diff --git a/block/blk-core.c b/block/blk-core.c
index c00e0bdeab4a..853f92749202 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -693,11 +693,20 @@ blk_init_queue_node(request_fn_proc *rfn, spinlock_t *lock, int node_id)
 	if (!uninit_q)
 		return NULL;
 
+	uninit_q->flush_rq = kzalloc(sizeof(struct request), GFP_KERNEL);
+	if (!uninit_q->flush_rq)
+		goto out_cleanup_queue;
+
 	q = blk_init_allocated_queue(uninit_q, rfn, lock);
 	if (!q)
-		blk_cleanup_queue(uninit_q);
-
+		goto out_free_flush_rq;
 	return q;
+
+out_free_flush_rq:
+	kfree(uninit_q->flush_rq);
+out_cleanup_queue:
+	blk_cleanup_queue(uninit_q);
+	return NULL;
 }
 EXPORT_SYMBOL(blk_init_queue_node);
 
@@ -1127,7 +1136,7 @@ static struct request *blk_old_get_request(struct request_queue *q, int rw,
 struct request *blk_get_request(struct request_queue *q, int rw, gfp_t gfp_mask)
 {
 	if (q->mq_ops)
-		return blk_mq_alloc_request(q, rw, gfp_mask, false);
+		return blk_mq_alloc_request(q, rw, gfp_mask);
 	else
 		return blk_old_get_request(q, rw, gfp_mask);
 }
@@ -1278,6 +1287,11 @@ void __blk_put_request(struct request_queue *q, struct request *req)
 	if (unlikely(!q))
 		return;
 
+	if (q->mq_ops) {
+		blk_mq_free_request(req);
+		return;
+	}
+
 	blk_pm_put_request(req);
 
 	elv_completed_request(q, req);
diff --git a/block/blk-exec.c b/block/blk-exec.c
index bbfc072a79c2..c68613bb4c79 100644
--- a/block/blk-exec.c
+++ b/block/blk-exec.c
@@ -65,7 +65,7 @@ void blk_execute_rq_nowait(struct request_queue *q, struct gendisk *bd_disk,
 	 * be resued after dying flag is set
 	 */
 	if (q->mq_ops) {
-		blk_mq_insert_request(q, rq, true);
+		blk_mq_insert_request(q, rq, at_head, true);
 		return;
 	}
 
diff --git a/block/blk-flush.c b/block/blk-flush.c
index 9288aaf35c21..66e2b697f5db 100644
--- a/block/blk-flush.c
+++ b/block/blk-flush.c
@@ -130,20 +130,26 @@ static void blk_flush_restore_request(struct request *rq)
 	blk_clear_rq_complete(rq);
 }
 
-static void mq_flush_data_run(struct work_struct *work)
+static void mq_flush_run(struct work_struct *work)
 {
 	struct request *rq;
 
-	rq = container_of(work, struct request, mq_flush_data);
+	rq = container_of(work, struct request, mq_flush_work);
 
 	memset(&rq->csd, 0, sizeof(rq->csd));
 	blk_mq_run_request(rq, true, false);
 }
 
-static void blk_mq_flush_data_insert(struct request *rq)
+static bool blk_flush_queue_rq(struct request *rq)
 {
-	INIT_WORK(&rq->mq_flush_data, mq_flush_data_run);
-	kblockd_schedule_work(rq->q, &rq->mq_flush_data);
+	if (rq->q->mq_ops) {
+		INIT_WORK(&rq->mq_flush_work, mq_flush_run);
+		kblockd_schedule_work(rq->q, &rq->mq_flush_work);
+		return false;
+	} else {
+		list_add_tail(&rq->queuelist, &rq->q->queue_head);
+		return true;
+	}
 }
 
 /**
@@ -187,12 +193,7 @@ static bool blk_flush_complete_seq(struct request *rq, unsigned int seq,
 
 	case REQ_FSEQ_DATA:
 		list_move_tail(&rq->flush.list, &q->flush_data_in_flight);
-		if (q->mq_ops)
-			blk_mq_flush_data_insert(rq);
-		else {
-			list_add(&rq->queuelist, &q->queue_head);
-			queued = true;
-		}
+		queued = blk_flush_queue_rq(rq);
 		break;
 
 	case REQ_FSEQ_DONE:
@@ -216,9 +217,6 @@ static bool blk_flush_complete_seq(struct request *rq, unsigned int seq,
 	}
 
 	kicked = blk_kick_flush(q);
-	/* blk_mq_run_flush will run queue */
-	if (q->mq_ops)
-		return queued;
 	return kicked | queued;
 }
 
@@ -230,10 +228,9 @@ static void flush_end_io(struct request *flush_rq, int error)
 	struct request *rq, *n;
 	unsigned long flags = 0;
 
-	if (q->mq_ops) {
-		blk_mq_free_request(flush_rq);
+	if (q->mq_ops)
 		spin_lock_irqsave(&q->mq_flush_lock, flags);
-	}
+
 	running = &q->flush_queue[q->flush_running_idx];
 	BUG_ON(q->flush_pending_idx == q->flush_running_idx);
 
@@ -263,49 +260,14 @@ static void flush_end_io(struct request *flush_rq, int error)
 	 * kblockd.
 	 */
 	if (queued || q->flush_queue_delayed) {
-		if (!q->mq_ops)
-			blk_run_queue_async(q);
-		else
-		/*
-		 * This can be optimized to only run queues with requests
-		 * queued if necessary.
-		 */
-			blk_mq_run_queues(q, true);
+		WARN_ON(q->mq_ops);
+		blk_run_queue_async(q);
 	}
 	q->flush_queue_delayed = 0;
 	if (q->mq_ops)
 		spin_unlock_irqrestore(&q->mq_flush_lock, flags);
 }
 
-static void mq_flush_work(struct work_struct *work)
-{
-	struct request_queue *q;
-	struct request *rq;
-
-	q = container_of(work, struct request_queue, mq_flush_work);
-
-	/* We don't need set REQ_FLUSH_SEQ, it's for consistency */
-	rq = blk_mq_alloc_request(q, WRITE_FLUSH|REQ_FLUSH_SEQ,
-		__GFP_WAIT|GFP_ATOMIC, true);
-	rq->cmd_type = REQ_TYPE_FS;
-	rq->end_io = flush_end_io;
-
-	blk_mq_run_request(rq, true, false);
-}
-
-/*
- * We can't directly use q->flush_rq, because it doesn't have tag and is not in
- * hctx->rqs[]. so we must allocate a new request, since we can't sleep here,
- * so offload the work to workqueue.
- *
- * Note: we assume a flush request finished in any hardware queue will flush
- * the whole disk cache.
- */
-static void mq_run_flush(struct request_queue *q)
-{
-	kblockd_schedule_work(q, &q->mq_flush_work);
-}
-
 /**
  * blk_kick_flush - consider issuing flush request
  * @q: request_queue being kicked
@@ -340,19 +302,31 @@ static bool blk_kick_flush(struct request_queue *q)
 	 * different from running_idx, which means flush is in flight.
 	 */
 	q->flush_pending_idx ^= 1;
+
 	if (q->mq_ops) {
-		mq_run_flush(q);
-		return true;
+		struct blk_mq_ctx *ctx = first_rq->mq_ctx;
+		struct blk_mq_hw_ctx *hctx = q->mq_ops->map_queue(q, ctx->cpu);
+
+		blk_mq_rq_init(hctx, q->flush_rq);
+		q->flush_rq->mq_ctx = ctx;
+
+		/*
+		 * Reuse the tag value from the fist waiting request,
+		 * with blk-mq the tag is generated during request
+		 * allocation and drivers can rely on it being inside
+		 * the range they asked for.
+		 */
+		q->flush_rq->tag = first_rq->tag;
+	} else {
+		blk_rq_init(q, q->flush_rq);
 	}
 
