bcache: Prune struct btree_op

Eventual goal is for struct btree_op to contain only what is necessary
for traversing the btree.
Signed-off-by: Kent Overstreet <kmo@daterainc.com>

drivers/md/bcache/bset.c

@@ -1197,7 +1197,6 @@ int bch_bset_print_stats(struct cache_set *c, char *buf)
 
 	memset(&t, 0, sizeof(struct bset_stats));
 	bch_btree_op_init_stack(&t.op);
-	t.op.c = c;
 
 	ret = bch_btree_map_nodes(&t.op, c, &ZERO_KEY, btree_bset_stats);
 	if (ret < 0)

drivers/md/bcache/btree.c

@@ -503,7 +503,7 @@ static void btree_node_write_work(struct work_struct *w)
 	rw_unlock(true, b);
 }
 
-static void bch_btree_leaf_dirty(struct btree *b, struct btree_op *op)
+static void bch_btree_leaf_dirty(struct btree *b, atomic_t *journal_ref)
 {
 	struct bset *i = b->sets[b->nsets].data;
 	struct btree_write *w = btree_current_write(b);
@@ -516,15 +516,15 @@ static void bch_btree_leaf_dirty(struct btree *b, struct btree_op *op)
 
 	set_btree_node_dirty(b);
 
-	if (op->journal) {
+	if (journal_ref) {
 		if (w->journal &&
-		    journal_pin_cmp(b->c, w, op)) {
+		    journal_pin_cmp(b->c, w->journal, journal_ref)) {
 			atomic_dec_bug(w->journal);
 			w->journal = NULL;
 		}
 
 		if (!w->journal) {
-			w->journal = op->journal;
+			w->journal = journal_ref;
 			atomic_inc(w->journal);
 		}
 	}
@@ -1663,13 +1663,16 @@ static int bch_btree_check_recurse(struct btree *b, struct btree_op *op,
 	return 0;
 }
 
-int bch_btree_check(struct cache_set *c, struct btree_op *op)
+int bch_btree_check(struct cache_set *c)
 {
 	int ret = -ENOMEM;
 	unsigned i;
 	unsigned long *seen[MAX_CACHES_PER_SET];
+	struct btree_op op;
 
 	memset(seen, 0, sizeof(seen));
+	bch_btree_op_init_stack(&op);
+	op.lock = SHRT_MAX;
 
 	for (i = 0; c->cache[i]; i++) {
 		size_t n = DIV_ROUND_UP(c->cache[i]->sb.nbuckets, 8);
@@ -1681,7 +1684,7 @@ int bch_btree_check(struct cache_set *c, struct btree_op *op)
 		memset(seen[i], 0xFF, n);
 	}
 
-	ret = btree_root(check_recurse, c, op, seen);
+	ret = btree_root(check_recurse, c, &op, seen);
 err:
 	for (i = 0; i < MAX_CACHES_PER_SET; i++)
 		kfree(seen[i]);
@@ -2091,7 +2094,8 @@ static int btree_split(struct btree *b, struct btree_op *op,
 }
 
 static int bch_btree_insert_node(struct btree *b, struct btree_op *op,
-				 struct keylist *insert_keys)
+				 struct keylist *insert_keys,
+				 atomic_t *journal_ref)
 {
 	int ret = 0;
 	struct keylist split_keys;
@@ -2123,7 +2127,7 @@ static int bch_btree_insert_node(struct btree *b, struct btree_op *op,
 
 			if (bch_btree_insert_keys(b, op, insert_keys)) {
 				if (!b->level)
-					bch_btree_leaf_dirty(b, op);
+					bch_btree_leaf_dirty(b, journal_ref);
 				else
 					bch_btree_node_write(b, &op->cl);
 			}
@@ -2162,7 +2166,7 @@ int bch_btree_insert_check_key(struct btree *b, struct btree_op *op,
 
 	BUG_ON(op->type != BTREE_INSERT);
 
-	ret = bch_btree_insert_node(b, op, &insert);
+	ret = bch_btree_insert_node(b, op, &insert, NULL);
 
 	BUG_ON(!ret && !bch_keylist_empty(&insert));
 out:
@@ -2172,7 +2176,7 @@ int bch_btree_insert_check_key(struct btree *b, struct btree_op *op,
 }
 
 static int bch_btree_insert_recurse(struct btree *b, struct btree_op *op,
-				    struct keylist *keys)
+				    struct keylist *keys, atomic_t *journal_ref)
 {
 	if (bch_keylist_empty(keys))
 		return 0;
@@ -2189,14 +2193,14 @@ static int bch_btree_insert_recurse(struct btree *b, struct btree_op *op,
 			return -EIO;
 		}
 
-		return btree(insert_recurse, k, b, op, keys);
+		return btree(insert_recurse, k, b, op, keys, journal_ref);
 	} else {
-		return bch_btree_insert_node(b, op, keys);
+		return bch_btree_insert_node(b, op, keys, journal_ref);
 	}
 }
 
 int bch_btree_insert(struct btree_op *op, struct cache_set *c,
-		     struct keylist *keys)
+		     struct keylist *keys, atomic_t *journal_ref)
 {
 	int ret = 0;
 
@@ -2210,7 +2214,7 @@ int bch_btree_insert(struct btree_op *op, struct cache_set *c,
 
 	while (!bch_keylist_empty(keys)) {
 		op->lock = 0;
-		ret = btree_root(insert_recurse, c, op, keys);
+		ret = btree_root(insert_recurse, c, op, keys, journal_ref);
 
 		if (ret == -EAGAIN) {
 			ret = 0;

drivers/md/bcache/btree.h

@@ -238,17 +238,6 @@ void __bkey_put(struct cache_set *c, struct bkey *k);
 
 struct btree_op {
 	struct closure		cl;
-	struct cache_set	*c;
-
-	/* Journal entry we have a refcount on */
-	atomic_t		*journal;
-
-	/* Bio to be inserted into the cache */
-	struct bio		*cache_bio;
-
-	unsigned		inode;
-
-	uint16_t		write_prio;
 
