diff --git a/block/blk-map.c b/block/blk-map.c
index 5d5dbe47c228..e663ac2d8e68 100644
--- a/block/blk-map.c
+++ b/block/blk-map.c
@@ -201,12 +201,13 @@ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
 	for (i = 0; i < iov_count; i++) {
 		unsigned long uaddr = (unsigned long)iov[i].iov_base;
 
+		if (!iov[i].iov_len)
+			return -EINVAL;
+
 		if (uaddr & queue_dma_alignment(q)) {
 			unaligned = 1;
 			break;
 		}
-		if (!iov[i].iov_len)
-			return -EINVAL;
 	}
 
 	if (unaligned || (q->dma_pad_mask & len) || map_data)
diff --git a/block/blk-merge.c b/block/blk-merge.c
index 77b7c26df6b5..74bc4a768f32 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -21,7 +21,7 @@ static unsigned int __blk_recalc_rq_segments(struct request_queue *q,
 		return 0;
 
 	fbio = bio;
-	cluster = test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags);
+	cluster = blk_queue_cluster(q);
 	seg_size = 0;
 	nr_phys_segs = 0;
 	for_each_bio(bio) {
@@ -87,7 +87,7 @@ EXPORT_SYMBOL(blk_recount_segments);
 static int blk_phys_contig_segment(struct request_queue *q, struct bio *bio,
 				   struct bio *nxt)
 {
-	if (!test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags))
+	if (!blk_queue_cluster(q))
 		return 0;
 
 	if (bio->bi_seg_back_size + nxt->bi_seg_front_size >
@@ -123,7 +123,7 @@ int blk_rq_map_sg(struct request_queue *q, struct request *rq,
 	int nsegs, cluster;
 
 	nsegs = 0;
-	cluster = test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags);
+	cluster = blk_queue_cluster(q);
 
 	/*
 	 * for each bio in rq
diff --git a/block/blk-settings.c b/block/blk-settings.c
index 701859fb9647..36c8c1f2af18 100644
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -126,7 +126,7 @@ void blk_set_default_limits(struct queue_limits *lim)
 	lim->alignment_offset = 0;
 	lim->io_opt = 0;
 	lim->misaligned = 0;
-	lim->no_cluster = 0;
+	lim->cluster = 1;
 }
 EXPORT_SYMBOL(blk_set_default_limits);
 
@@ -229,8 +229,8 @@ void blk_queue_bounce_limit(struct request_queue *q, u64 dma_mask)
 EXPORT_SYMBOL(blk_queue_bounce_limit);
 
 /**
- * blk_queue_max_hw_sectors - set max sectors for a request for this queue
- * @q:  the request queue for the device
+ * blk_limits_max_hw_sectors - set hard and soft limit of max sectors for request
+ * @limits: the queue limits
  * @max_hw_sectors:  max hardware sectors in the usual 512b unit
  *
  * Description:
@@ -244,7 +244,7 @@ EXPORT_SYMBOL(blk_queue_bounce_limit);
  *    per-device basis in /sys/block/<device>/queue/max_sectors_kb.
  *    The soft limit can not exceed max_hw_sectors.
  **/
-void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_sectors)
+void blk_limits_max_hw_sectors(struct queue_limits *limits, unsigned int max_hw_sectors)
 {
 	if ((max_hw_sectors << 9) < PAGE_CACHE_SIZE) {
 		max_hw_sectors = 1 << (PAGE_CACHE_SHIFT - 9);
@@ -252,9 +252,23 @@ void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_secto
 		       __func__, max_hw_sectors);
 	}
 
-	q->limits.max_hw_sectors = max_hw_sectors;
-	q->limits.max_sectors = min_t(unsigned int, max_hw_sectors,
-				      BLK_DEF_MAX_SECTORS);
+	limits->max_hw_sectors = max_hw_sectors;
+	limits->max_sectors = min_t(unsigned int, max_hw_sectors,
+				    BLK_DEF_MAX_SECTORS);
+}
+EXPORT_SYMBOL(blk_limits_max_hw_sectors);
+
+/**
+ * blk_queue_max_hw_sectors - set max sectors for a request for this queue
+ * @q:  the request queue for the device
+ * @max_hw_sectors:  max hardware sectors in the usual 512b unit
+ *
+ * Description:
+ *    See description for blk_limits_max_hw_sectors().
+ **/
+void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_sectors)
+{
+	blk_limits_max_hw_sectors(&q->limits, max_hw_sectors);
 }
 EXPORT_SYMBOL(blk_queue_max_hw_sectors);
 
