/sysfs.c
@@ -9,7 +9,9 @@
 #include "sysfs.h"
 #include "btree.h"
 #include "request.h"
+#include "writeback.h"
 
+#include <linux/blkdev.h>
 #include <linux/sort.h>
 
 static const char * const cache_replacement_policies[] = {
@@ -79,6 +81,9 @@ rw_attribute(writeback_rate_p_term_inverse);
 rw_attribute(writeback_rate_d_smooth);
 read_attribute(writeback_rate_debug);
 
+read_attribute(stripe_size);
+read_attribute(partial_stripes_expensive);
+
 rw_attribute(synchronous);
 rw_attribute(journal_delay_ms);
 rw_attribute(discard);
@@ -127,7 +132,7 @@ SHOW(__bch_cached_dev)
 		char derivative[20];
 		char target[20];
 		bch_hprint(dirty,
-		       atomic_long_read(&dc->disk.sectors_dirty) << 9);
+			   bcache_dev_sectors_dirty(&dc->disk) << 9);
 		bch_hprint(derivative,	dc->writeback_rate_derivative << 9);
 		bch_hprint(target,	dc->writeback_rate_target << 9);
 
@@ -143,7 +148,10 @@ SHOW(__bch_cached_dev)
 	}
 
 	sysfs_hprint(dirty_data,
-		     atomic_long_read(&dc->disk.sectors_dirty) << 9);
+		     bcache_dev_sectors_dirty(&dc->disk) << 9);
+
+	sysfs_hprint(stripe_size,	(1 << dc->disk.stripe_size_bits) << 9);
+	var_printf(partial_stripes_expensive,	"%u");
 
 	var_printf(sequential_merge,	"%i");
 	var_hprint(sequential_cutoff);
@@ -170,6 +178,7 @@ STORE(__cached_dev)
 					     disk.kobj);
 	unsigned v = size;
 	struct cache_set *c;
+	struct kobj_uevent_env *env;
 
 #define d_strtoul(var)		sysfs_strtoul(var, dc->var)
 #define d_strtoi_h(var)		sysfs_hatoi(var, dc->var)
@@ -214,6 +223,7 @@ STORE(__cached_dev)
 	}
 
 	if (attr == &sysfs_label) {
+		/* note: endlines are preserved */
 		memcpy(dc->sb.label, buf, SB_LABEL_SIZE);
 		bch_write_bdev_super(dc, NULL);
 		if (dc->disk.c) {
@@ -221,6 +231,15 @@ STORE(__cached_dev)
 			       buf, SB_LABEL_SIZE);
 			bch_uuid_write(dc->disk.c);
 		}
+		env = kzalloc(sizeof(struct kobj_uevent_env), GFP_KERNEL);
+		if (!env)
+			return -ENOMEM;
+		add_uevent_var(env, "DRIVER=bcache");
+		add_uevent_var(env, "CACHED_UUID=%pU", dc->sb.uuid),
+		add_uevent_var(env, "CACHED_LABEL=%s", buf);
+		kobject_uevent_env(
+			&disk_to_dev(dc->disk.disk)->kobj, KOBJ_CHANGE, env->envp);
+		kfree(env);
 	}
 
 	if (attr == &sysfs_attach) {
@@ -284,6 +303,8 @@ static struct attribute *bch_cached_dev_files[] = {
 	&sysfs_writeback_rate_d_smooth,
 	&sysfs_writeback_rate_debug,
 	&sysfs_dirty_data,
+	&sysfs_stripe_size,
+	&sysfs_partial_stripes_expensive,
 	&sysfs_sequential_cutoff,
 	&sysfs_sequential_merge,
 	&sysfs_clear_stats,
@@ -665,12 +686,10 @@ SHOW(__bch_cache)
 		int cmp(const void *l, const void *r)
 		{	return *((uint16_t *) r) - *((uint16_t *) l); }
 
-		/* Number of quantiles we compute */
-		const unsigned nq = 31;
-
 		size_t n = ca->sb.nbuckets, i, unused, btree;
 		uint64_t sum = 0;
-		uint16_t q[nq], *p, *cached;
+		/* Compute 31 quantiles */
+		uint16_t q[31], *p, *cached;
 		ssize_t ret;
 
 		cached = p = vmalloc(ca->sb.nbuckets * sizeof(uint16_t));
@@ -703,26 +722,29 @@ SHOW(__bch_cache)
 		if (n)
 			do_div(sum, n);
 