 	/* Btree level at which we start taking write locks */
 	short			lock;
@@ -259,11 +248,6 @@ struct btree_op {
 		BTREE_REPLACE
 	} type:8;
 
-	unsigned		csum:1;
-	unsigned		bypass:1;
-	unsigned		flush_journal:1;
-
-	unsigned		insert_data_done:1;
 	unsigned		insert_collision:1;
 
 	BKEY_PADDED(replace);
@@ -303,12 +287,13 @@ struct btree *bch_btree_node_get(struct cache_set *, struct bkey *, int, bool);
 
 int bch_btree_insert_check_key(struct btree *, struct btree_op *,
 			       struct bkey *);
-int bch_btree_insert(struct btree_op *, struct cache_set *, struct keylist *);
+int bch_btree_insert(struct btree_op *, struct cache_set *,
+		     struct keylist *, atomic_t *);
 
 int bch_gc_thread_start(struct cache_set *);
 size_t bch_btree_gc_finish(struct cache_set *);
 void bch_moving_gc(struct cache_set *);
-int bch_btree_check(struct cache_set *, struct btree_op *);
+int bch_btree_check(struct cache_set *);
 uint8_t __bch_btree_mark_key(struct cache_set *, int, struct bkey *);
 
 static inline void wake_up_gc(struct cache_set *c)

drivers/md/bcache/journal.c

@@ -30,17 +30,20 @@ static void journal_read_endio(struct bio *bio, int error)
 }
 
 static int journal_read_bucket(struct cache *ca, struct list_head *list,
-			       struct btree_op *op, unsigned bucket_index)
+			       unsigned bucket_index)
 {
 	struct journal_device *ja = &ca->journal;
 	struct bio *bio = &ja->bio;
 
 	struct journal_replay *i;
 	struct jset *j, *data = ca->set->journal.w[0].data;
+	struct closure cl;
 	unsigned len, left, offset = 0;
 	int ret = 0;
 	sector_t bucket = bucket_to_sector(ca->set, ca->sb.d[bucket_index]);
 
+	closure_init_stack(&cl);
+
 	pr_debug("reading %llu", (uint64_t) bucket);
 
 	while (offset < ca->sb.bucket_size) {
@@ -54,11 +57,11 @@ reread:		left = ca->sb.bucket_size - offset;
 		bio->bi_size	= len << 9;
 
 		bio->bi_end_io	= journal_read_endio;
-		bio->bi_private = &op->cl;
+		bio->bi_private = &cl;
 		bch_bio_map(bio, data);
 
-		closure_bio_submit(bio, &op->cl, ca);
-		closure_sync(&op->cl);
+		closure_bio_submit(bio, &cl, ca);
+		closure_sync(&cl);
 
 		/* This function could be simpler now since we no longer write
 		 * journal entries that overlap bucket boundaries; this means
@@ -128,12 +131,11 @@ reread:		left = ca->sb.bucket_size - offset;
 	return ret;
 }
 
-int bch_journal_read(struct cache_set *c, struct list_head *list,
-			struct btree_op *op)
+int bch_journal_read(struct cache_set *c, struct list_head *list)
 {
 #define read_bucket(b)							\
 	({								\
-		int ret = journal_read_bucket(ca, list, op, b);		\
+		int ret = journal_read_bucket(ca, list, b);		\
 		__set_bit(b, bitmap);					\
 		if (ret < 0)						\
 			return ret;					\
@@ -291,8 +293,7 @@ void bch_journal_mark(struct cache_set *c, struct list_head *list)
 	}
 }
 
-int bch_journal_replay(struct cache_set *s, struct list_head *list,
-			  struct btree_op *op)
+int bch_journal_replay(struct cache_set *s, struct list_head *list)
 {
 	int ret = 0, keys = 0, entries = 0;
 	struct bkey *k;
@@ -301,8 +302,11 @@ int bch_journal_replay(struct cache_set *s, struct list_head *list,
 
 	uint64_t start = i->j.last_seq, end = i->j.seq, n = start;
 	struct keylist keylist;
+	struct btree_op op;
 
 	bch_keylist_init(&keylist);
+	bch_btree_op_init_stack(&op);
+	op.lock = SHRT_MAX;
 
 	list_for_each_entry(i, list, list) {
 		BUG_ON(i->pin && atomic_read(i->pin) != 1);
@@ -319,9 +323,7 @@ int bch_journal_replay(struct cache_set *s, struct list_head *list,
 			bkey_copy(keylist.top, k);
 			bch_keylist_push(&keylist);
 
-			op->journal = i->pin;
-
-			ret = bch_btree_insert(op, s, &keylist);
+			ret = bch_btree_insert(&op, s, &keylist, i->pin);
 			if (ret)
 				goto err;
 
@@ -346,7 +348,7 @@ int bch_journal_replay(struct cache_set *s, struct list_head *list,
 		kfree(i);
 	}
 err:
-	closure_sync(&op->cl);
+	closure_sync(&op.cl);
 	return ret;
 }
 