@@ -464,15 +478,6 @@ EXPORT_SYMBOL(blk_queue_io_opt);
 void blk_queue_stack_limits(struct request_queue *t, struct request_queue *b)
 {
 	blk_stack_limits(&t->limits, &b->limits, 0);
-
-	if (!t->queue_lock)
-		WARN_ON_ONCE(1);
-	else if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags)) {
-		unsigned long flags;
-		spin_lock_irqsave(t->queue_lock, flags);
-		queue_flag_clear(QUEUE_FLAG_CLUSTER, t);
-		spin_unlock_irqrestore(t->queue_lock, flags);
-	}
 }
 EXPORT_SYMBOL(blk_queue_stack_limits);
 
@@ -545,7 +550,7 @@ int blk_stack_limits(struct queue_limits *t, struct queue_limits *b,
 	t->io_min = max(t->io_min, b->io_min);
 	t->io_opt = lcm(t->io_opt, b->io_opt);
 
-	t->no_cluster |= b->no_cluster;
+	t->cluster &= b->cluster;
 	t->discard_zeroes_data &= b->discard_zeroes_data;
 
 	/* Physical block size a multiple of the logical block size? */
@@ -641,7 +646,6 @@ void disk_stack_limits(struct gendisk *disk, struct block_device *bdev,
 		       sector_t offset)
 {
 	struct request_queue *t = disk->queue;
-	struct request_queue *b = bdev_get_queue(bdev);
 
 	if (bdev_stack_limits(&t->limits, bdev, offset >> 9) < 0) {
 		char top[BDEVNAME_SIZE], bottom[BDEVNAME_SIZE];
@@ -652,17 +656,6 @@ void disk_stack_limits(struct gendisk *disk, struct block_device *bdev,
 		printk(KERN_NOTICE "%s: Warning: Device %s is misaligned\n",
 		       top, bottom);
 	}
-
-	if (!t->queue_lock)
-		WARN_ON_ONCE(1);
-	else if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags)) {
-		unsigned long flags;
-
-		spin_lock_irqsave(t->queue_lock, flags);
-		if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags))
-			queue_flag_clear(QUEUE_FLAG_CLUSTER, t);
-		spin_unlock_irqrestore(t->queue_lock, flags);
-	}
 }
 EXPORT_SYMBOL(disk_stack_limits);
 
diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
index 013457f47fdc..41fb69150b4d 100644
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c
@@ -119,7 +119,7 @@ static ssize_t queue_max_integrity_segments_show(struct request_queue *q, char *
 
 static ssize_t queue_max_segment_size_show(struct request_queue *q, char *page)
 {
-	if (test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags))
+	if (blk_queue_cluster(q))
 		return queue_var_show(queue_max_segment_size(q), (page));
 
 	return queue_var_show(PAGE_CACHE_SIZE, (page));
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 004be80fd894..381b09bb562b 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -355,6 +355,12 @@ throtl_start_new_slice(struct throtl_data *td, struct throtl_grp *tg, bool rw)
 			tg->slice_end[rw], jiffies);
 }
 
+static inline void throtl_set_slice_end(struct throtl_data *td,
+		struct throtl_grp *tg, bool rw, unsigned long jiffy_end)
+{
+	tg->slice_end[rw] = roundup(jiffy_end, throtl_slice);
+}
+
 static inline void throtl_extend_slice(struct throtl_data *td,
 		struct throtl_grp *tg, bool rw, unsigned long jiffy_end)
 {
@@ -391,6 +397,16 @@ throtl_trim_slice(struct throtl_data *td, struct throtl_grp *tg, bool rw)
 	if (throtl_slice_used(td, tg, rw))
 		return;
 
+	/*
+	 * A bio has been dispatched. Also adjust slice_end. It might happen
+	 * that initially cgroup limit was very low resulting in high
+	 * slice_end, but later limit was bumped up and bio was dispached
+	 * sooner, then we need to reduce slice_end. A high bogus slice_end
+	 * is bad because it does not allow new slice to start.
+	 */
+
+	throtl_set_slice_end(td, tg, rw, jiffies + throtl_slice);
+
 	time_elapsed = jiffies - tg->slice_start[rw];
 
 	nr_slices = time_elapsed / throtl_slice;
@@ -709,26 +725,21 @@ static void throtl_process_limit_change(struct throtl_data *td)
 	struct throtl_grp *tg;
 	struct hlist_node *pos, *n;
 
-	/*
-	 * Make sure atomic_inc() effects from
-	 * throtl_update_blkio_group_read_bps(), group of functions are
-	 * visible.
-	 * Is this required or smp_mb__after_atomic_inc() was suffcient
-	 * after the atomic_inc().
-	 */
-	smp_rmb();
 	if (!atomic_read(&td->limits_changed))
 		return;
 
 	throtl_log(td, "limit changed =%d", atomic_read(&td->limits_changed));
 
