Merge branch 'for-3.16/core' of git://git.kernel.dk/linux-block into next

Pull block core updates from Jens Axboe:
 "It's a big(ish) round this time, lots of development effort has gone
  into blk-mq in the last 3 months.  Generally we're heading to where
  3.16 will be a feature complete and performant blk-mq.  scsi-mq is
  progressing nicely and will hopefully be in 3.17.  A nvme port is in
  progress, and the Micron pci-e flash driver, mtip32xx, is converted
  and will be sent in with the driver pull request for 3.16.

  This pull request contains:

   - Lots of prep and support patches for scsi-mq have been integrated.
     All from Christoph.

   - API and code cleanups for blk-mq from Christoph.

   - Lots of good corner case and error handling cleanup fixes for
     blk-mq from Ming Lei.

   - A flew of blk-mq updates from me:

     * Provide strict mappings so that the driver can rely on the CPU
       to queue mapping.  This enables optimizations in the driver.

     * Provided a bitmap tagging instead of percpu_ida, which never
       really worked well for blk-mq.  percpu_ida relies on the fact
       that we have a lot more tags available than we really need, it
       fails miserably for cases where we exhaust (or are close to
       exhausting) the tag space.

     * Provide sane support for shared tag maps, as utilized by scsi-mq

     * Various fixes for IO timeouts.

     * API cleanups, and lots of perf tweaks and optimizations.

   - Remove 'buffer' from struct request.  This is ancient code, from
     when requests were always virtually mapped.  Kill it, to reclaim
     some space in struct request.  From me.

   - Remove 'magic' from blk_plug.  Since we store these on the stack
     and since we've never caught any actual bugs with this, lets just
     get rid of it.  From me.

   - Only call part_in_flight() once for IO completion, as includes two
     atomic reads.  Hopefully we'll get a better implementation soon, as
     the part IO stats are now one of the more expensive parts of doing
     IO on blk-mq.  From me.

   - File migration of block code from {mm,fs}/ to block/.  This
     includes bio.c, bio-integrity.c, bounce.c, and ioprio.c.  From me,
     from a discussion on lkml.

  That should describe the meat of the pull request.  Also has various
  little fixes and cleanups from Dave Jones, Shaohua Li, Duan Jiong,
  Fengguang Wu, Fabian Frederick, Randy Dunlap, Robert Elliott, and Sam
  Bradshaw"

* 'for-3.16/core' of git://git.kernel.dk/linux-block: (100 commits)
  blk-mq: push IPI or local end_io decision to __blk_mq_complete_request()
  blk-mq: remember to start timeout handler for direct queue
  block: ensure that the timer is always added
  blk-mq: blk_mq_unregister_hctx() can be static
  blk-mq: make the sysfs mq/ layout reflect current mappings
  blk-mq: blk_mq_tag_to_rq should handle flush request
  block: remove dead code in scsi_ioctl:blk_verify_command
  blk-mq: request initialization optimizations
  block: add queue flag for disabling SG merging
  block: remove 'magic' from struct blk_plug
  blk-mq: remove alloc_hctx and free_hctx methods
  blk-mq: add file comments and update copyright notices
  blk-mq: remove blk_mq_alloc_request_pinned
  blk-mq: do not use blk_mq_alloc_request_pinned in blk_mq_map_request
  blk-mq: remove blk_mq_wait_for_tags
  blk-mq: initialize request in __blk_mq_alloc_request
  blk-mq: merge blk_mq_alloc_reserved_request into blk_mq_alloc_request
  blk-mq: add helper to insert requests from irq context
  blk-mq: remove stale comment for blk_mq_complete_request()
  blk-mq: allow non-softirq completions
  ...

16 files changed, 124 insertions, 149 deletions

diff --git a/drivers/block/amiflop.c b/drivers/block/amiflop.c
index 748dea4f34dc..758da2287d9a 100644
--- a/drivers/block/amiflop.c
+++ b/drivers/block/amiflop.c
@@ -1406,7 +1406,7 @@ next_segment:
 