@@ -368,8 +370,8 @@ static void btree_flush_write(struct cache_set *c)
 			if (!best)
 				best = b;
 			else if (journal_pin_cmp(c,
-						 btree_current_write(best),
-						 btree_current_write(b))) {
+					btree_current_write(best)->journal,
+					btree_current_write(b)->journal)) {
 				best = b;
 			}
 		}

drivers/md/bcache/journal.h

@@ -189,8 +189,7 @@ struct journal_device {
 };
 
 #define journal_pin_cmp(c, l, r)				\
-	(fifo_idx(&(c)->journal.pin, (l)->journal) >		\
-	 fifo_idx(&(c)->journal.pin, (r)->journal))
+	(fifo_idx(&(c)->journal.pin, (l)) > fifo_idx(&(c)->journal.pin, (r)))
 
 #define JOURNAL_PIN	20000
 
@@ -206,10 +205,8 @@ atomic_t *bch_journal(struct cache_set *, struct keylist *, struct closure *);
 void bch_journal_next(struct journal *);
 void bch_journal_mark(struct cache_set *, struct list_head *);
 void bch_journal_meta(struct cache_set *, struct closure *);
-int bch_journal_read(struct cache_set *, struct list_head *,
-			struct btree_op *);
-int bch_journal_replay(struct cache_set *, struct list_head *,
-			  struct btree_op *);
+int bch_journal_read(struct cache_set *, struct list_head *);
+int bch_journal_replay(struct cache_set *, struct list_head *);
 
 void bch_journal_free(struct cache_set *);
 int bch_journal_alloc(struct cache_set *);

drivers/md/bcache/movinggc.c

@@ -55,9 +55,9 @@ static void write_moving_finish(struct closure *cl)
 	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);
+	bch_keybuf_del(&io->s.c->moving_gc_keys, io->w);
 
-	up(&io->s.op.c->moving_in_flight);
+	up(&io->s.c->moving_in_flight);
 
 	closure_return_with_destructor(cl, moving_io_destructor);
 }
@@ -70,7 +70,7 @@ static void read_moving_endio(struct bio *bio, int error)
 	if (error)
 		io->s.error = error;
 
-	bch_bbio_endio(io->s.op.c, bio, error, "reading data to move");
+	bch_bbio_endio(io->s.c, bio, error, "reading data to move");
 }
 
 static void moving_init(struct moving_io *io)
@@ -99,11 +99,11 @@ static void write_moving(struct closure *cl)
 
 		io->bio.bio.bi_sector	= KEY_START(&io->w->key);
 		s->op.lock		= -1;
-		s->op.write_prio	= 1;
-		s->op.cache_bio		= &io->bio.bio;
+		s->write_prio		= 1;
+		s->cache_bio		= &io->bio.bio;
 
 		s->writeback		= KEY_DIRTY(&io->w->key);
-		s->op.csum		= KEY_CSUM(&io->w->key);
+		s->csum			= KEY_CSUM(&io->w->key);
 
 		s->op.type = BTREE_REPLACE;
 		bkey_copy(&s->op.replace, &io->w->key);
@@ -121,7 +121,7 @@ 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;
 
-	bch_submit_bbio(bio, s->op.c, &io->w->key, 0);
+	bch_submit_bbio(bio, s->c, &io->w->key, 0);
 
 	continue_at(cl, write_moving, system_wq);
 }
@@ -151,8 +151,8 @@ static void read_moving(struct cache_set *c)
 
 		w->private	= io;
 		io->w		= w;
-		io->s.op.inode	= KEY_INODE(&w->key);
-		io->s.op.c	= c;
+		io->s.inode	= KEY_INODE(&w->key);
+		io->s.c		= c;
 
 		moving_init(io);
 		bio = &io->bio.bio;

drivers/md/bcache/request.c

@@ -217,6 +217,7 @@ static void bch_data_insert_keys(struct closure *cl)
 {
 	struct btree_op *op = container_of(cl, struct btree_op, cl);
 	struct search *s = container_of(op, struct search, op);
+	atomic_t *journal_ref = NULL;
 
 	/*
 	 * If we're looping, might already be waiting on
@@ -231,20 +232,19 @@ static void bch_data_insert_keys(struct closure *cl)
 #endif
 
 	if (s->write)
-		op->journal = bch_journal(op->c, &s->insert_keys,
-					  op->flush_journal
+		journal_ref = bch_journal(s->c, &s->insert_keys,
+					  s->flush_journal
 					  ? &s->cl : NULL);
 
-	if (bch_btree_insert(op, op->c, &s->insert_keys)) {
+	if (bch_btree_insert(op, s->c, &s->insert_keys, journal_ref)) {
 		s->error		= -ENOMEM;
-		op->insert_data_done	= true;
+		s->insert_data_done	= true;
 	}
 
-	if (op->journal)
-		atomic_dec_bug(op->journal);
-	op->journal = NULL;
+	if (journal_ref)
+		atomic_dec_bug(journal_ref);
 
-	if (!op->insert_data_done)
+	if (!s->insert_data_done)
 		continue_at(cl, bch_data_insert_start, bcache_wq);
 
 	bch_keylist_free(&s->insert_keys);
@@ -347,7 +347,7 @@ static struct open_bucket *pick_data_bucket(struct cache_set *c,
 static bool bch_alloc_sectors(struct bkey *k, unsigned sectors,
 			      struct search *s)
 {
-	struct cache_set *c = s->op.c;
+	struct cache_set *c = s->c;
 	struct open_bucket *b;
 	BKEY_PADDED(key) alloc;
 	unsigned i;
@@ -363,7 +363,7 @@ static bool bch_alloc_sectors(struct bkey *k, unsigned sectors,
 	spin_lock(&c->data_bucket_lock);
 
 	while (!(b = pick_data_bucket(c, k, s->task, &alloc.key))) {
-		unsigned watermark = s->op.write_prio
+		unsigned watermark = s->write_prio
 			? WATERMARK_MOVINGGC
 			: WATERMARK_NONE;
 