-	blk_rq_init(q, &q->flush_rq);
-	q->flush_rq.cmd_type = REQ_TYPE_FS;
-	q->flush_rq.cmd_flags = WRITE_FLUSH | REQ_FLUSH_SEQ;
-	q->flush_rq.rq_disk = first_rq->rq_disk;
-	q->flush_rq.end_io = flush_end_io;
+	q->flush_rq->cmd_type = REQ_TYPE_FS;
+	q->flush_rq->cmd_flags = WRITE_FLUSH | REQ_FLUSH_SEQ;
+	q->flush_rq->rq_disk = first_rq->rq_disk;
+	q->flush_rq->end_io = flush_end_io;
 
-	list_add_tail(&q->flush_rq.queuelist, &q->queue_head);
-	return true;
+	return blk_flush_queue_rq(q->flush_rq);
 }
 
 static void flush_data_end_io(struct request *rq, int error)
@@ -558,5 +532,4 @@ EXPORT_SYMBOL(blkdev_issue_flush);
 void blk_mq_init_flush(struct request_queue *q)
 {
 	spin_lock_init(&q->mq_flush_lock);
-	INIT_WORK(&q->mq_flush_work, mq_flush_work);
 }
diff --git a/block/blk-lib.c b/block/blk-lib.c
index 2da76c999ef3..97a733cf3d5f 100644
--- a/block/blk-lib.c
+++ b/block/blk-lib.c
@@ -119,6 +119,14 @@ int blkdev_issue_discard(struct block_device *bdev, sector_t sector,
 
 		atomic_inc(&bb.done);
 		submit_bio(type, bio);
+
+		/*
+		 * We can loop for a long time in here, if someone does
+		 * full device discards (like mkfs). Be nice and allow
+		 * us to schedule out to avoid softlocking if preempt
+		 * is disabled.
+		 */
+		cond_resched();
 	}
 	blk_finish_plug(&plug);
 
diff --git a/block/blk-merge.c b/block/blk-merge.c
index 8f8adaa95466..6c583f9c5b65 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -21,6 +21,16 @@ static unsigned int __blk_recalc_rq_segments(struct request_queue *q,
 	if (!bio)
 		return 0;
 
+	/*
+	 * This should probably be returning 0, but blk_add_request_payload()
+	 * (Christoph!!!!)
+	 */
+	if (bio->bi_rw & REQ_DISCARD)
+		return 1;
+
+	if (bio->bi_rw & REQ_WRITE_SAME)
+		return 1;
+
 	fbio = bio;
 	cluster = blk_queue_cluster(q);
 	seg_size = 0;
@@ -161,6 +171,48 @@ new_segment:
 	*bvprv = *bvec;
 }
 
+static int __blk_bios_map_sg(struct request_queue *q, struct bio *bio,
+			     struct scatterlist *sglist,
+			     struct scatterlist **sg)
+{
+	struct bio_vec bvec, bvprv = { NULL };
+	struct bvec_iter iter;
+	int nsegs, cluster;
+
+	nsegs = 0;
+	cluster = blk_queue_cluster(q);
+
+	if (bio->bi_rw & REQ_DISCARD) {
+		/*
+		 * This is a hack - drivers should be neither modifying the
+		 * biovec, nor relying on bi_vcnt - but because of
+		 * blk_add_request_payload(), a discard bio may or may not have
+		 * a payload we need to set up here (thank you Christoph) and
+		 * bi_vcnt is really the only way of telling if we need to.
+		 */
+
+		if (bio->bi_vcnt)
+			goto single_segment;
+
+		return 0;
+	}
+
+	if (bio->bi_rw & REQ_WRITE_SAME) {
+single_segment:
+		*sg = sglist;
+		bvec = bio_iovec(bio);
+		sg_set_page(*sg, bvec.bv_page, bvec.bv_len, bvec.bv_offset);
+		return 1;
+	}
+
+	for_each_bio(bio)
+		bio_for_each_segment(bvec, bio, iter)
+			__blk_segment_map_sg(q, &bvec, sglist, &bvprv, sg,
+					     &nsegs, &cluster);
+
+	return nsegs;
+}
+
 /*
  * map a request to scatterlist, return number of sg entries setup. Caller
  * must make sure sg can hold rq->nr_phys_segments entries
@@ -168,23 +220,11 @@ new_segment:
 int blk_rq_map_sg(struct request_queue *q, struct request *rq,
 		  struct scatterlist *sglist)
 {
-	struct bio_vec bvec, bvprv = { NULL };
-	struct req_iterator iter;
-	struct scatterlist *sg;
-	int nsegs, cluster;
-
-	nsegs = 0;
-	cluster = blk_queue_cluster(q);
-
-	/*
-	 * for each bio in rq
-	 */
-	sg = NULL;
-	rq_for_each_segment(bvec, rq, iter) {
-		__blk_segment_map_sg(q, &bvec, sglist, &bvprv, &sg,
-				     &nsegs, &cluster);
-	} /* segments in rq */
+	struct scatterlist *sg = NULL;
+	int nsegs = 0;
 
+	if (rq->bio)
+		nsegs = __blk_bios_map_sg(q, rq->bio, sglist, &sg);
 
 	if (unlikely(rq->cmd_flags & REQ_COPY_USER) &&
 	    (blk_rq_bytes(rq) & q->dma_pad_mask)) {
@@ -230,20 +270,13 @@ EXPORT_SYMBOL(blk_rq_map_sg);
 int blk_bio_map_sg(struct request_queue *q, struct bio *bio,
 		   struct scatterlist *sglist)
 {
-	struct bio_vec bvec, bvprv = { NULL };
-	struct scatterlist *sg;
-	int nsegs, cluster;
-	struct bvec_iter iter;
-
-	nsegs = 0;
-	cluster = blk_queue_cluster(q);
-
-	sg = NULL;
-	bio_for_each_segment(bvec, bio, iter) {
-		__blk_segment_map_sg(q, &bvec, sglist, &bvprv, &sg,
-				     &nsegs, &cluster);
-	} /* segments in bio */
+	struct scatterlist *sg = NULL;
+	int nsegs;
+	struct bio *next = bio->bi_next;
+	bio->bi_next = NULL;
 
+	nsegs = __blk_bios_map_sg(q, bio, sglist, &sg);
+	bio->bi_next = next;
 	if (sg)
 		sg_mark_end(sg);
 
diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
index 5d70edc9855f..83ae96c51a27 100644
--- a/block/blk-mq-tag.c
+++ b/block/blk-mq-tag.c
@@ -184,7 +184,7 @@ void blk_mq_free_tags(struct blk_mq_tags *tags)
 ssize_t blk_mq_tag_sysfs_show(struct blk_mq_tags *tags, char *page)
 {
 	char *orig_page = page;
-	int cpu;
+	unsigned int cpu;
 
 	if (!tags)
 		return 0;
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 57039fcd9c93..1fa9dd153fde 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -226,15 +226,14 @@ static struct request *blk_mq_alloc_request_pinned(struct request_queue *q,
 	return rq;
 }
 
-struct request *blk_mq_alloc_request(struct request_queue *q, int rw,
-		gfp_t gfp, bool reserved)
+struct request *blk_mq_alloc_request(struct request_queue *q, int rw, gfp_t gfp)
 {
 	struct request *rq;
 
 	if (blk_mq_queue_enter(q))
 		return NULL;
 
-	rq = blk_mq_alloc_request_pinned(q, rw, gfp, reserved);
+	rq = blk_mq_alloc_request_pinned(q, rw, gfp, false);
 	if (rq)
 		blk_mq_put_ctx(rq->mq_ctx);
 	return rq;
@@ -258,7 +257,7 @@ EXPORT_SYMBOL(blk_mq_alloc_reserved_request);
 /*
  * Re-init and set pdu, if we have it
  */
-static void blk_mq_rq_init(struct blk_mq_hw_ctx *hctx, struct request *rq)
+void blk_mq_rq_init(struct blk_mq_hw_ctx *hctx, struct request *rq)
 {
 	blk_rq_init(hctx->queue, rq);
 
@@ -305,7 +304,7 @@ static void blk_mq_bio_endio(struct request *rq, struct bio *bio, int error)
 		bio_endio(bio, error);
 }
 