 		track = block / (floppy->dtype->sects * floppy->type->sect_mult);
 		sector = block % (floppy->dtype->sects * floppy->type->sect_mult);
-		data = rq->buffer + 512 * cnt;
+		data = bio_data(rq->bio) + 512 * cnt;
 #ifdef DEBUG
 		printk("access to track %d, sector %d, with buffer at "
 		       "0x%08lx\n", track, sector, data);
diff --git a/drivers/block/ataflop.c b/drivers/block/ataflop.c
index cfa64bdf01c9..2104b1b4ccda 100644
--- a/drivers/block/ataflop.c
+++ b/drivers/block/ataflop.c
@@ -1484,7 +1484,7 @@ repeat:
 	ReqCnt = 0;
 	ReqCmd = rq_data_dir(fd_request);
 	ReqBlock = blk_rq_pos(fd_request);
-	ReqBuffer = fd_request->buffer;
+	ReqBuffer = bio_data(fd_request->bio);
 	setup_req_params( drive );
 	do_fd_action( drive );
 
diff --git a/drivers/block/floppy.c b/drivers/block/floppy.c
index fa9bb742df6e..dc3a41c82b38 100644
--- a/drivers/block/floppy.c
+++ b/drivers/block/floppy.c
@@ -2351,7 +2351,7 @@ static void rw_interrupt(void)
 	}
 
 	if (CT(COMMAND) != FD_READ ||
-	    raw_cmd->kernel_data == current_req->buffer) {
+	    raw_cmd->kernel_data == bio_data(current_req->bio)) {
 		/* transfer directly from buffer */
 		cont->done(1);
 	} else if (CT(COMMAND) == FD_READ) {
@@ -2640,7 +2640,7 @@ static int make_raw_rw_request(void)
 		raw_cmd->flags &= ~FD_RAW_WRITE;
 		raw_cmd->flags |= FD_RAW_READ;
 		COMMAND = FM_MODE(_floppy, FD_READ);
-	} else if ((unsigned long)current_req->buffer < MAX_DMA_ADDRESS) {
+	} else if ((unsigned long)bio_data(current_req->bio) < MAX_DMA_ADDRESS) {
 		unsigned long dma_limit;
 		int direct, indirect;
 
@@ -2654,13 +2654,13 @@ static int make_raw_rw_request(void)
 		 */
 		max_size = buffer_chain_size();
 		dma_limit = (MAX_DMA_ADDRESS -
-			     ((unsigned long)current_req->buffer)) >> 9;
+			     ((unsigned long)bio_data(current_req->bio))) >> 9;
 		if ((unsigned long)max_size > dma_limit)
 			max_size = dma_limit;
 		/* 64 kb boundaries */
-		if (CROSS_64KB(current_req->buffer, max_size << 9))
+		if (CROSS_64KB(bio_data(current_req->bio), max_size << 9))
 			max_size = (K_64 -
-				    ((unsigned long)current_req->buffer) %
+				    ((unsigned long)bio_data(current_req->bio)) %
 				    K_64) >> 9;
 		direct = transfer_size(ssize, max_sector, max_size) - fsector_t;
 		/*
@@ -2677,7 +2677,7 @@ static int make_raw_rw_request(void)
 		       (DP->read_track & (1 << DRS->probed_format)))))) {
 			max_size = blk_rq_sectors(current_req);
 		} else {
-			raw_cmd->kernel_data = current_req->buffer;
+			raw_cmd->kernel_data = bio_data(current_req->bio);
 			raw_cmd->length = current_count_sectors << 9;
 			if (raw_cmd->length == 0) {
 				DPRINT("%s: zero dma transfer attempted\n", __func__);
@@ -2731,7 +2731,7 @@ static int make_raw_rw_request(void)
 	raw_cmd->length = ((raw_cmd->length - 1) | (ssize - 1)) + 1;
 	raw_cmd->length <<= 9;
 	if ((raw_cmd->length < current_count_sectors << 9) ||
-	    (raw_cmd->kernel_data != current_req->buffer &&
+	    (raw_cmd->kernel_data != bio_data(current_req->bio) &&
 	     CT(COMMAND) == FD_WRITE &&
 	     (aligned_sector_t + (raw_cmd->length >> 9) > buffer_max ||
 	      aligned_sector_t < buffer_min)) ||
@@ -2739,7 +2739,7 @@ static int make_raw_rw_request(void)
 	    raw_cmd->length <= 0 || current_count_sectors <= 0) {
 		DPRINT("fractionary current count b=%lx s=%lx\n",
 		       raw_cmd->length, current_count_sectors);
-		if (raw_cmd->kernel_data != current_req->buffer)
+		if (raw_cmd->kernel_data != bio_data(current_req->bio))
 			pr_info("addr=%d, length=%ld\n",
 				(int)((raw_cmd->kernel_data -
 				       floppy_track_buffer) >> 9),
@@ -2756,7 +2756,7 @@ static int make_raw_rw_request(void)
 		return 0;
 	}
 