@@ -435,7 +435,7 @@ static void bch_data_invalidate(struct closure *cl)
 {
 	struct btree_op *op = container_of(cl, struct btree_op, cl);
 	struct search *s = container_of(op, struct search, op);
-	struct bio *bio = op->cache_bio;
+	struct bio *bio = s->cache_bio;
 
 	pr_debug("invalidating %i sectors from %llu",
 		 bio_sectors(bio), (uint64_t) bio->bi_sector);
@@ -443,17 +443,17 @@ static void bch_data_invalidate(struct closure *cl)
 	while (bio_sectors(bio)) {
 		unsigned len = min(bio_sectors(bio), 1U << 14);
 
-		if (bch_keylist_realloc(&s->insert_keys, 0, op->c))
+		if (bch_keylist_realloc(&s->insert_keys, 0, s->c))
 			goto out;
 
 		bio->bi_sector	+= len;
 		bio->bi_size	-= len << 9;
 
 		bch_keylist_add(&s->insert_keys,
-				&KEY(op->inode, bio->bi_sector, len));
+				&KEY(s->inode, bio->bi_sector, len));
 	}
 
-	op->insert_data_done = true;
+	s->insert_data_done = true;
 	bio_put(bio);
 out:
 	continue_at(cl, bch_data_insert_keys, bcache_wq);
@@ -506,21 +506,21 @@ static void bch_data_insert_endio(struct bio *bio, int error)
 			set_closure_fn(cl, NULL, NULL);
 	}
 
-	bch_bbio_endio(op->c, bio, error, "writing data to cache");
+	bch_bbio_endio(s->c, bio, error, "writing data to cache");
 }
 
 static void bch_data_insert_start(struct closure *cl)
 {
 	struct btree_op *op = container_of(cl, struct btree_op, cl);
 	struct search *s = container_of(op, struct search, op);
-	struct bio *bio = op->cache_bio, *n;
+	struct bio *bio = s->cache_bio, *n;
 
-	if (op->bypass)
+	if (s->bypass)
 		return bch_data_invalidate(cl);
 
-	if (atomic_sub_return(bio_sectors(bio), &op->c->sectors_to_gc) < 0) {
-		set_gc_sectors(op->c);
-		wake_up_gc(op->c);
+	if (atomic_sub_return(bio_sectors(bio), &s->c->sectors_to_gc) < 0) {
+		set_gc_sectors(s->c);
+		wake_up_gc(s->c);
 	}
 
 	/*
@@ -533,17 +533,17 @@ static void bch_data_insert_start(struct closure *cl)
 		unsigned i;
 		struct bkey *k;
 		struct bio_set *split = s->d
-			? s->d->bio_split : op->c->bio_split;
+			? s->d->bio_split : s->c->bio_split;
 
 		/* 1 for the device pointer and 1 for the chksum */
 		if (bch_keylist_realloc(&s->insert_keys,
-					1 + (op->csum ? 1 : 0),
-					op->c))
+					1 + (s->csum ? 1 : 0),
+					s->c))
 			continue_at(cl, bch_data_insert_keys, bcache_wq);
 
 		k = s->insert_keys.top;
 		bkey_init(k);
-		SET_KEY_INODE(k, op->inode);
+		SET_KEY_INODE(k, s->inode);
 		SET_KEY_OFFSET(k, bio->bi_sector);
 
 		if (!bch_alloc_sectors(k, bio_sectors(bio), s))
@@ -558,11 +558,11 @@ static void bch_data_insert_start(struct closure *cl)
 			SET_KEY_DIRTY(k, true);
 
 			for (i = 0; i < KEY_PTRS(k); i++)
-				SET_GC_MARK(PTR_BUCKET(op->c, k, i),
+				SET_GC_MARK(PTR_BUCKET(s->c, k, i),
 					    GC_MARK_DIRTY);
 		}
 
-		SET_KEY_CSUM(k, op->csum);
+		SET_KEY_CSUM(k, s->csum);
 		if (KEY_CSUM(k))
 			bio_csum(n, k);
 
@@ -570,10 +570,10 @@ static void bch_data_insert_start(struct closure *cl)
 		bch_keylist_push(&s->insert_keys);
 
 		n->bi_rw |= REQ_WRITE;
-		bch_submit_bbio(n, op->c, k, 0);
+		bch_submit_bbio(n, s->c, k, 0);
 	} while (n != bio);
 
-	op->insert_data_done = true;
+	s->insert_data_done = true;
 	continue_at(cl, bch_data_insert_keys, bcache_wq);
 err:
 	/* bch_alloc_sectors() blocks if s->writeback = true */
@@ -592,14 +592,14 @@ static void bch_data_insert_start(struct closure *cl)
 		 * we wait for buckets to be freed up, so just invalidate the
 		 * rest of the write.
 		 */
-		op->bypass = true;
+		s->bypass = true;
 		return bch_data_invalidate(cl);
 	} else {
 		/*
 		 * From a cache miss, we can just insert the keys for the data
 		 * we have written or bail out if we didn't do anything.
 		 */
-		op->insert_data_done = true;
+		s->insert_data_done = true;
 		bio_put(bio);
 