-void blk_mq_complete_request(struct request *rq, int error)
+void blk_mq_end_io(struct request *rq, int error)
 {
 	struct bio *bio = rq->bio;
 	unsigned int bytes = 0;
@@ -330,48 +329,55 @@ void blk_mq_complete_request(struct request *rq, int error)
 	else
 		blk_mq_free_request(rq);
 }
+EXPORT_SYMBOL(blk_mq_end_io);
 
-void __blk_mq_end_io(struct request *rq, int error)
-{
-	if (!blk_mark_rq_complete(rq))
-		blk_mq_complete_request(rq, error);
-}
-
-static void blk_mq_end_io_remote(void *data)
+static void __blk_mq_complete_request_remote(void *data)
 {
 	struct request *rq = data;
 
-	__blk_mq_end_io(rq, rq->errors);
+	rq->q->softirq_done_fn(rq);
 }
 
-/*
- * End IO on this request on a multiqueue enabled driver. We'll either do
- * it directly inline, or punt to a local IPI handler on the matching
- * remote CPU.
- */
-void blk_mq_end_io(struct request *rq, int error)
+void __blk_mq_complete_request(struct request *rq)
 {
 	struct blk_mq_ctx *ctx = rq->mq_ctx;
 	int cpu;
 
-	if (!ctx->ipi_redirect)
-		return __blk_mq_end_io(rq, error);
+	if (!ctx->ipi_redirect) {
+		rq->q->softirq_done_fn(rq);
+		return;
+	}
 
 	cpu = get_cpu();
 	if (cpu != ctx->cpu && cpu_online(ctx->cpu)) {
-		rq->errors = error;
-		rq->csd.func = blk_mq_end_io_remote;
+		rq->csd.func = __blk_mq_complete_request_remote;
 		rq->csd.info = rq;
 		rq->csd.flags = 0;
 		__smp_call_function_single(ctx->cpu, &rq->csd, 0);
 	} else {
-		__blk_mq_end_io(rq, error);
+		rq->q->softirq_done_fn(rq);
 	}
 	put_cpu();
 }
-EXPORT_SYMBOL(blk_mq_end_io);
 
-static void blk_mq_start_request(struct request *rq)
+/**
+ * blk_mq_complete_request - end I/O on a request
+ * @rq:		the request being processed
+ *
+ * Description:
+ *	Ends all I/O on a request. It does not handle partial completions.
+ *	The actual completion happens out-of-order, through a IPI handler.
+ **/
+void blk_mq_complete_request(struct request *rq)
+{
+	if (unlikely(blk_should_fake_timeout(rq->q)))
+		return;
+	if (!blk_mark_rq_complete(rq))
+		__blk_mq_complete_request(rq);
+}
+EXPORT_SYMBOL(blk_mq_complete_request);
+
+static void blk_mq_start_request(struct request *rq, bool last)
 {
 	struct request_queue *q = rq->q;
 
@@ -384,6 +390,25 @@ static void blk_mq_start_request(struct request *rq)
 	 */
 	rq->deadline = jiffies + q->rq_timeout;
 	set_bit(REQ_ATOM_STARTED, &rq->atomic_flags);
+
+	if (q->dma_drain_size && blk_rq_bytes(rq)) {
+		/*
+		 * Make sure space for the drain appears.  We know we can do
+		 * this because max_hw_segments has been adjusted to be one
+		 * fewer than the device can handle.
+		 */
+		rq->nr_phys_segments++;
+	}
+
+	/*
+	 * Flag the last request in the series so that drivers know when IO
+	 * should be kicked off, if they don't do it on a per-request basis.
+	 *
+	 * Note: the flag isn't the only condition drivers should do kick off.
+	 * If drive is busy, the last request might not have the bit set.
+	 */
+	if (last)
+		rq->cmd_flags |= REQ_END;
 }
 
 static void blk_mq_requeue_request(struct request *rq)
@@ -392,6 +417,11 @@ static void blk_mq_requeue_request(struct request *rq)
 
 	trace_block_rq_requeue(q, rq);
 	clear_bit(REQ_ATOM_STARTED, &rq->atomic_flags);
+
+	rq->cmd_flags &= ~REQ_END;
+
+	if (q->dma_drain_size && blk_rq_bytes(rq))
+		rq->nr_phys_segments--;
 }
 
 struct blk_mq_timeout_data {
@@ -559,19 +589,8 @@ static void __blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx)
 
 		rq = list_first_entry(&rq_list, struct request, queuelist);
 		list_del_init(&rq->queuelist);
-		blk_mq_start_request(rq);
 
-		/*
-		 * Last request in the series. Flag it as such, this
-		 * enables drivers to know when IO should be kicked off,
-		 * if they don't do it on a per-request basis.
-		 *
-		 * Note: the flag isn't the only condition drivers
-		 * should do kick off. If drive is busy, the last
-		 * request might not have the bit set.
-		 */
-		if (list_empty(&rq_list))
-			rq->cmd_flags |= REQ_END;
+		blk_mq_start_request(rq, list_empty(&rq_list));
 
 		ret = q->mq_ops->queue_rq(hctx, rq);
 		switch (ret) {
@@ -589,8 +608,8 @@ static void __blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx)
 			break;
 		default:
 			pr_err("blk-mq: bad return on queue: %d\n", ret);
-			rq->errors = -EIO;
 		case BLK_MQ_RQ_QUEUE_ERROR:
+			rq->errors = -EIO;
 			blk_mq_end_io(rq, rq->errors);
 			break;
 		}
@@ -693,13 +712,16 @@ static void blk_mq_work_fn(struct work_struct *work)
 }
 
 static void __blk_mq_insert_request(struct blk_mq_hw_ctx *hctx,
-				    struct request *rq)
+				    struct request *rq, bool at_head)
 {
 	struct blk_mq_ctx *ctx = rq->mq_ctx;
 
 	trace_block_rq_insert(hctx->queue, rq);
 
-	list_add_tail(&rq->queuelist, &ctx->rq_list);
+	if (at_head)
+		list_add(&rq->queuelist, &ctx->rq_list);
+	else
+		list_add_tail(&rq->queuelist, &ctx->rq_list);
 	blk_mq_hctx_mark_pending(hctx, ctx);
 
 	/*
@@ -709,7 +731,7 @@ static void __blk_mq_insert_request(struct blk_mq_hw_ctx *hctx,
 }
 
 void blk_mq_insert_request(struct request_queue *q, struct request *rq,
-			   bool run_queue)
+			   bool at_head, bool run_queue)
 {
 	struct blk_mq_hw_ctx *hctx;
 	struct blk_mq_ctx *ctx, *current_ctx;
@@ -728,7 +750,7 @@ void blk_mq_insert_request(struct request_queue *q, struct request *rq,
 			rq->mq_ctx = ctx;
 		}
 		spin_lock(&ctx->lock);
-		__blk_mq_insert_request(hctx, rq);
+		__blk_mq_insert_request(hctx, rq, at_head);
 		spin_unlock(&ctx->lock);
 
 		blk_mq_put_ctx(current_ctx);
@@ -760,7 +782,7 @@ void blk_mq_run_request(struct request *rq, bool run_queue, bool async)
 
 	/* ctx->cpu might be offline */
 	spin_lock(&ctx->lock);
-	__blk_mq_insert_request(hctx, rq);
+	__blk_mq_insert_request(hctx, rq, false);
 	spin_unlock(&ctx->lock);
 
 	blk_mq_put_ctx(current_ctx);
@@ -798,7 +820,7 @@ static void blk_mq_insert_requests(struct request_queue *q,
 		rq = list_first_entry(list, struct request, queuelist);
 		list_del_init(&rq->queuelist);
 		rq->mq_ctx = ctx;
-		__blk_mq_insert_request(hctx, rq);
+		__blk_mq_insert_request(hctx, rq, false);
 	}
 	spin_unlock(&ctx->lock);
 
@@ -888,6 +910,11 @@ static void blk_mq_make_request(struct request_queue *q, struct bio *bio)
 
 	blk_queue_bounce(q, &bio);
 
+	if (bio_integrity_enabled(bio) && bio_integrity_prep(bio)) {
+		bio_endio(bio, -EIO);
+		return;
+	}
+
 	if (use_plug && blk_attempt_plug_merge(q, bio, &request_count))
 		return;
 