-		for (i = 0; i < nq; i++)
-			q[i] = INITIAL_PRIO - cached[n * (i + 1) / (nq + 1)];
+		for (i = 0; i < ARRAY_SIZE(q); i++)
+			q[i] = INITIAL_PRIO - cached[n * (i + 1) /
+				(ARRAY_SIZE(q) + 1)];
 
 		vfree(p);
 
-		ret = snprintf(buf, PAGE_SIZE,
-			       "Unused:		%zu%%\n"
-			       "Metadata:	%zu%%\n"
-			       "Average:	%llu\n"
-			       "Sectors per Q:	%zu\n"
-			       "Quantiles:	[",
-			       unused * 100 / (size_t) ca->sb.nbuckets,
-			       btree * 100 / (size_t) ca->sb.nbuckets, sum,
-			       n * ca->sb.bucket_size / (nq + 1));
-
-		for (i = 0; i < nq && ret < (ssize_t) PAGE_SIZE; i++)
-			ret += snprintf(buf + ret, PAGE_SIZE - ret,
-					i < nq - 1 ? "%u " : "%u]\n", q[i]);
-
-		buf[PAGE_SIZE - 1] = '\0';
+		ret = scnprintf(buf, PAGE_SIZE,
+				"Unused:		%zu%%\n"
+				"Metadata:	%zu%%\n"
+				"Average:	%llu\n"
+				"Sectors per Q:	%zu\n"
+				"Quantiles:	[",
+				unused * 100 / (size_t) ca->sb.nbuckets,
+				btree * 100 / (size_t) ca->sb.nbuckets, sum,
+				n * ca->sb.bucket_size / (ARRAY_SIZE(q) + 1));
+
+		for (i = 0; i < ARRAY_SIZE(q); i++)
+			ret += scnprintf(buf + ret, PAGE_SIZE - ret,
+					 "%u ", q[i]);
+		ret--;
+
+		ret += scnprintf(buf + ret, PAGE_SIZE - ret, "]\n");
+
 		return ret;
 	}
 
diff --git a/drivers/md/bcache/trace.c b/drivers/md/bcache/trace.c
index 983f9bb411bc..f7b6c197f90f 100644
--- a/drivers/md/bcache/trace.c
+++ b/drivers/md/bcache/trace.c
@@ -2,6 +2,7 @@
 #include "btree.h"
 #include "request.h"
 
+#include <linux/blktrace_api.h>
 #include <linux/module.h>
 
 #define CREATE_TRACE_POINTS
@@ -9,18 +10,44 @@
 
 EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_request_start);
 EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_request_end);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_passthrough);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_hit);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_miss);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_bypass_sequential);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_bypass_congested);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_read);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_write);
 EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_read_retry);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writethrough);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writeback);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_write_skip);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_insert);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_replay_key);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_write);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_full);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_entry_full);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_cache_cannibalize);
+
 EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_read);
 EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_write);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_write_dirty);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_read_dirty);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_journal_write);
-EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_cache_insert);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_alloc);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_alloc_fail);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_free);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_gc_coalesce);
 EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_start);
 EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_end);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_copy);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_gc_copy_collision);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_insert_key);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_split);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_node_compact);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_btree_set_root);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_alloc_invalidate);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_alloc_fail);
+
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writeback);
+EXPORT_TRACEPOINT_SYMBOL_GPL(bcache_writeback_collision);
diff --git a/drivers/md/bcache/util.c b/drivers/md/bcache/util.c
index da3a99e85b1e..98eb81159a22 100644
--- a/drivers/md/bcache/util.c
+++ b/drivers/md/bcache/util.c
@@ -228,23 +228,6 @@ start:		bv->bv_len	= min_t(size_t, PAGE_SIZE - bv->bv_offset,
 	}
 }
 