 		if (!bch_keylist_empty(&s->insert_keys))
@@ -622,11 +622,11 @@ static void bch_data_insert_start(struct closure *cl)
  * data is written it calls bch_journal, and after the keys have been added to
  * the next journal write they're inserted into the btree.
  *
- * It inserts the data in op->cache_bio; bi_sector is used for the key offset,
+ * It inserts the data in s->cache_bio; bi_sector is used for the key offset,
  * and op->inode is used for the key inode.
  *
- * If op->bypass is true, instead of inserting the data it invalidates the
- * region of the cache represented by op->cache_bio and op->inode.
+ * If s->bypass is true, instead of inserting the data it invalidates the
+ * region of the cache represented by s->cache_bio and op->inode.
  */
 void bch_data_insert(struct closure *cl)
 {
@@ -634,7 +634,7 @@ void bch_data_insert(struct closure *cl)
 	struct search *s = container_of(op, struct search, op);
 
 	bch_keylist_init(&s->insert_keys);
-	bio_get(op->cache_bio);
+	bio_get(s->cache_bio);
 	bch_data_insert_start(cl);
 }
 
@@ -655,12 +655,12 @@ static void bch_cache_read_endio(struct bio *bio, int error)
 
 	if (error)
 		s->error = error;
-	else if (ptr_stale(s->op.c, &b->key, 0)) {
-		atomic_long_inc(&s->op.c->cache_read_races);
+	else if (ptr_stale(s->c, &b->key, 0)) {
+		atomic_long_inc(&s->c->cache_read_races);
 		s->error = -EINTR;
 	}
 
-	bch_bbio_endio(s->op.c, bio, error, "reading from cache");
+	bch_bbio_endio(s->c, bio, error, "reading from cache");
 }
 
 /*
@@ -674,13 +674,13 @@ static int cache_lookup_fn(struct btree_op *op, struct btree *b, struct bkey *k)
 	struct bkey *bio_key;
 	unsigned ptr;
 
-	if (bkey_cmp(k, &KEY(op->inode, bio->bi_sector, 0)) <= 0)
+	if (bkey_cmp(k, &KEY(s->inode, bio->bi_sector, 0)) <= 0)
 		return MAP_CONTINUE;
 
-	if (KEY_INODE(k) != s->op.inode ||
+	if (KEY_INODE(k) != s->inode ||
 	    KEY_START(k) > bio->bi_sector) {
 		unsigned bio_sectors = bio_sectors(bio);
-		unsigned sectors = KEY_INODE(k) == s->op.inode
+		unsigned sectors = KEY_INODE(k) == s->inode
 			? min_t(uint64_t, INT_MAX,
 				KEY_START(k) - bio->bi_sector)
 			: INT_MAX;
@@ -708,8 +708,8 @@ static int cache_lookup_fn(struct btree_op *op, struct btree *b, struct bkey *k)
 	bio_key = &container_of(n, struct bbio, bio)->key;
 	bch_bkey_copy_single_ptr(bio_key, k, ptr);
 
-	bch_cut_front(&KEY(s->op.inode, n->bi_sector, 0), bio_key);
-	bch_cut_back(&KEY(s->op.inode, bio_end_sector(n), 0), bio_key);
+	bch_cut_front(&KEY(s->inode, n->bi_sector, 0), bio_key);
+	bch_cut_back(&KEY(s->inode, bio_end_sector(n), 0), bio_key);
 
 	n->bi_end_io	= bch_cache_read_endio;
 	n->bi_private	= &s->cl;
@@ -735,8 +735,8 @@ static void cache_lookup(struct closure *cl)
 	struct search *s = container_of(op, struct search, op);
 	struct bio *bio = &s->bio.bio;
 
-	int ret = bch_btree_map_keys(op, op->c,
-				     &KEY(op->inode, bio->bi_sector, 0),
+	int ret = bch_btree_map_keys(op, s->c,
+				     &KEY(s->inode, bio->bi_sector, 0),
 				     cache_lookup_fn, MAP_END_KEY);
 	if (ret == -EAGAIN)
 		continue_at(cl, cache_lookup, bcache_wq);
@@ -793,8 +793,8 @@ static void search_free(struct closure *cl)
 	struct search *s = container_of(cl, struct search, cl);
 	bio_complete(s);
 
-	if (s->op.cache_bio)
-		bio_put(s->op.cache_bio);
+	if (s->cache_bio)
+		bio_put(s->cache_bio);
 
 	if (s->unaligned_bvec)
 		mempool_free(s->bio.bio.bi_io_vec, s->d->unaligned_bvec);
@@ -813,14 +813,14 @@ static struct search *search_alloc(struct bio *bio, struct bcache_device *d)
 
 	__closure_init(&s->cl, NULL);
 
-	s->op.inode		= d->id;
-	s->op.c			= d->c;
+	s->inode		= d->id;
+	s->c			= d->c;
 	s->d			= d;
 	s->op.lock		= -1;
 	s->task			= current;
 	s->orig_bio		= bio;
 	s->write		= (bio->bi_rw & REQ_WRITE) != 0;
-	s->op.flush_journal	= (bio->bi_rw & (REQ_FLUSH|REQ_FUA)) != 0;
+	s->flush_journal	= (bio->bi_rw & (REQ_FLUSH|REQ_FUA)) != 0;
 	s->recoverable		= 1;
 	s->start_time		= jiffies;
 	do_bio_hook(s);
@@ -891,7 +891,7 @@ static struct hlist_head *iohash(struct cached_dev *dc, uint64_t k)
 
 static bool check_should_bypass(struct cached_dev *dc, struct search *s)
 {
-	struct cache_set *c = s->op.c;
+	struct cache_set *c = s->c;
 	struct bio *bio = &s->bio.bio;
 	unsigned mode = cache_mode(dc, bio);
 	unsigned sectors, congested = bch_get_congested(c);
@@ -985,11 +985,11 @@ static void cached_dev_cache_miss_done(struct closure *cl)
 	if (s->op.insert_collision)
 		bch_mark_cache_miss_collision(s);
 