-	if (raw_cmd->kernel_data != current_req->buffer) {
+	if (raw_cmd->kernel_data != bio_data(current_req->bio)) {
 		if (raw_cmd->kernel_data < floppy_track_buffer ||
 		    current_count_sectors < 0 ||
 		    raw_cmd->length < 0 ||
diff --git a/drivers/block/hd.c b/drivers/block/hd.c
index bf397bf108b7..8a290c08262f 100644
--- a/drivers/block/hd.c
+++ b/drivers/block/hd.c
@@ -464,11 +464,11 @@ static void read_intr(void)
 
 ok_to_read:
 	req = hd_req;
-	insw(HD_DATA, req->buffer, 256);
+	insw(HD_DATA, bio_data(req->bio), 256);
 #ifdef DEBUG
 	printk("%s: read: sector %ld, remaining = %u, buffer=%p\n",
 	       req->rq_disk->disk_name, blk_rq_pos(req) + 1,
-	       blk_rq_sectors(req) - 1, req->buffer+512);
+	       blk_rq_sectors(req) - 1, bio_data(req->bio)+512);
 #endif
 	if (hd_end_request(0, 512)) {
 		SET_HANDLER(&read_intr);
@@ -505,7 +505,7 @@ static void write_intr(void)
 ok_to_write:
 	if (hd_end_request(0, 512)) {
 		SET_HANDLER(&write_intr);
-		outsw(HD_DATA, req->buffer, 256);
+		outsw(HD_DATA, bio_data(req->bio), 256);
 		return;
 	}
 
@@ -624,7 +624,7 @@ repeat:
 	printk("%s: %sing: CHS=%d/%d/%d, sectors=%d, buffer=%p\n",
 		req->rq_disk->disk_name,
 		req_data_dir(req) == READ ? "read" : "writ",
-		cyl, head, sec, nsect, req->buffer);
+		cyl, head, sec, nsect, bio_data(req->bio));
 #endif
 	if (req->cmd_type == REQ_TYPE_FS) {
 		switch (rq_data_dir(req)) {
@@ -643,7 +643,7 @@ repeat:
 				bad_rw_intr();
 				goto repeat;
 			}
-			outsw(HD_DATA, req->buffer, 256);
+			outsw(HD_DATA, bio_data(req->bio), 256);
 			break;
 		default:
 			printk("unknown hd-command\n");
diff --git a/drivers/block/mg_disk.c b/drivers/block/mg_disk.c
index eb59b1241366..e352cac707e8 100644
--- a/drivers/block/mg_disk.c
+++ b/drivers/block/mg_disk.c
@@ -479,7 +479,7 @@ static unsigned int mg_out(struct mg_host *host,
 
 static void mg_read_one(struct mg_host *host, struct request *req)
 {
-	u16 *buff = (u16 *)req->buffer;
+	u16 *buff = (u16 *)bio_data(req->bio);
 	u32 i;
 
 	for (i = 0; i < MG_SECTOR_SIZE >> 1; i++)
@@ -496,7 +496,7 @@ static void mg_read(struct request *req)
 		mg_bad_rw_intr(host);
 
 	MG_DBG("requested %d sects (from %ld), buffer=0x%p\n",
-	       blk_rq_sectors(req), blk_rq_pos(req), req->buffer);
+	       blk_rq_sectors(req), blk_rq_pos(req), bio_data(req->bio));
 
 	do {
 		if (mg_wait(host, ATA_DRQ,
@@ -514,7 +514,7 @@ static void mg_read(struct request *req)
 
 static void mg_write_one(struct mg_host *host, struct request *req)
 {
-	u16 *buff = (u16 *)req->buffer;
+	u16 *buff = (u16 *)bio_data(req->bio);
 	u32 i;
 
 	for (i = 0; i < MG_SECTOR_SIZE >> 1; i++)
@@ -534,7 +534,7 @@ static void mg_write(struct request *req)
 	}
 