-int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp)
-{
-	int i;
-	struct bio_vec *bv;
-
-	bio_for_each_segment(bv, bio, i) {
-		bv->bv_page = alloc_page(gfp);
-		if (!bv->bv_page) {
-			while (bv-- != bio->bi_io_vec + bio->bi_idx)
-				__free_page(bv->bv_page);
-			return -ENOMEM;
-		}
-	}
-
-	return 0;
-}
-
 /*
  * Portions Copyright (c) 1996-2001, PostgreSQL Global Development Group (Any
  * use permitted, subject to terms of PostgreSQL license; see.)
diff --git a/drivers/md/bcache/util.h b/drivers/md/bcache/util.h
index 577393e38c3a..1ae2a73ad85f 100644
--- a/drivers/md/bcache/util.h
+++ b/drivers/md/bcache/util.h
@@ -15,8 +15,6 @@
 
 struct closure;
 
-#include <trace/events/bcache.h>
-
 #ifdef CONFIG_BCACHE_EDEBUG
 
 #define atomic_dec_bug(v)	BUG_ON(atomic_dec_return(v) < 0)
@@ -566,12 +564,8 @@ static inline unsigned fract_exp_two(unsigned x, unsigned fract_bits)
 	return x;
 }
 
-#define bio_end(bio)	((bio)->bi_sector + bio_sectors(bio))
-
 void bch_bio_map(struct bio *bio, void *base);
 
-int bch_bio_alloc_pages(struct bio *bio, gfp_t gfp);
-
 static inline sector_t bdev_sectors(struct block_device *bdev)
 {
 	return bdev->bd_inode->i_size >> 9;
diff --git a/drivers/md/bcache/writeback.c b/drivers/md/bcache/writeback.c
index 2714ed3991d1..22cbff551628 100644
--- a/drivers/md/bcache/writeback.c
+++ b/drivers/md/bcache/writeback.c
@@ -9,6 +9,9 @@
 #include "bcache.h"
 #include "btree.h"
 #include "debug.h"
+#include "writeback.h"
+
+#include <trace/events/bcache.h>
 
 static struct workqueue_struct *dirty_wq;
 
@@ -36,7 +39,7 @@ static void __update_writeback_rate(struct cached_dev *dc)
 
 	int change = 0;
 	int64_t error;
-	int64_t dirty = atomic_long_read(&dc->disk.sectors_dirty);
+	int64_t dirty = bcache_dev_sectors_dirty(&dc->disk);
 	int64_t derivative = dirty - dc->disk.sectors_dirty_last;
 
 	dc->disk.sectors_dirty_last = dirty;
@@ -105,6 +108,31 @@ static bool dirty_pred(struct keybuf *buf, struct bkey *k)
 	return KEY_DIRTY(k);
 }
 
+static bool dirty_full_stripe_pred(struct keybuf *buf, struct bkey *k)
+{
+	uint64_t stripe;
+	unsigned nr_sectors = KEY_SIZE(k);
+	struct cached_dev *dc = container_of(buf, struct cached_dev,
+					     writeback_keys);
+	unsigned stripe_size = 1 << dc->disk.stripe_size_bits;
+
+	if (!KEY_DIRTY(k))
+		return false;
+
+	stripe = KEY_START(k) >> dc->disk.stripe_size_bits;
+	while (1) {
+		if (atomic_read(dc->disk.stripe_sectors_dirty + stripe) !=
+		    stripe_size)
+			return false;
+
+		if (nr_sectors <= stripe_size)
+			return true;
+
+		nr_sectors -= stripe_size;
+		stripe++;
+	}
+}
+
 static void dirty_init(struct keybuf_key *w)
 {
 	struct dirty_io *io = w->private;
@@ -149,7 +177,22 @@ static void refill_dirty(struct closure *cl)
 		searched_from_start = true;
 	}
 
-	bch_refill_keybuf(dc->disk.c, buf, &end);
+	if (dc->partial_stripes_expensive) {
+		uint64_t i;
+
+		for (i = 0; i < dc->disk.nr_stripes; i++)
+			if (atomic_read(dc->disk.stripe_sectors_dirty + i) ==
+			    1 << dc->disk.stripe_size_bits)
+				goto full_stripes;
+
+		goto normal_refill;
+full_stripes:
+		bch_refill_keybuf(dc->disk.c, buf, &end,
+				  dirty_full_stripe_pred);
+	} else {
+normal_refill:
+		bch_refill_keybuf(dc->disk.c, buf, &end, dirty_pred);
+	}
 
 	if (bkey_cmp(&buf->last_scanned, &end) >= 0 && searched_from_start) {
 		/* Searched the entire btree  - delay awhile */
@@ -181,10 +224,8 @@ void bch_writeback_queue(struct cached_dev *dc)
 	}
 }
 