@@ -950,7 +977,7 @@ static void blk_mq_make_request(struct request_queue *q, struct bio *bio)
 		__blk_mq_free_request(hctx, ctx, rq);
 	else {
 		blk_mq_bio_to_request(rq, bio);
-		__blk_mq_insert_request(hctx, rq);
+		__blk_mq_insert_request(hctx, rq, false);
 	}
 
 	spin_unlock(&ctx->lock);
@@ -1309,15 +1336,6 @@ struct request_queue *blk_mq_init_queue(struct blk_mq_reg *reg,
 		reg->queue_depth = BLK_MQ_MAX_DEPTH;
 	}
 
-	/*
-	 * Set aside a tag for flush requests.  It will only be used while
-	 * another flush request is in progress but outside the driver.
-	 *
-	 * TODO: only allocate if flushes are supported
-	 */
-	reg->queue_depth++;
-	reg->reserved_tags++;
-
 	if (reg->queue_depth < (reg->reserved_tags + BLK_MQ_TAG_MIN))
 		return ERR_PTR(-EINVAL);
 
@@ -1360,17 +1378,27 @@ struct request_queue *blk_mq_init_queue(struct blk_mq_reg *reg,
 	q->mq_ops = reg->ops;
 	q->queue_flags |= QUEUE_FLAG_MQ_DEFAULT;
 
+	q->sg_reserved_size = INT_MAX;
+
 	blk_queue_make_request(q, blk_mq_make_request);
 	blk_queue_rq_timed_out(q, reg->ops->timeout);
 	if (reg->timeout)
 		blk_queue_rq_timeout(q, reg->timeout);
 
+	if (reg->ops->complete)
+		blk_queue_softirq_done(q, reg->ops->complete);
+
 	blk_mq_init_flush(q);
 	blk_mq_init_cpu_queues(q, reg->nr_hw_queues);
 
-	if (blk_mq_init_hw_queues(q, reg, driver_data))
+	q->flush_rq = kzalloc(round_up(sizeof(struct request) + reg->cmd_size,
+				cache_line_size()), GFP_KERNEL);
+	if (!q->flush_rq)
 		goto err_hw;
 
+	if (blk_mq_init_hw_queues(q, reg, driver_data))
+		goto err_flush_rq;
+
 	blk_mq_map_swqueue(q);
 
 	mutex_lock(&all_q_mutex);
@@ -1378,6 +1406,9 @@ struct request_queue *blk_mq_init_queue(struct blk_mq_reg *reg,
 	mutex_unlock(&all_q_mutex);
 
 	return q;
+
+err_flush_rq:
+	kfree(q->flush_rq);
 err_hw:
 	kfree(q->mq_map);
 err_map:
diff --git a/block/blk-mq.h b/block/blk-mq.h
index 5c3917984b00..ed0035cd458e 100644
--- a/block/blk-mq.h
+++ b/block/blk-mq.h
@@ -22,13 +22,13 @@ struct blk_mq_ctx {
 	struct kobject		kobj;
 };
 
-void __blk_mq_end_io(struct request *rq, int error);
-void blk_mq_complete_request(struct request *rq, int error);
+void __blk_mq_complete_request(struct request *rq);
 void blk_mq_run_request(struct request *rq, bool run_queue, bool async);
 void blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async);
 void blk_mq_init_flush(struct request_queue *q);
 void blk_mq_drain_queue(struct request_queue *q);
 void blk_mq_free_queue(struct request_queue *q);
+void blk_mq_rq_init(struct blk_mq_hw_ctx *hctx, struct request *rq);
 
 /*
  * CPU hotplug helpers
diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c
index 8095c4a21fc0..7500f876dae4 100644
--- a/block/blk-sysfs.c
+++ b/block/blk-sysfs.c
@@ -549,6 +549,8 @@ static void blk_release_queue(struct kobject *kobj)
 	if (q->mq_ops)
 		blk_mq_free_queue(q);
 
+	kfree(q->flush_rq);
+
 	blk_trace_shutdown(q);
 
 	bdi_destroy(&q->backing_dev_info);
diff --git a/block/blk-timeout.c b/block/blk-timeout.c
index bba81c9348e1..d96f7061c6fd 100644
--- a/block/blk-timeout.c
+++ b/block/blk-timeout.c
@@ -91,7 +91,7 @@ static void blk_rq_timed_out(struct request *req)
 	case BLK_EH_HANDLED:
 		/* Can we use req->errors here? */
 		if (q->mq_ops)
-			blk_mq_complete_request(req, req->errors);
+			__blk_mq_complete_request(req);
 		else
 			__blk_complete_request(req);
 		break;
diff --git a/block/blk.h b/block/blk.h
index c90e1d8f7a2b..d23b415b8a28 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -113,7 +113,7 @@ static inline struct request *__elv_next_request(struct request_queue *q)
 			q->flush_queue_delayed = 1;
 			return NULL;
 		}
-		if (unlikely(blk_queue_dying(q)) ||
+		if (unlikely(blk_queue_bypass(q)) ||
 		    !q->elevator->type->ops.elevator_dispatch_fn(q, 0))
 			return NULL;
 	}
diff --git a/drivers/block/null_blk.c b/drivers/block/null_blk.c
index 3107282a9741..091b9ea14feb 100644
--- a/drivers/block/null_blk.c
+++ b/drivers/block/null_blk.c
@@ -60,7 +60,9 @@ enum {
 	NULL_IRQ_NONE		= 0,
 	NULL_IRQ_SOFTIRQ	= 1,
 	NULL_IRQ_TIMER		= 2,
+};
 
+enum {
 	NULL_Q_BIO		= 0,
 	NULL_Q_RQ		= 1,
 	NULL_Q_MQ		= 2,
@@ -172,18 +174,20 @@ static struct nullb_cmd *alloc_cmd(struct nullb_queue *nq, int can_wait)
 
 static void end_cmd(struct nullb_cmd *cmd)
 {
-	if (cmd->rq) {
-		if (queue_mode == NULL_Q_MQ)
-			blk_mq_end_io(cmd->rq, 0);
-		else {
-			INIT_LIST_HEAD(&cmd->rq->queuelist);
-			blk_end_request_all(cmd->rq, 0);
-		}
-	} else if (cmd->bio)
+	switch (queue_mode)  {
+	case NULL_Q_MQ:
+		blk_mq_end_io(cmd->rq, 0);
+		return;
+	case NULL_Q_RQ:
+		INIT_LIST_HEAD(&cmd->rq->queuelist);
+		blk_end_request_all(cmd->rq, 0);
+		break;
+	case NULL_Q_BIO:
 		bio_endio(cmd->bio, 0);
+		break;
+	}
 
-	if (queue_mode != NULL_Q_MQ)
-		free_cmd(cmd);
+	free_cmd(cmd);
 }
 
 static enum hrtimer_restart null_cmd_timer_expired(struct hrtimer *timer)
@@ -195,6 +199,7 @@ static enum hrtimer_restart null_cmd_timer_expired(struct hrtimer *timer)
 	cq = &per_cpu(completion_queues, smp_processor_id());
 
 	while ((entry = llist_del_all(&cq->list)) != NULL) {
+		entry = llist_reverse_order(entry);
 		do {
 			cmd = container_of(entry, struct nullb_cmd, ll_list);
 			end_cmd(cmd);
@@ -221,62 +226,32 @@ static void null_cmd_end_timer(struct nullb_cmd *cmd)
 
 static void null_softirq_done_fn(struct request *rq)
 {
-	blk_end_request_all(rq, 0);
+	end_cmd(rq->special);
 }
 