 	MG_DBG("requested %d sects (from %ld), buffer=0x%p\n",
-	       rem, blk_rq_pos(req), req->buffer);
+	       rem, blk_rq_pos(req), bio_data(req->bio));
 
 	if (mg_wait(host, ATA_DRQ,
 		    MG_TMAX_WAIT_WR_DRQ) != MG_ERR_NONE) {
@@ -585,7 +585,7 @@ ok_to_read:
 	mg_read_one(host, req);
 
 	MG_DBG("sector %ld, remaining=%ld, buffer=0x%p\n",
-	       blk_rq_pos(req), blk_rq_sectors(req) - 1, req->buffer);
+	       blk_rq_pos(req), blk_rq_sectors(req) - 1, bio_data(req->bio));
 
 	/* send read confirm */
 	outb(MG_CMD_RD_CONF, (unsigned long)host->dev_base + MG_REG_COMMAND);
@@ -624,7 +624,7 @@ ok_to_write:
 		/* write 1 sector and set handler if remains */
 		mg_write_one(host, req);
 		MG_DBG("sector %ld, remaining=%ld, buffer=0x%p\n",
-		       blk_rq_pos(req), blk_rq_sectors(req), req->buffer);
+		       blk_rq_pos(req), blk_rq_sectors(req), bio_data(req->bio));
 		host->mg_do_intr = mg_write_intr;
 		mod_timer(&host->timer, jiffies + 3 * HZ);
 	}
diff --git a/drivers/block/null_blk.c b/drivers/block/null_blk.c
index 091b9ea14feb..b40af63a5476 100644
--- a/drivers/block/null_blk.c
+++ b/drivers/block/null_blk.c
@@ -32,6 +32,7 @@ struct nullb {
 	unsigned int index;
 	struct request_queue *q;
 	struct gendisk *disk;
+	struct blk_mq_tag_set tag_set;
 	struct hrtimer timer;
 	unsigned int queue_depth;
 	spinlock_t lock;
@@ -226,7 +227,7 @@ static void null_cmd_end_timer(struct nullb_cmd *cmd)
 
 static void null_softirq_done_fn(struct request *rq)
 {
-	end_cmd(rq->special);
+	end_cmd(blk_mq_rq_to_pdu(rq));
 }
 
 static inline void null_handle_cmd(struct nullb_cmd *cmd)
@@ -311,7 +312,7 @@ static void null_request_fn(struct request_queue *q)
 
 static int null_queue_rq(struct blk_mq_hw_ctx *hctx, struct request *rq)
 {
-	struct nullb_cmd *cmd = rq->special;
+	struct nullb_cmd *cmd = blk_mq_rq_to_pdu(rq);
 
 	cmd->rq = rq;
 	cmd->nq = hctx->driver_data;
@@ -320,46 +321,6 @@ static int null_queue_rq(struct blk_mq_hw_ctx *hctx, struct request *rq)
 	return BLK_MQ_RQ_QUEUE_OK;
 }
 
-static struct blk_mq_hw_ctx *null_alloc_hctx(struct blk_mq_reg *reg, unsigned int hctx_index)
-{
-	int b_size = DIV_ROUND_UP(reg->nr_hw_queues, nr_online_nodes);
-	int tip = (reg->nr_hw_queues % nr_online_nodes);
-	int node = 0, i, n;
-
-	/*
-	 * Split submit queues evenly wrt to the number of nodes. If uneven,
-	 * fill the first buckets with one extra, until the rest is filled with
-	 * no extra.
-	 */
-	for (i = 0, n = 1; i < hctx_index; i++, n++) {
-		if (n % b_size == 0) {
-			n = 0;
-			node++;
-
-			tip--;
-			if (!tip)
-				b_size = reg->nr_hw_queues / nr_online_nodes;
-		}
-	}
-
-	/*
-	 * A node might not be online, therefore map the relative node id to the
-	 * real node id.
-	 */
-	for_each_online_node(n) {
-		if (!node)
-			break;
-		node--;
-	}
-
-	return kzalloc_node(sizeof(struct blk_mq_hw_ctx), GFP_KERNEL, n);
-}
-
-static void null_free_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_index)
-{
-	kfree(hctx);
-}
-
 static void null_init_queue(struct nullb *nullb, struct nullb_queue *nq)
 {
 	BUG_ON(!nullb);
@@ -389,19 +350,14 @@ static struct blk_mq_ops null_mq_ops = {
 	.complete	= null_softirq_done_fn,
 };
 