-	hlist_for_each_entry_safe(tg, pos, n, &td->tg_list, tg_node) {
-		/*
-		 * Do I need an smp_rmb() here to make sure tg->limits_changed
-		 * update is visible. I am relying on smp_rmb() at the
-		 * beginning of function and not putting a new one here.
-		 */
+	/*
+	 * Make sure updates from throtl_update_blkio_group_read_bps() group
+	 * of functions to tg->limits_changed are visible. We do not
+	 * want update td->limits_changed to be visible but update to
+	 * tg->limits_changed not being visible yet on this cpu. Hence
+	 * the read barrier.
+	 */
+	smp_rmb();
 
+	hlist_for_each_entry_safe(tg, pos, n, &td->tg_list, tg_node) {
 		if (throtl_tg_on_rr(tg) && tg->limits_changed) {
 			throtl_log_tg(td, tg, "limit change rbps=%llu wbps=%llu"
 				" riops=%u wiops=%u", tg->bps[READ],
diff --git a/drivers/block/cciss.c b/drivers/block/cciss.c
index f291587d753e..233e06c29ff4 100644
--- a/drivers/block/cciss.c
+++ b/drivers/block/cciss.c
@@ -2834,6 +2834,8 @@ static int cciss_revalidate(struct gendisk *disk)
 	InquiryData_struct *inq_buff = NULL;
 
 	for (logvol = 0; logvol < CISS_MAX_LUN; logvol++) {
+		if (!h->drv[logvol])
+			continue
 		if (memcmp(h->drv[logvol]->LunID, drv->LunID,
 			sizeof(drv->LunID)) == 0) {
 			FOUND = 1;
diff --git a/drivers/block/drbd/drbd_receiver.c b/drivers/block/drbd/drbd_receiver.c
index 89d8a7cc4054..24487d4fb202 100644
--- a/drivers/block/drbd/drbd_receiver.c
+++ b/drivers/block/drbd/drbd_receiver.c
@@ -3627,17 +3627,19 @@ static void drbdd(struct drbd_conf *mdev)
 		}
 
 		shs = drbd_cmd_handler[cmd].pkt_size - sizeof(union p_header);
-		rv = drbd_recv(mdev, &header->h80.payload, shs);
-		if (unlikely(rv != shs)) {
-			dev_err(DEV, "short read while reading sub header: rv=%d\n", rv);
-			goto err_out;
-		}
-
 		if (packet_size - shs > 0 && !drbd_cmd_handler[cmd].expect_payload) {
 			dev_err(DEV, "No payload expected %s l:%d\n", cmdname(cmd), packet_size);
 			goto err_out;
 		}
 
+		if (shs) {
+			rv = drbd_recv(mdev, &header->h80.payload, shs);
+			if (unlikely(rv != shs)) {
+				dev_err(DEV, "short read while reading sub header: rv=%d\n", rv);
+				goto err_out;
+			}
+		}
+
 		rv = drbd_cmd_handler[cmd].function(mdev, cmd, packet_size - shs);
 
 		if (unlikely(!rv)) {
diff --git a/drivers/block/drbd/drbd_req.h b/drivers/block/drbd/drbd_req.h
index 181ea0364822..ab2bd09d54b4 100644
--- a/drivers/block/drbd/drbd_req.h
+++ b/drivers/block/drbd/drbd_req.h
@@ -339,7 +339,8 @@ static inline int _req_mod(struct drbd_request *req, enum drbd_req_event what)
 }
 
 /* completion of master bio is outside of spinlock.
- * If you need it irqsave, do it your self! */
+ * If you need it irqsave, do it your self!
+ * Which means: don't use from bio endio callback. */
 static inline int req_mod(struct drbd_request *req,
 		enum drbd_req_event what)
 {
diff --git a/drivers/block/drbd/drbd_worker.c b/drivers/block/drbd/drbd_worker.c
index 47d223c2409c..34f224b018b3 100644
--- a/drivers/block/drbd/drbd_worker.c
+++ b/drivers/block/drbd/drbd_worker.c
@@ -193,8 +193,10 @@ void drbd_endio_sec(struct bio *bio, int error)
  */
 void drbd_endio_pri(struct bio *bio, int error)
 {
+	unsigned long flags;
 	struct drbd_request *req = bio->bi_private;
 	struct drbd_conf *mdev = req->mdev;
+	struct bio_and_error m;
 	enum drbd_req_event what;
 	int uptodate = bio_flagged(bio, BIO_UPTODATE);
 
@@ -220,7 +222,13 @@ void drbd_endio_pri(struct bio *bio, int error)
 	bio_put(req->private_bio);
 	req->private_bio = ERR_PTR(error);
 