-#ifdef CONFIG_SMP
-
-static void null_ipi_cmd_end_io(void *data)
-{
-	struct completion_queue *cq;
-	struct llist_node *entry, *next;
-	struct nullb_cmd *cmd;
-
-	cq = &per_cpu(completion_queues, smp_processor_id());
-
-	entry = llist_del_all(&cq->list);
-
-	while (entry) {
-		next = entry->next;
-		cmd = llist_entry(entry, struct nullb_cmd, ll_list);
-		end_cmd(cmd);
-		entry = next;
-	}
-}
-
-static void null_cmd_end_ipi(struct nullb_cmd *cmd)
-{
-	struct call_single_data *data = &cmd->csd;
-	int cpu = get_cpu();
-	struct completion_queue *cq = &per_cpu(completion_queues, cpu);
-
-	cmd->ll_list.next = NULL;
-
-	if (llist_add(&cmd->ll_list, &cq->list)) {
-		data->func = null_ipi_cmd_end_io;
-		data->flags = 0;
-		__smp_call_function_single(cpu, data, 0);
-	}
-
-	put_cpu();
-}
-
-#endif /* CONFIG_SMP */
-
 static inline void null_handle_cmd(struct nullb_cmd *cmd)
 {
 	/* Complete IO by inline, softirq or timer */
 	switch (irqmode) {
+	case NULL_IRQ_SOFTIRQ:
+		switch (queue_mode)  {
+		case NULL_Q_MQ:
+			blk_mq_complete_request(cmd->rq);
+			break;
+		case NULL_Q_RQ:
+			blk_complete_request(cmd->rq);
+			break;
+		case NULL_Q_BIO:
+			/*
+			 * XXX: no proper submitting cpu information available.
+			 */
+			end_cmd(cmd);
+			break;
+		}
+		break;
 	case NULL_IRQ_NONE:
 		end_cmd(cmd);
 		break;
-	case NULL_IRQ_SOFTIRQ:
-#ifdef CONFIG_SMP
-		null_cmd_end_ipi(cmd);
-#else
-		end_cmd(cmd);
-#endif
-		break;
 	case NULL_IRQ_TIMER:
 		null_cmd_end_timer(cmd);
 		break;
@@ -411,6 +386,7 @@ static struct blk_mq_ops null_mq_ops = {
 	.queue_rq       = null_queue_rq,
 	.map_queue      = blk_mq_map_queue,
 	.init_hctx	= null_init_hctx,
+	.complete	= null_softirq_done_fn,
 };
 
 static struct blk_mq_reg null_mq_reg = {
@@ -609,13 +585,6 @@ static int __init null_init(void)
 {
 	unsigned int i;
 
-#if !defined(CONFIG_SMP)
-	if (irqmode == NULL_IRQ_SOFTIRQ) {
-		pr_warn("null_blk: softirq completions not available.\n");
-		pr_warn("null_blk: using direct completions.\n");
-		irqmode = NULL_IRQ_NONE;
-	}
-#endif
 	if (bs > PAGE_SIZE) {
 		pr_warn("null_blk: invalid block size\n");
 		pr_warn("null_blk: defaults block size to %lu\n", PAGE_SIZE);
diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index 6a680d4de7f1..b1cb3f4c4db4 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -110,9 +110,9 @@ static int __virtblk_add_req(struct virtqueue *vq,
 	return virtqueue_add_sgs(vq, sgs, num_out, num_in, vbr, GFP_ATOMIC);
 }
 
-static inline void virtblk_request_done(struct virtblk_req *vbr)
+static inline void virtblk_request_done(struct request *req)
 {
-	struct request *req = vbr->req;
+	struct virtblk_req *vbr = req->special;
 	int error = virtblk_result(vbr);
 
 	if (req->cmd_type == REQ_TYPE_BLOCK_PC) {
@@ -138,7 +138,7 @@ static void virtblk_done(struct virtqueue *vq)
 	do {
 		virtqueue_disable_cb(vq);
 		while ((vbr = virtqueue_get_buf(vblk->vq, &len)) != NULL) {
-			virtblk_request_done(vbr);
+			blk_mq_complete_request(vbr->req);
 			req_done = true;
 		}
 		if (unlikely(virtqueue_is_broken(vq)))
@@ -479,6 +479,7 @@ static struct blk_mq_ops virtio_mq_ops = {
 	.map_queue	= blk_mq_map_queue,
 	.alloc_hctx	= blk_mq_alloc_single_hw_queue,
 	.free_hctx	= blk_mq_free_single_hw_queue,
+	.complete	= virtblk_request_done,
 };
 
 static struct blk_mq_reg virtio_mq_reg = {
diff --git a/drivers/block/xen-blkback/blkback.c b/drivers/block/xen-blkback/blkback.c
index 4b97b86da926..64c60edcdfbc 100644
--- a/drivers/block/xen-blkback/blkback.c
+++ b/drivers/block/xen-blkback/blkback.c
@@ -299,7 +299,7 @@ static void free_persistent_gnts(struct xen_blkif *blkif, struct rb_root *root,
 	BUG_ON(num != 0);
 }
 
-static void unmap_purged_grants(struct work_struct *work)
+void xen_blkbk_unmap_purged_grants(struct work_struct *work)
 {
 	struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 	struct page *pages[BLKIF_MAX_SEGMENTS_PER_REQUEST];
@@ -375,7 +375,7 @@ static void purge_persistent_gnt(struct xen_blkif *blkif)
 
 	pr_debug(DRV_PFX "Going to purge %u persistent grants\n", num_clean);
 
-	INIT_LIST_HEAD(&blkif->persistent_purge_list);
+	BUG_ON(!list_empty(&blkif->persistent_purge_list));
 	root = &blkif->persistent_gnts;
 purge_list:
 	foreach_grant_safe(persistent_gnt, n, root, node) {
@@ -420,7 +420,6 @@ finished:
 	blkif->vbd.overflow_max_grants = 0;
 
 	/* We can defer this work */
-	INIT_WORK(&blkif->persistent_purge_work, unmap_purged_grants);
 	schedule_work(&blkif->persistent_purge_work);
 	pr_debug(DRV_PFX "Purged %u/%u\n", (total - num_clean), total);
 	return;
@@ -625,16 +624,8 @@ purge_gnt_list:
 			print_stats(blkif);
 	}
 
-	/* Since we are shutting down remove all pages from the buffer */
-	shrink_free_pagepool(blkif, 0 /* All */);
-
-	/* Free all persistent grant pages */
-	if (!RB_EMPTY_ROOT(&blkif->persistent_gnts))
-		free_persistent_gnts(blkif, &blkif->persistent_gnts,
-			blkif->persistent_gnt_c);
-
-	BUG_ON(!RB_EMPTY_ROOT(&blkif->persistent_gnts));
-	blkif->persistent_gnt_c = 0;
+	/* Drain pending purge work */
+	flush_work(&blkif->persistent_purge_work);
 
 	if (log_stats)
 		print_stats(blkif);
@@ -645,6 +636,23 @@ purge_gnt_list:
 	return 0;
 }
 
+/*
+ * Remove persistent grants and empty the pool of free pages
+ */
+void xen_blkbk_free_caches(struct xen_blkif *blkif)
+{
+	/* Free all persistent grant pages */
+	if (!RB_EMPTY_ROOT(&blkif->persistent_gnts))
+		free_persistent_gnts(blkif, &blkif->persistent_gnts,
+			blkif->persistent_gnt_c);
+
+	BUG_ON(!RB_EMPTY_ROOT(&blkif->persistent_gnts));
+	blkif->persistent_gnt_c = 0;
+
+	/* Since we are shutting down remove all pages from the buffer */
+	shrink_free_pagepool(blkif, 0 /* All */);
+}
+
 /*
  * Unmap the grant references, and also remove the M2P over-rides
  * used in the 'pending_req'.
@@ -838,7 +846,7 @@ static int xen_blkbk_parse_indirect(struct blkif_request *req,
 	struct grant_page **pages = pending_req->indirect_pages;
 	struct xen_blkif *blkif = pending_req->blkif;
 	int indirect_grefs, rc, n, nseg, i;
-	struct blkif_request_segment_aligned *segments = NULL;
+	struct blkif_request_segment *segments = NULL;
 