-	if (s->op.cache_bio) {
+	if (s->cache_bio) {
 		int i;
 		struct bio_vec *bv;
 
-		__bio_for_each_segment(bv, s->op.cache_bio, i, 0)
+		bio_for_each_segment_all(bv, s->cache_bio, i)
 			__free_page(bv->bv_page);
 	}
 
@@ -1042,14 +1042,15 @@ static void cached_dev_read_done(struct closure *cl)
 	 * to the buffers the original bio pointed to:
 	 */
 
-	if (s->op.cache_bio) {
-		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);
+	if (s->cache_bio) {
+		bio_reset(s->cache_bio);
+		s->cache_bio->bi_sector =
+			s->cache_miss->bi_sector;
+		s->cache_bio->bi_bdev = s->cache_miss->bi_bdev;
+		s->cache_bio->bi_size = s->cache_bio_sectors << 9;
+		bch_bio_map(s->cache_bio, NULL);
 
-		bio_copy_data(s->cache_miss, s->op.cache_bio);
+		bio_copy_data(s->cache_miss, s->cache_bio);
 
 		bio_put(s->cache_miss);
 		s->cache_miss = NULL;
@@ -1060,8 +1061,8 @@ static void cached_dev_read_done(struct closure *cl)
 
 	bio_complete(s);
 
-	if (s->op.cache_bio &&
-	    !test_bit(CACHE_SET_STOPPING, &s->op.c->flags)) {
+	if (s->cache_bio &&
+	    !test_bit(CACHE_SET_STOPPING, &s->c->flags)) {
 		s->op.type = BTREE_REPLACE;
 		closure_call(&s->op.cl, bch_data_insert, NULL, cl);
 	}
@@ -1074,12 +1075,12 @@ static void cached_dev_read_done_bh(struct closure *cl)
 	struct search *s = container_of(cl, struct search, cl);
 	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
 
-	bch_mark_cache_accounting(s, !s->cache_miss, s->op.bypass);
-	trace_bcache_read(s->orig_bio, !s->cache_miss, s->op.bypass);
+	bch_mark_cache_accounting(s, !s->cache_miss, s->bypass);
+	trace_bcache_read(s->orig_bio, !s->cache_miss, s->bypass);
 
 	if (s->error)
 		continue_at_nobarrier(cl, cached_dev_read_error, bcache_wq);
-	else if (s->op.cache_bio || verify(dc, &s->bio.bio))
+	else if (s->cache_bio || verify(dc, &s->bio.bio))
 		continue_at_nobarrier(cl, cached_dev_read_done, bcache_wq);
 	else
 		continue_at_nobarrier(cl, cached_dev_bio_complete, NULL);
@@ -1093,7 +1094,7 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s,
 	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
 	struct bio *miss, *cache_bio;
 
-	if (s->cache_miss || s->op.bypass) {
+	if (s->cache_miss || s->bypass) {
 		miss = bch_bio_split(bio, sectors, GFP_NOIO, s->d->bio_split);
 		ret = miss == bio ? MAP_DONE : MAP_CONTINUE;
 		goto out_submit;
@@ -1101,13 +1102,13 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s,
 
 	if (!(bio->bi_rw & REQ_RAHEAD) &&
 	    !(bio->bi_rw & REQ_META) &&
-	    s->op.c->gc_stats.in_use < CUTOFF_CACHE_READA)
+	    s->c->gc_stats.in_use < CUTOFF_CACHE_READA)
 		reada = min_t(sector_t, dc->readahead >> 9,
 			      bdev_sectors(bio->bi_bdev) - bio_end_sector(bio));
 
 	s->cache_bio_sectors = min(sectors, bio_sectors(bio) + reada);
 
-	s->op.replace = KEY(s->op.inode, bio->bi_sector +
+	s->op.replace = KEY(s->inode, bio->bi_sector +
 			    s->cache_bio_sectors, s->cache_bio_sectors);
 
 	ret = bch_btree_insert_check_key(b, &s->op, &s->op.replace);
@@ -1137,7 +1138,7 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s,
 		goto out_put;
 
 	s->cache_miss	= miss;
-	s->op.cache_bio = cache_bio;
+	s->cache_bio = cache_bio;
 	bio_get(cache_bio);
 	closure_bio_submit(cache_bio, &s->cl, s->d);
 
@@ -1177,7 +1178,7 @@ static void cached_dev_write(struct cached_dev *dc, struct search *s)
 	struct bkey start = KEY(dc->disk.id, bio->bi_sector, 0);
 	struct bkey end = KEY(dc->disk.id, bio_end_sector(bio), 0);
 
-	bch_keybuf_check_overlapping(&s->op.c->moving_gc_keys, &start, &end);
+	bch_keybuf_check_overlapping(&s->c->moving_gc_keys, &start, &end);
 
 	down_read_non_owner(&dc->writeback_lock);
 	if (bch_keybuf_check_overlapping(&dc->writeback_keys, &start, &end)) {
@@ -1185,7 +1186,7 @@ static void cached_dev_write(struct cached_dev *dc, struct search *s)
 		 * We overlap with some dirty data undergoing background
 		 * writeback, force this write to writeback
 		 */
-		s->op.bypass	= false;
+		s->bypass	= false;
 		s->writeback	= true;
 	}
 