-void bch_writeback_add(struct cached_dev *dc, unsigned sectors)
+void bch_writeback_add(struct cached_dev *dc)
 {
-	atomic_long_add(sectors, &dc->disk.sectors_dirty);
-
 	if (!atomic_read(&dc->has_dirty) &&
 	    !atomic_xchg(&dc->has_dirty, 1)) {
 		atomic_inc(&dc->count);
@@ -203,6 +244,34 @@ void bch_writeback_add(struct cached_dev *dc, unsigned sectors)
 	}
 }
 
+void bcache_dev_sectors_dirty_add(struct cache_set *c, unsigned inode,
+				  uint64_t offset, int nr_sectors)
+{
+	struct bcache_device *d = c->devices[inode];
+	unsigned stripe_size, stripe_offset;
+	uint64_t stripe;
+
+	if (!d)
+		return;
+
+	stripe_size = 1 << d->stripe_size_bits;
+	stripe = offset >> d->stripe_size_bits;
+	stripe_offset = offset & (stripe_size - 1);
+
+	while (nr_sectors) {
+		int s = min_t(unsigned, abs(nr_sectors),
+			      stripe_size - stripe_offset);
+
+		if (nr_sectors < 0)
+			s = -s;
+
+		atomic_add(s, d->stripe_sectors_dirty + stripe);
+		nr_sectors -= s;
+		stripe_offset = 0;
+		stripe++;
+	}
+}
+
 /* Background writeback - IO loop */
 
 static void dirty_io_destructor(struct closure *cl)
@@ -216,9 +285,10 @@ static void write_dirty_finish(struct closure *cl)
 	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
 	struct keybuf_key *w = io->bio.bi_private;
 	struct cached_dev *dc = io->dc;
-	struct bio_vec *bv = bio_iovec_idx(&io->bio, io->bio.bi_vcnt);
+	struct bio_vec *bv;
+	int i;
 
-	while (bv-- != io->bio.bi_io_vec)
+	bio_for_each_segment_all(bv, &io->bio, i)
 		__free_page(bv->bv_page);
 
 	/* This is kind of a dumb way of signalling errors. */
@@ -236,10 +306,12 @@ static void write_dirty_finish(struct closure *cl)
 		for (i = 0; i < KEY_PTRS(&w->key); i++)
 			atomic_inc(&PTR_BUCKET(dc->disk.c, &w->key, i)->pin);
 
-		pr_debug("clearing %s", pkey(&w->key));
 		bch_btree_insert(&op, dc->disk.c);
 		closure_sync(&op.cl);
 
+		if (op.insert_collision)
+			trace_bcache_writeback_collision(&w->key);
+
 		atomic_long_inc(op.insert_collision
 				? &dc->disk.c->writeback_keys_failed
 				: &dc->disk.c->writeback_keys_done);
@@ -275,7 +347,6 @@ static void write_dirty(struct closure *cl)
 	io->bio.bi_bdev		= io->dc->bdev;
 	io->bio.bi_end_io	= dirty_endio;
 
-	trace_bcache_write_dirty(&io->bio);
 	closure_bio_submit(&io->bio, cl, &io->dc->disk);
 
 	continue_at(cl, write_dirty_finish, dirty_wq);
@@ -296,7 +367,6 @@ static void read_dirty_submit(struct closure *cl)
 {
 	struct dirty_io *io = container_of(cl, struct dirty_io, cl);
 
-	trace_bcache_read_dirty(&io->bio);
 	closure_bio_submit(&io->bio, cl, &io->dc->disk);
 
 	continue_at(cl, write_dirty, dirty_wq);
@@ -349,10 +419,10 @@ static void read_dirty(struct closure *cl)
 		io->bio.bi_rw		= READ;
 		io->bio.bi_end_io	= read_dirty_endio;
 
-		if (bch_bio_alloc_pages(&io->bio, GFP_KERNEL))
+		if (bio_alloc_pages(&io->bio, GFP_KERNEL))
 			goto err_free;
 
-		pr_debug("%s", pkey(&w->key));
+		trace_bcache_writeback(&w->key);
 
 		closure_call(&io->cl, read_dirty_submit, NULL, &dc->disk.cl);
 