 	nseg = pending_req->nr_pages;
 	indirect_grefs = INDIRECT_PAGES(nseg);
@@ -934,9 +942,7 @@ static void xen_blk_drain_io(struct xen_blkif *blkif)
 {
 	atomic_set(&blkif->drain, 1);
 	do {
-		/* The initial value is one, and one refcnt taken at the
-		 * start of the xen_blkif_schedule thread. */
-		if (atomic_read(&blkif->refcnt) <= 2)
+		if (atomic_read(&blkif->inflight) == 0)
 			break;
 		wait_for_completion_interruptible_timeout(
 				&blkif->drain_complete, HZ);
@@ -976,17 +982,30 @@ static void __end_block_io_op(struct pending_req *pending_req, int error)
 	 * the proper response on the ring.
 	 */
 	if (atomic_dec_and_test(&pending_req->pendcnt)) {
-		xen_blkbk_unmap(pending_req->blkif,
+		struct xen_blkif *blkif = pending_req->blkif;
+
+		xen_blkbk_unmap(blkif,
 		                pending_req->segments,
 		                pending_req->nr_pages);
-		make_response(pending_req->blkif, pending_req->id,
+		make_response(blkif, pending_req->id,
 			      pending_req->operation, pending_req->status);
-		xen_blkif_put(pending_req->blkif);
-		if (atomic_read(&pending_req->blkif->refcnt) <= 2) {
-			if (atomic_read(&pending_req->blkif->drain))
-				complete(&pending_req->blkif->drain_complete);
+		free_req(blkif, pending_req);
+		/*
+		 * Make sure the request is freed before releasing blkif,
+		 * or there could be a race between free_req and the
+		 * cleanup done in xen_blkif_free during shutdown.
+		 *
+		 * NB: The fact that we might try to wake up pending_free_wq
+		 * before drain_complete (in case there's a drain going on)
+		 * it's not a problem with our current implementation
+		 * because we can assure there's no thread waiting on
+		 * pending_free_wq if there's a drain going on, but it has
+		 * to be taken into account if the current model is changed.
+		 */
+		if (atomic_dec_and_test(&blkif->inflight) && atomic_read(&blkif->drain)) {
+			complete(&blkif->drain_complete);
 		}
-		free_req(pending_req->blkif, pending_req);
+		xen_blkif_put(blkif);
 	}
 }
 
@@ -1240,6 +1259,7 @@ static int dispatch_rw_block_io(struct xen_blkif *blkif,
 	 * below (in "!bio") if we are handling a BLKIF_OP_DISCARD.
 	 */
 	xen_blkif_get(blkif);
+	atomic_inc(&blkif->inflight);
 
 	for (i = 0; i < nseg; i++) {
 		while ((bio == NULL) ||
diff --git a/drivers/block/xen-blkback/common.h b/drivers/block/xen-blkback/common.h
index 8d8807563d99..be052773ad03 100644
--- a/drivers/block/xen-blkback/common.h
+++ b/drivers/block/xen-blkback/common.h
@@ -57,7 +57,7 @@
 #define MAX_INDIRECT_SEGMENTS 256
 
 #define SEGS_PER_INDIRECT_FRAME \
-	(PAGE_SIZE/sizeof(struct blkif_request_segment_aligned))
+	(PAGE_SIZE/sizeof(struct blkif_request_segment))
 #define MAX_INDIRECT_PAGES \
 	((MAX_INDIRECT_SEGMENTS + SEGS_PER_INDIRECT_FRAME - 1)/SEGS_PER_INDIRECT_FRAME)
 #define INDIRECT_PAGES(_segs) \
@@ -278,6 +278,7 @@ struct xen_blkif {
 	/* for barrier (drain) requests */
 	struct completion	drain_complete;
 	atomic_t		drain;
+	atomic_t		inflight;
 	/* One thread per one blkif. */
 	struct task_struct	*xenblkd;
 	unsigned int		waiting_reqs;
@@ -376,6 +377,7 @@ int xen_blkif_xenbus_init(void);
 irqreturn_t xen_blkif_be_int(int irq, void *dev_id);
 int xen_blkif_schedule(void *arg);
 int xen_blkif_purge_persistent(void *arg);
+void xen_blkbk_free_caches(struct xen_blkif *blkif);
 
 int xen_blkbk_flush_diskcache(struct xenbus_transaction xbt,
 			      struct backend_info *be, int state);
@@ -383,6 +385,7 @@ int xen_blkbk_flush_diskcache(struct xenbus_transaction xbt,
 int xen_blkbk_barrier(struct xenbus_transaction xbt,
 		      struct backend_info *be, int state);
 struct xenbus_device *xen_blkbk_xenbus(struct backend_info *be);
+void xen_blkbk_unmap_purged_grants(struct work_struct *work);
 
 static inline void blkif_get_x86_32_req(struct blkif_request *dst,
 					struct blkif_x86_32_request *src)
diff --git a/drivers/block/xen-blkback/xenbus.c b/drivers/block/xen-blkback/xenbus.c
index c2014a0aa206..9a547e6b6ebf 100644
--- a/drivers/block/xen-blkback/xenbus.c
+++ b/drivers/block/xen-blkback/xenbus.c
@@ -125,8 +125,11 @@ static struct xen_blkif *xen_blkif_alloc(domid_t domid)
 	blkif->persistent_gnts.rb_node = NULL;
 	spin_lock_init(&blkif->free_pages_lock);
 	INIT_LIST_HEAD(&blkif->free_pages);
+	INIT_LIST_HEAD(&blkif->persistent_purge_list);
 	blkif->free_pages_num = 0;
 	atomic_set(&blkif->persistent_gnt_in_use, 0);
+	atomic_set(&blkif->inflight, 0);
+	INIT_WORK(&blkif->persistent_purge_work, xen_blkbk_unmap_purged_grants);
 
 	INIT_LIST_HEAD(&blkif->pending_free);
 
@@ -259,6 +262,17 @@ static void xen_blkif_free(struct xen_blkif *blkif)
 	if (!atomic_dec_and_test(&blkif->refcnt))
 		BUG();
 
+	/* Remove all persistent grants and the cache of ballooned pages. */
+	xen_blkbk_free_caches(blkif);
+
+	/* Make sure everything is drained before shutting down */
+	BUG_ON(blkif->persistent_gnt_c != 0);
+	BUG_ON(atomic_read(&blkif->persistent_gnt_in_use) != 0);
+	BUG_ON(blkif->free_pages_num != 0);
+	BUG_ON(!list_empty(&blkif->persistent_purge_list));
+	BUG_ON(!list_empty(&blkif->free_pages));
+	BUG_ON(!RB_EMPTY_ROOT(&blkif->persistent_gnts));
+
 	/* Check that there is no request in use */
 	list_for_each_entry_safe(req, n, &blkif->pending_free, free_list) {
 		list_del(&req->free_list);
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 8dcfb54f1603..efe1b4761735 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -162,7 +162,7 @@ static DEFINE_SPINLOCK(minor_lock);
 #define DEV_NAME	"xvd"	/* name in /dev */
 
 #define SEGS_PER_INDIRECT_FRAME \
-	(PAGE_SIZE/sizeof(struct blkif_request_segment_aligned))
+	(PAGE_SIZE/sizeof(struct blkif_request_segment))
 #define INDIRECT_GREFS(_segs) \
 	((_segs + SEGS_PER_INDIRECT_FRAME - 1)/SEGS_PER_INDIRECT_FRAME)
 
@@ -393,7 +393,7 @@ static int blkif_queue_request(struct request *req)
 	unsigned long id;
 	unsigned int fsect, lsect;
 	int i, ref, n;
-	struct blkif_request_segment_aligned *segments = NULL;
+	struct blkif_request_segment *segments = NULL;
 
 	/*
 	 * Used to store if we are able to queue the request by just using
@@ -550,7 +550,7 @@ static int blkif_queue_request(struct request *req)
 			} else {
 				n = i % SEGS_PER_INDIRECT_FRAME;
 				segments[n] =
-					(struct blkif_request_segment_aligned) {
+					(struct blkif_request_segment) {
 							.gref       = ref,
 							.first_sect = fsect,
 							.last_sect  = lsect };
@@ -1904,13 +1904,16 @@ static void blkback_changed(struct xenbus_device *dev,
 	case XenbusStateReconfiguring:
 	case XenbusStateReconfigured:
 	case XenbusStateUnknown:
-	case XenbusStateClosed:
 		break;
 
 	case XenbusStateConnected:
 		blkfront_connect(info);
 		break;
 