-	req_mod(req, what);
+	/* not req_mod(), we need irqsave here! */
+	spin_lock_irqsave(&mdev->req_lock, flags);
+	__req_mod(req, what, &m);
+	spin_unlock_irqrestore(&mdev->req_lock, flags);
+
+	if (m.bio)
+		complete_master_bio(mdev, &m);
 }
 
 int w_read_retry_remote(struct drbd_conf *mdev, struct drbd_work *w, int cancel)
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index 90267f8d64ee..4d705cea0f8c 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -517,9 +517,8 @@ int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev,
 	 */
 
 	if (q->merge_bvec_fn && !ti->type->merge)
-		limits->max_sectors =
-			min_not_zero(limits->max_sectors,
-				     (unsigned int) (PAGE_SIZE >> 9));
+		blk_limits_max_hw_sectors(limits,
+					  (unsigned int) (PAGE_SIZE >> 9));
 	return 0;
 }
 EXPORT_SYMBOL_GPL(dm_set_device_limits);
@@ -1131,11 +1130,6 @@ void dm_table_set_restrictions(struct dm_table *t, struct request_queue *q,
 	 */
 	q->limits = *limits;
 
-	if (limits->no_cluster)
-		queue_flag_clear_unlocked(QUEUE_FLAG_CLUSTER, q);
-	else
-		queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, q);
-
 	if (!dm_table_supports_discards(t))
 		queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, q);
 	else
diff --git a/drivers/md/md.c b/drivers/md/md.c
index e71c5fa527f5..175c424f201f 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -4295,9 +4295,6 @@ static int md_alloc(dev_t dev, char *name)
 		goto abort;
 	mddev->queue->queuedata = mddev;
 
-	/* Can be unlocked because the queue is new: no concurrency */
-	queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, mddev->queue);
-
 	blk_queue_make_request(mddev->queue, md_make_request);
 
 	disk = alloc_disk(1 << shift);
diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
index 5b6bbaea59fe..4a3842212c50 100644
--- a/drivers/scsi/scsi_lib.c
+++ b/drivers/scsi/scsi_lib.c
@@ -1637,9 +1637,8 @@ struct request_queue *__scsi_alloc_queue(struct Scsi_Host *shost,
 
 	blk_queue_max_segment_size(q, dma_get_max_seg_size(dev));
 
-	/* New queue, no concurrency on queue_flags */
 	if (!shost->use_clustering)
-		queue_flag_clear_unlocked(QUEUE_FLAG_CLUSTER, q);
+		q->limits.cluster = 0;
 
 	/*
 	 * set a reasonable default alignment on word boundaries: the
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index aae86fd10c4f..36ab42c9bb99 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -250,7 +250,7 @@ struct queue_limits {
 
 	unsigned char		misaligned;
 	unsigned char		discard_misaligned;
-	unsigned char		no_cluster;
+	unsigned char		cluster;
 	signed char		discard_zeroes_data;
 };
 
@@ -380,7 +380,6 @@ struct request_queue
 #endif
 };
 
-#define QUEUE_FLAG_CLUSTER	0	/* cluster several segments into 1 */
 #define QUEUE_FLAG_QUEUED	1	/* uses generic tag queueing */
 #define QUEUE_FLAG_STOPPED	2	/* queue is stopped */
 #define	QUEUE_FLAG_SYNCFULL	3	/* read queue has been filled */
@@ -403,7 +402,6 @@ struct request_queue
 #define QUEUE_FLAG_SECDISCARD  19	/* supports SECDISCARD */
 
 #define QUEUE_FLAG_DEFAULT	((1 << QUEUE_FLAG_IO_STAT) |		\
-				 (1 << QUEUE_FLAG_CLUSTER) |		\
 				 (1 << QUEUE_FLAG_STACKABLE)	|	\
 				 (1 << QUEUE_FLAG_SAME_COMP)	|	\
 				 (1 << QUEUE_FLAG_ADD_RANDOM))
@@ -510,6 +508,11 @@ static inline void queue_flag_clear(unsigned int flag, struct request_queue *q)
 
 #define rq_data_dir(rq)		((rq)->cmd_flags & 1)
 
+static inline unsigned int blk_queue_cluster(struct request_queue *q)
+{
+	return q->limits.cluster;
+}
+
 /*
  * We regard a request as sync, if either a read or a sync write
  */
@@ -805,6 +808,7 @@ extern struct request_queue *blk_init_allocated_queue(struct request_queue *,
 extern void blk_cleanup_queue(struct request_queue *);
 extern void blk_queue_make_request(struct request_queue *, make_request_fn *);
 extern void blk_queue_bounce_limit(struct request_queue *, u64);
+extern void blk_limits_max_hw_sectors(struct queue_limits *, unsigned int);
 extern void blk_queue_max_hw_sectors(struct request_queue *, unsigned int);
 extern void blk_queue_max_segments(struct request_queue *, unsigned short);
 extern void blk_queue_max_segment_size(struct request_queue *, unsigned int);