@@ -1197,27 +1198,27 @@ static void cached_dev_write(struct cached_dev *dc, struct search *s)
 	 * so we still want to call it.
 	 */
 	if (bio->bi_rw & REQ_DISCARD)
-		s->op.bypass = true;
+		s->bypass = true;
 
 	if (should_writeback(dc, s->orig_bio,
 			     cache_mode(dc, bio),
-			     s->op.bypass)) {
-		s->op.bypass = false;
+			     s->bypass)) {
+		s->bypass = false;
 		s->writeback = true;
 	}
 
-	trace_bcache_write(s->orig_bio, s->writeback, s->op.bypass);
+	trace_bcache_write(s->orig_bio, s->writeback, s->bypass);
 
-	if (s->op.bypass) {
-		s->op.cache_bio = s->orig_bio;
-		bio_get(s->op.cache_bio);
+	if (s->bypass) {
+		s->cache_bio = s->orig_bio;
+		bio_get(s->cache_bio);
 
 		if (!(bio->bi_rw & REQ_DISCARD) ||
 		    blk_queue_discard(bdev_get_queue(dc->bdev)))
 			closure_bio_submit(bio, cl, s->d);
 	} else if (s->writeback) {
 		bch_writeback_add(dc);
-		s->op.cache_bio = bio;
+		s->cache_bio = bio;
 
 		if (bio->bi_rw & REQ_FLUSH) {
 			/* Also need to send a flush to the backing device */
@@ -1232,8 +1233,8 @@ static void cached_dev_write(struct cached_dev *dc, struct search *s)
 			closure_bio_submit(flush, cl, s->d);
 		}
 	} else {
-		s->op.cache_bio = bio_clone_bioset(bio, GFP_NOIO,
-						   dc->disk.bio_split);
+		s->cache_bio = bio_clone_bioset(bio, GFP_NOIO,
+						dc->disk.bio_split);
 
 		closure_bio_submit(bio, cl, s->d);
 	}
@@ -1247,8 +1248,8 @@ static void cached_dev_nodata(struct closure *cl)
 	struct search *s = container_of(cl, struct search, cl);
 	struct bio *bio = &s->bio.bio;
 
-	if (s->op.flush_journal)
-		bch_journal_meta(s->op.c, cl);
+	if (s->flush_journal)
+		bch_journal_meta(s->c, cl);
 
 	/* If it's a flush, we send the flush to the backing device too */
 	closure_bio_submit(bio, cl, s->d);
@@ -1286,7 +1287,7 @@ static void cached_dev_make_request(struct request_queue *q, struct bio *bio)
 					      cached_dev_nodata,
 					      bcache_wq);
 		} else {
-			s->op.bypass = check_should_bypass(dc, s);
+			s->bypass = check_should_bypass(dc, s);
 
 			if (rw)
 				cached_dev_write(dc, s);
@@ -1376,8 +1377,8 @@ static void flash_dev_nodata(struct closure *cl)
 {
 	struct search *s = container_of(cl, struct search, cl);
 
-	if (s->op.flush_journal)
-		bch_journal_meta(s->op.c, cl);
+	if (s->flush_journal)
+		bch_journal_meta(s->c, cl);
 
 	continue_at(cl, search_free, NULL);
 }
@@ -1409,13 +1410,13 @@ static void flash_dev_make_request(struct request_queue *q, struct bio *bio)
 				      flash_dev_nodata,
 				      bcache_wq);
 	} else if (rw) {
-		bch_keybuf_check_overlapping(&s->op.c->moving_gc_keys,
+		bch_keybuf_check_overlapping(&s->c->moving_gc_keys,
 					&KEY(d->id, bio->bi_sector, 0),
 					&KEY(d->id, bio_end_sector(bio), 0));
 
-		s->op.bypass	= (bio->bi_rw & REQ_DISCARD) != 0;
+		s->bypass	= (bio->bi_rw & REQ_DISCARD) != 0;
 		s->writeback	= true;
-		s->op.cache_bio	= bio;
+		s->cache_bio	= bio;
 
 		closure_call(&s->op.cl, bch_data_insert, NULL, cl);
 	} else {

drivers/md/bcache/request.h

@@ -8,19 +8,33 @@ struct search {
 	struct closure		cl;
 
 	struct bcache_device	*d;
+	struct cache_set	*c;
 	struct task_struct	*task;
 
 	struct bbio		bio;
 	struct bio		*orig_bio;
 	struct bio		*cache_miss;
+
+	/* Bio to be inserted into the cache */
+	struct bio		*cache_bio;
 	unsigned		cache_bio_sectors;
 
+	unsigned		inode;
+
 	unsigned		recoverable:1;
 	unsigned		unaligned_bvec:1;
 
 	unsigned		write:1;
 	unsigned		writeback:1;
 
+	unsigned		csum:1;
+	unsigned		bypass:1;
+	unsigned		flush_journal:1;
+
+	unsigned		insert_data_done:1;
+
+	uint16_t		write_prio;
+
 	/* IO error returned to s->bio */
 	short			error;
 	unsigned long		start_time;

drivers/md/bcache/stats.c

@@ -200,7 +200,7 @@ void bch_mark_cache_accounting(struct search *s, bool hit, bool bypass)
 {
 	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
 	mark_cache_stats(&dc->accounting.collector, hit, bypass);
-	mark_cache_stats(&s->op.c->accounting.collector, hit, bypass);
+	mark_cache_stats(&s->c->accounting.collector, hit, bypass);
 #ifdef CONFIG_CGROUP_BCACHE
 	mark_cache_stats(&(bch_bio_to_cgroup(s->orig_bio)->stats), hit, bypass);
 #endif
@@ -210,21 +210,21 @@ void bch_mark_cache_readahead(struct search *s)
 {
 	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
 	atomic_inc(&dc->accounting.collector.cache_readaheads);
-	atomic_inc(&s->op.c->accounting.collector.cache_readaheads);
+	atomic_inc(&s->c->accounting.collector.cache_readaheads);
 }
 