+	case XenbusStateClosed:
+		if (dev->state == XenbusStateClosed)
+			break;
+		/* Missed the backend's Closing state -- fallthrough */
 	case XenbusStateClosing:
 		blkfront_closing(info);
 		break;
diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index 0c707e4f4eaf..a4c7306ff43d 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -210,7 +210,9 @@ BITMASK(GC_MARK,	 struct bucket, gc_mark, 0, 2);
 #define GC_MARK_RECLAIMABLE	0
 #define GC_MARK_DIRTY		1
 #define GC_MARK_METADATA	2
-BITMASK(GC_SECTORS_USED, struct bucket, gc_mark, 2, 13);
+#define GC_SECTORS_USED_SIZE	13
+#define MAX_GC_SECTORS_USED	(~(~0ULL << GC_SECTORS_USED_SIZE))
+BITMASK(GC_SECTORS_USED, struct bucket, gc_mark, 2, GC_SECTORS_USED_SIZE);
 BITMASK(GC_MOVE, struct bucket, gc_mark, 15, 1);
 
 #include "journal.h"
diff --git a/drivers/md/bcache/bset.c b/drivers/md/bcache/bset.c
index 4f6b5940e609..3f74b4b0747b 100644
--- a/drivers/md/bcache/bset.c
+++ b/drivers/md/bcache/bset.c
@@ -23,7 +23,7 @@ void bch_dump_bset(struct btree_keys *b, struct bset *i, unsigned set)
 	for (k = i->start; k < bset_bkey_last(i); k = next) {
 		next = bkey_next(k);
 
-		printk(KERN_ERR "block %u key %zi/%u: ", set,
+		printk(KERN_ERR "block %u key %li/%u: ", set,
 		       (uint64_t *) k - i->d, i->keys);
 
 		if (b->ops->key_dump)
@@ -1185,9 +1185,12 @@ static void __btree_sort(struct btree_keys *b, struct btree_iter *iter,
 	struct bset *out = (void *) __get_free_pages(__GFP_NOWARN|GFP_NOIO,
 						     order);
 	if (!out) {
+		struct page *outp;
+
 		BUG_ON(order > state->page_order);
 
-		out = page_address(mempool_alloc(state->pool, GFP_NOIO));
+		outp = mempool_alloc(state->pool, GFP_NOIO);
+		out = page_address(outp);
 		used_mempool = true;
 		order = state->page_order;
 	}
diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index 98cc0a810a36..5f9c2a665ca5 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -1167,7 +1167,7 @@ uint8_t __bch_btree_mark_key(struct cache_set *c, int level, struct bkey *k)
 		/* guard against overflow */
 		SET_GC_SECTORS_USED(g, min_t(unsigned,
 					     GC_SECTORS_USED(g) + KEY_SIZE(k),
-					     (1 << 14) - 1));
+					     MAX_GC_SECTORS_USED));
 
 		BUG_ON(!GC_SECTORS_USED(g));
 	}
@@ -1805,7 +1805,7 @@ static bool btree_insert_key(struct btree *b, struct bkey *k,
 
 static size_t insert_u64s_remaining(struct btree *b)
 {
-	ssize_t ret = bch_btree_keys_u64s_remaining(&b->keys);
+	long ret = bch_btree_keys_u64s_remaining(&b->keys);
 
 	/*
 	 * Might land in the middle of an existing extent and have to split it
diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index 72cd213f213f..5d5d031cf381 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -353,14 +353,14 @@ static void bch_data_insert_start(struct closure *cl)
 	struct data_insert_op *op = container_of(cl, struct data_insert_op, cl);
 	struct bio *bio = op->bio, *n;
 
-	if (op->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 (op->bypass)
+		return bch_data_invalidate(cl);
+
 	/*
 	 * Journal writes are marked REQ_FLUSH; if the original write was a
 	 * flush, it'll wait on the journal write.
diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c
index c6ab69333a6d..d8458d477a12 100644
--- a/drivers/md/bcache/sysfs.c
+++ b/drivers/md/bcache/sysfs.c
@@ -416,7 +416,7 @@ static int btree_bset_stats(struct btree_op *b_op, struct btree *b)
 	return MAP_CONTINUE;
 }
 
-int bch_bset_print_stats(struct cache_set *c, char *buf)
+static int bch_bset_print_stats(struct cache_set *c, char *buf)
 {
 	struct bset_stats_op op;
 	int ret;
diff --git a/fs/bio-integrity.c b/fs/bio-integrity.c
index 0bad24ddc2e7..0129b78a6908 100644
--- a/fs/bio-integrity.c
+++ b/fs/bio-integrity.c
@@ -114,6 +114,14 @@ void bio_integrity_free(struct bio *bio)
 }
 EXPORT_SYMBOL(bio_integrity_free);
 
+static inline unsigned int bip_integrity_vecs(struct bio_integrity_payload *bip)
+{
+	if (bip->bip_slab == BIO_POOL_NONE)
+		return BIP_INLINE_VECS;
+
+	return bvec_nr_vecs(bip->bip_slab);
+}
+
 /**
  * bio_integrity_add_page - Attach integrity metadata
  * @bio:	bio to update
@@ -129,7 +137,7 @@ int bio_integrity_add_page(struct bio *bio, struct page *page,
 	struct bio_integrity_payload *bip = bio->bi_integrity;
 	struct bio_vec *iv;
 
-	if (bip->bip_vcnt >= bvec_nr_vecs(bip->bip_slab)) {
+	if (bip->bip_vcnt >= bip_integrity_vecs(bip)) {
 		printk(KERN_ERR "%s: bip_vec full\n", __func__);
 		return 0;
 	}
@@ -226,7 +234,8 @@ unsigned int bio_integrity_tag_size(struct bio *bio)
 }
 EXPORT_SYMBOL(bio_integrity_tag_size);
 
-int bio_integrity_tag(struct bio *bio, void *tag_buf, unsigned int len, int set)
+static int bio_integrity_tag(struct bio *bio, void *tag_buf, unsigned int len,
+			     int set)
 {
 	struct bio_integrity_payload *bip = bio->bi_integrity;
 	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
diff --git a/fs/bio.c b/fs/bio.c
index 75c49a382239..8754e7b6eb49 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -611,7 +611,6 @@ EXPORT_SYMBOL(bio_clone_fast);
 struct bio *bio_clone_bioset(struct bio *bio_src, gfp_t gfp_mask,
 			     struct bio_set *bs)
 {
-	unsigned nr_iovecs = 0;
 	struct bvec_iter iter;
 	struct bio_vec bv;
 	struct bio *bio;
@@ -638,10 +637,7 @@ struct bio *bio_clone_bioset(struct bio *bio_src, gfp_t gfp_mask,
 	 *    __bio_clone_fast() anyways.
 	 */
 
-	bio_for_each_segment(bv, bio_src, iter)
-		nr_iovecs++;
-
-	bio = bio_alloc_bioset(gfp_mask, nr_iovecs, bs);
+	bio = bio_alloc_bioset(gfp_mask, bio_segments(bio_src), bs);
 	if (!bio)
 		return NULL;
 
@@ -650,9 +646,18 @@ struct bio *bio_clone_bioset(struct bio *bio_src, gfp_t gfp_mask,
 	bio->bi_iter.bi_sector	= bio_src->bi_iter.bi_sector;
 	bio->bi_iter.bi_size	= bio_src->bi_iter.bi_size;
 
+	if (bio->bi_rw & REQ_DISCARD)
+		goto integrity_clone;
+
+	if (bio->bi_rw & REQ_WRITE_SAME) {
+		bio->bi_io_vec[bio->bi_vcnt++] = bio_src->bi_io_vec[0];
+		goto integrity_clone;
+	}
+
 	bio_for_each_segment(bv, bio_src, iter)
 		bio->bi_io_vec[bio->bi_vcnt++] = bv;
 
+integrity_clone:
 	if (bio_integrity(bio_src)) {
 		int ret;
 
diff --git a/include/linux/bio.h b/include/linux/bio.h
index 70654521dab6..5a4d39b4686b 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -250,6 +250,17 @@ static inline unsigned bio_segments(struct bio *bio)
 	struct bio_vec bv;
 	struct bvec_iter iter;
 