@@ -375,12 +445,49 @@ err:
 	refill_dirty(cl);
 }
 
+/* Init */
+
+static int bch_btree_sectors_dirty_init(struct btree *b, struct btree_op *op,
+					struct cached_dev *dc)
+{
+	struct bkey *k;
+	struct btree_iter iter;
+
+	bch_btree_iter_init(b, &iter, &KEY(dc->disk.id, 0, 0));
+	while ((k = bch_btree_iter_next_filter(&iter, b, bch_ptr_bad)))
+		if (!b->level) {
+			if (KEY_INODE(k) > dc->disk.id)
+				break;
+
+			if (KEY_DIRTY(k))
+				bcache_dev_sectors_dirty_add(b->c, dc->disk.id,
+							     KEY_START(k),
+							     KEY_SIZE(k));
+		} else {
+			btree(sectors_dirty_init, k, b, op, dc);
+			if (KEY_INODE(k) > dc->disk.id)
+				break;
+
+			cond_resched();
+		}
+
+	return 0;
+}
+
+void bch_sectors_dirty_init(struct cached_dev *dc)
+{
+	struct btree_op op;
+
+	bch_btree_op_init_stack(&op);
+	btree_root(sectors_dirty_init, dc->disk.c, &op, dc);
+}
+
 void bch_cached_dev_writeback_init(struct cached_dev *dc)
 {
 	closure_init_unlocked(&dc->writeback);
 	init_rwsem(&dc->writeback_lock);
 
-	bch_keybuf_init(&dc->writeback_keys, dirty_pred);
+	bch_keybuf_init(&dc->writeback_keys);
 
 	dc->writeback_metadata		= true;
 	dc->writeback_running		= true;
diff --git a/drivers/md/bcache/writeback.h b/drivers/md/bcache/writeback.h
new file mode 100644
index 000000000000..c91f61bb95b6
--- /dev/null
+++ b/drivers/md/bcache/writeback.h
@@ -0,0 +1,64 @@
+#ifndef _BCACHE_WRITEBACK_H
+#define _BCACHE_WRITEBACK_H
+
+#define CUTOFF_WRITEBACK	40
+#define CUTOFF_WRITEBACK_SYNC	70
+
+static inline uint64_t bcache_dev_sectors_dirty(struct bcache_device *d)
+{
+	uint64_t i, ret = 0;
+
+	for (i = 0; i < d->nr_stripes; i++)
+		ret += atomic_read(d->stripe_sectors_dirty + i);
+
+	return ret;
+}
+
+static inline bool bcache_dev_stripe_dirty(struct bcache_device *d,
+					   uint64_t offset,
+					   unsigned nr_sectors)
+{
+	uint64_t stripe = offset >> d->stripe_size_bits;
+
+	while (1) {
+		if (atomic_read(d->stripe_sectors_dirty + stripe))
+			return true;
+
+		if (nr_sectors <= 1 << d->stripe_size_bits)
+			return false;
+
+		nr_sectors -= 1 << d->stripe_size_bits;
+		stripe++;
+	}
+}
+
+static inline bool should_writeback(struct cached_dev *dc, struct bio *bio,
+				    unsigned cache_mode, bool would_skip)
+{
+	unsigned in_use = dc->disk.c->gc_stats.in_use;
+
+	if (cache_mode != CACHE_MODE_WRITEBACK ||
+	    atomic_read(&dc->disk.detaching) ||
+	    in_use > CUTOFF_WRITEBACK_SYNC)
+		return false;
+
+	if (dc->partial_stripes_expensive &&
+	    bcache_dev_stripe_dirty(&dc->disk, bio->bi_sector,
+				    bio_sectors(bio)))
+		return true;
+
+	if (would_skip)
+		return false;
+
+	return bio->bi_rw & REQ_SYNC ||
+		in_use <= CUTOFF_WRITEBACK;
+}
+
+void bcache_dev_sectors_dirty_add(struct cache_set *, unsigned, uint64_t, int);
+void bch_writeback_queue(struct cached_dev *);
+void bch_writeback_add(struct cached_dev *);
+
+void bch_sectors_dirty_init(struct cached_dev *dc);
+void bch_cached_dev_writeback_init(struct cached_dev *);
+
+#endif