 void bch_mark_cache_miss_collision(struct search *s)
 {
 	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
 	atomic_inc(&dc->accounting.collector.cache_miss_collisions);
-	atomic_inc(&s->op.c->accounting.collector.cache_miss_collisions);
+	atomic_inc(&s->c->accounting.collector.cache_miss_collisions);
 }
 
 void bch_mark_sectors_bypassed(struct search *s, int sectors)
 {
 	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
 	atomic_add(sectors, &dc->accounting.collector.sectors_bypassed);
-	atomic_add(sectors, &s->op.c->accounting.collector.sectors_bypassed);
+	atomic_add(sectors, &s->c->accounting.collector.sectors_bypassed);
 }
 
 void bch_cache_accounting_init(struct cache_accounting *acc,

drivers/md/bcache/super.c

@@ -1493,11 +1493,10 @@ static void run_cache_set(struct cache_set *c)
 	const char *err = "cannot allocate memory";
 	struct cached_dev *dc, *t;
 	struct cache *ca;
+	struct closure cl;
 	unsigned i;
 
-	struct btree_op op;
-	bch_btree_op_init_stack(&op);
-	op.lock = SHRT_MAX;
+	closure_init_stack(&cl);
 
 	for_each_cache(ca, c, i)
 		c->nbuckets += ca->sb.nbuckets;
@@ -1508,7 +1507,7 @@ static void run_cache_set(struct cache_set *c)
 		struct jset *j;
 
 		err = "cannot allocate memory for journal";
-		if (bch_journal_read(c, &journal, &op))
+		if (bch_journal_read(c, &journal))
 			goto err;
 
 		pr_debug("btree_journal_read() done");
@@ -1543,12 +1542,12 @@ static void run_cache_set(struct cache_set *c)
 		list_del_init(&c->root->list);
 		rw_unlock(true, c->root);
 
-		err = uuid_read(c, j, &op.cl);
+		err = uuid_read(c, j, &cl);
 		if (err)
 			goto err;
 
 		err = "error in recovery";
-		if (bch_btree_check(c, &op))
+		if (bch_btree_check(c))
 			goto err;
 
 		bch_journal_mark(c, &journal);
@@ -1580,7 +1579,7 @@ static void run_cache_set(struct cache_set *c)
 		if (j->version < BCACHE_JSET_VERSION_UUID)
 			__uuid_write(c);
 
-		bch_journal_replay(c, &journal, &op);
+		bch_journal_replay(c, &journal);
 	} else {
 		pr_notice("invalidating existing data");
 
@@ -1616,7 +1615,7 @@ static void run_cache_set(struct cache_set *c)
 			goto err;
 
 		bkey_copy_key(&c->root->key, &MAX_KEY);
-		bch_btree_node_write(c->root, &op.cl);
+		bch_btree_node_write(c->root, &cl);
 
 		bch_btree_set_root(c->root);
 		rw_unlock(true, c->root);
@@ -1629,14 +1628,14 @@ static void run_cache_set(struct cache_set *c)
 		SET_CACHE_SYNC(&c->sb, true);
 
 		bch_journal_next(&c->journal);
-		bch_journal_meta(c, &op.cl);
+		bch_journal_meta(c, &cl);
 	}
 
 	err = "error starting gc thread";
 	if (bch_gc_thread_start(c))
 		goto err;
 
-	closure_sync(&op.cl);
+	closure_sync(&cl);
 	c->sb.last_mount = get_seconds();
 	bcache_write_super(c);
 
@@ -1647,7 +1646,7 @@ static void run_cache_set(struct cache_set *c)
 
 	return;
 err:
-	closure_sync(&op.cl);
+	closure_sync(&cl);
 	/* XXX: test this, it's broken */
 	bch_cache_set_error(c, err);
 }

drivers/md/bcache/writeback.c

@@ -155,7 +155,7 @@ 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);
 
-		bch_btree_insert(&op, dc->disk.c, &keys);
+		bch_btree_insert(&op, dc->disk.c, &keys, NULL);
 		closure_sync(&op.cl);
 
 		if (op.insert_collision)
@@ -433,9 +433,16 @@ static int bch_writeback_thread(void *arg)
 
 /* Init */
 
-static int sectors_dirty_init_fn(struct btree_op *op, struct btree *b,
+struct sectors_dirty_init {
+	struct btree_op	op;
+	unsigned	inode;
+};
+
+static int sectors_dirty_init_fn(struct btree_op *_op, struct btree *b,
 				 struct bkey *k)
 {
+	struct sectors_dirty_init *op = container_of(_op,
+						struct sectors_dirty_init, op);
 	if (KEY_INODE(k) > op->inode)
 		return MAP_DONE;
 
@@ -448,12 +455,12 @@ static int sectors_dirty_init_fn(struct btree_op *op, struct btree *b,
 
 void bch_sectors_dirty_init(struct cached_dev *dc)
 {
-	struct btree_op op;
+	struct sectors_dirty_init op;
 
-	bch_btree_op_init_stack(&op);
+	bch_btree_op_init_stack(&op.op);
 	op.inode = dc->disk.id;
 
-	bch_btree_map_keys(&op, dc->disk.c, &KEY(op.inode, 0, 0),
+	bch_btree_map_keys(&op.op, dc->disk.c, &KEY(op.inode, 0, 0),
 			   sectors_dirty_init_fn, 0);
 }