+	/*
+	 * We special case discard/write same, because they interpret bi_size
+	 * differently:
+	 */
+
+	if (bio->bi_rw & REQ_DISCARD)
+		return 1;
+
+	if (bio->bi_rw & REQ_WRITE_SAME)
+		return 1;
+
 	bio_for_each_segment(bv, bio, iter)
 		segs++;
 
@@ -332,6 +343,7 @@ extern struct bio *bio_clone_fast(struct bio *, gfp_t, struct bio_set *);
 extern struct bio *bio_clone_bioset(struct bio *, gfp_t, struct bio_set *bs);
 
 extern struct bio_set *fs_bio_set;
+unsigned int bio_integrity_tag_size(struct bio *bio);
 
 static inline struct bio *bio_alloc(gfp_t gfp_mask, unsigned int nr_iovecs)
 {
diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h
index 161b23105b1e..18ba8a627f46 100644
--- a/include/linux/blk-mq.h
+++ b/include/linux/blk-mq.h
@@ -83,6 +83,8 @@ struct blk_mq_ops {
 	 */
 	rq_timed_out_fn		*timeout;
 
+	softirq_done_fn		*complete;
+
 	/*
 	 * Override for hctx allocations (should probably go)
 	 */
@@ -119,11 +121,12 @@ void blk_mq_init_commands(struct request_queue *, void (*init)(void *data, struc
 
 void blk_mq_flush_plug_list(struct blk_plug *plug, bool from_schedule);
 
-void blk_mq_insert_request(struct request_queue *, struct request *, bool);
+void blk_mq_insert_request(struct request_queue *, struct request *,
+		bool, bool);
 void blk_mq_run_queues(struct request_queue *q, bool async);
 void blk_mq_free_request(struct request *rq);
 bool blk_mq_can_queue(struct blk_mq_hw_ctx *);
-struct request *blk_mq_alloc_request(struct request_queue *q, int rw, gfp_t gfp, bool reserved);
+struct request *blk_mq_alloc_request(struct request_queue *q, int rw, gfp_t gfp);
 struct request *blk_mq_alloc_reserved_request(struct request_queue *q, int rw, gfp_t gfp);
 struct request *blk_mq_rq_from_tag(struct request_queue *q, unsigned int tag);
 
@@ -133,6 +136,8 @@ void blk_mq_free_single_hw_queue(struct blk_mq_hw_ctx *, unsigned int);
 
 void blk_mq_end_io(struct request *rq, int error);
 
+void blk_mq_complete_request(struct request *rq);
+
 void blk_mq_stop_hw_queue(struct blk_mq_hw_ctx *hctx);
 void blk_mq_start_hw_queue(struct blk_mq_hw_ctx *hctx);
 void blk_mq_stop_hw_queues(struct request_queue *q);
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 8678c4322b44..4afa4f8f6090 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -98,7 +98,7 @@ struct request {
 	struct list_head queuelist;
 	union {
 		struct call_single_data csd;
-		struct work_struct mq_flush_data;
+		struct work_struct mq_flush_work;
 	};
 
 	struct request_queue *q;
@@ -448,13 +448,8 @@ struct request_queue {
 	unsigned long		flush_pending_since;
 	struct list_head	flush_queue[2];
 	struct list_head	flush_data_in_flight;
-	union {
-		struct request	flush_rq;
-		struct {
-			spinlock_t mq_flush_lock;
-			struct work_struct mq_flush_work;
-		};
-	};
+	struct request		*flush_rq;
+	spinlock_t		mq_flush_lock;
 
 	struct mutex		sysfs_lock;
 
diff --git a/include/xen/interface/io/blkif.h b/include/xen/interface/io/blkif.h
index ae665ac59c36..32ec05a6572f 100644
--- a/include/xen/interface/io/blkif.h
+++ b/include/xen/interface/io/blkif.h
@@ -113,13 +113,13 @@ typedef uint64_t blkif_sector_t;
  * it's less than the number provided by the backend. The indirect_grefs field
  * in blkif_request_indirect should be filled by the frontend with the
  * grant references of the pages that are holding the indirect segments.
- * This pages are filled with an array of blkif_request_segment_aligned
- * that hold the information about the segments. The number of indirect
- * pages to use is determined by the maximum number of segments
- * a indirect request contains. Every indirect page can contain a maximum
- * of 512 segments (PAGE_SIZE/sizeof(blkif_request_segment_aligned)),
- * so to calculate the number of indirect pages to use we have to do
- * ceil(indirect_segments/512).
+ * These pages are filled with an array of blkif_request_segment that hold the
+ * information about the segments. The number of indirect pages to use is
+ * determined by the number of segments an indirect request contains. Every
+ * indirect page can contain a maximum of
+ * (PAGE_SIZE / sizeof(struct blkif_request_segment)) segments, so to
+ * calculate the number of indirect pages to use we have to do
+ * ceil(indirect_segments / (PAGE_SIZE / sizeof(struct blkif_request_segment))).
  *
  * If a backend does not recognize BLKIF_OP_INDIRECT, it should *not*
  * create the "feature-max-indirect-segments" node!
@@ -135,13 +135,12 @@ typedef uint64_t blkif_sector_t;
 
 #define BLKIF_MAX_INDIRECT_PAGES_PER_REQUEST 8
 
-struct blkif_request_segment_aligned {
-	grant_ref_t gref;        /* reference to I/O buffer frame        */
-	/* @first_sect: first sector in frame to transfer (inclusive).   */
-	/* @last_sect: last sector in frame to transfer (inclusive).     */
-	uint8_t     first_sect, last_sect;
-	uint16_t    _pad; /* padding to make it 8 bytes, so it's cache-aligned */
-} __attribute__((__packed__));
+struct blkif_request_segment {
+		grant_ref_t gref;        /* reference to I/O buffer frame        */
+		/* @first_sect: first sector in frame to transfer (inclusive).   */
+		/* @last_sect: last sector in frame to transfer (inclusive).     */
+		uint8_t     first_sect, last_sect;
+};
 
 struct blkif_request_rw {
 	uint8_t        nr_segments;  /* number of segments                   */
@@ -151,12 +150,7 @@ struct blkif_request_rw {
 #endif
 	uint64_t       id;           /* private guest value, echoed in resp  */
 	blkif_sector_t sector_number;/* start sector idx on disk (r/w only)  */
-	struct blkif_request_segment {
-		grant_ref_t gref;        /* reference to I/O buffer frame        */
-		/* @first_sect: first sector in frame to transfer (inclusive).   */
-		/* @last_sect: last sector in frame to transfer (inclusive).     */
-		uint8_t     first_sect, last_sect;
-	} seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
+	struct blkif_request_segment seg[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 } __attribute__((__packed__));
 
 struct blkif_request_discard {
diff --git a/lib/percpu_ida.c b/lib/percpu_ida.c
index 7be235f1a70b..93d145e5539c 100644
--- a/lib/percpu_ida.c
+++ b/lib/percpu_ida.c
@@ -54,9 +54,7 @@ static inline void move_tags(unsigned *dst, unsigned *dst_nr,
 /*
  * Try to steal tags from a remote cpu's percpu freelist.
  *
- * We first check how many percpu freelists have tags - we don't steal tags
- * unless enough percpu freelists have tags on them that it's possible more than
- * half the total tags could be stuck on remote percpu freelists.
+ * We first check how many percpu freelists have tags
  *
  * Then we iterate through the cpus until we find some tags - we don't attempt
  * to find the "best" cpu to steal from, to keep cacheline bouncing to a
@@ -69,8 +67,7 @@ static inline void steal_tags(struct percpu_ida *pool,
 	struct percpu_ida_cpu *remote;
 
 	for (cpus_have_tags = cpumask_weight(&pool->cpus_have_tags);
-	     cpus_have_tags * pool->percpu_max_size > pool->nr_tags / 2;
-	     cpus_have_tags--) {
+	     cpus_have_tags; cpus_have_tags--) {
 		cpu = cpumask_next(cpu, &pool->cpus_have_tags);
 
 		if (cpu >= nr_cpu_ids) {