-static struct blk_mq_reg null_mq_reg = {
-	.ops		= &null_mq_ops,
-	.queue_depth	= 64,
-	.cmd_size	= sizeof(struct nullb_cmd),
-	.flags		= BLK_MQ_F_SHOULD_MERGE,
-};
-
 static void null_del_dev(struct nullb *nullb)
 {
 	list_del_init(&nullb->list);
 
 	del_gendisk(nullb->disk);
 	blk_cleanup_queue(nullb->q);
+	if (queue_mode == NULL_Q_MQ)
+		blk_mq_free_tag_set(&nullb->tag_set);
 	put_disk(nullb->disk);
 	kfree(nullb);
 }
@@ -506,7 +462,7 @@ static int null_add_dev(void)
 
 	nullb = kzalloc_node(sizeof(*nullb), GFP_KERNEL, home_node);
 	if (!nullb)
-		return -ENOMEM;
+		goto out;
 
 	spin_lock_init(&nullb->lock);
 
@@ -514,49 +470,44 @@ static int null_add_dev(void)
 		submit_queues = nr_online_nodes;
 
 	if (setup_queues(nullb))
-		goto err;
+		goto out_free_nullb;
 
 	if (queue_mode == NULL_Q_MQ) {
-		null_mq_reg.numa_node = home_node;
-		null_mq_reg.queue_depth = hw_queue_depth;
-		null_mq_reg.nr_hw_queues = submit_queues;
-
-		if (use_per_node_hctx) {
-			null_mq_reg.ops->alloc_hctx = null_alloc_hctx;
-			null_mq_reg.ops->free_hctx = null_free_hctx;
-		} else {
-			null_mq_reg.ops->alloc_hctx = blk_mq_alloc_single_hw_queue;
-			null_mq_reg.ops->free_hctx = blk_mq_free_single_hw_queue;
-		}
-
-		nullb->q = blk_mq_init_queue(&null_mq_reg, nullb);
+		nullb->tag_set.ops = &null_mq_ops;
+		nullb->tag_set.nr_hw_queues = submit_queues;
+		nullb->tag_set.queue_depth = hw_queue_depth;
+		nullb->tag_set.numa_node = home_node;
+		nullb->tag_set.cmd_size	= sizeof(struct nullb_cmd);
+		nullb->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
+		nullb->tag_set.driver_data = nullb;
+
+		if (blk_mq_alloc_tag_set(&nullb->tag_set))
+			goto out_cleanup_queues;
+
+		nullb->q = blk_mq_init_queue(&nullb->tag_set);
+		if (!nullb->q)
+			goto out_cleanup_tags;
 	} else if (queue_mode == NULL_Q_BIO) {
 		nullb->q = blk_alloc_queue_node(GFP_KERNEL, home_node);
+		if (!nullb->q)
+			goto out_cleanup_queues;
 		blk_queue_make_request(nullb->q, null_queue_bio);
 		init_driver_queues(nullb);
 	} else {
 		nullb->q = blk_init_queue_node(null_request_fn, &nullb->lock, home_node);
+		if (!nullb->q)
+			goto out_cleanup_queues;
 		blk_queue_prep_rq(nullb->q, null_rq_prep_fn);
-		if (nullb->q)
-			blk_queue_softirq_done(nullb->q, null_softirq_done_fn);
+		blk_queue_softirq_done(nullb->q, null_softirq_done_fn);
 		init_driver_queues(nullb);
 	}
 
-	if (!nullb->q)
-		goto queue_fail;
-
 	nullb->q->queuedata = nullb;
 	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, nullb->q);
 
 	disk = nullb->disk = alloc_disk_node(1, home_node);
-	if (!disk) {
-queue_fail:
-		blk_cleanup_queue(nullb->q);
-		cleanup_queues(nullb);
-err:
-		kfree(nullb);
-		return -ENOMEM;
-	}
+	if (!disk)
+		goto out_cleanup_blk_queue;
 
 	mutex_lock(&lock);
 	list_add_tail(&nullb->list, &nullb_list);
@@ -579,6 +530,18 @@ err:
 	sprintf(disk->disk_name, "nullb%d", nullb->index);
 	add_disk(disk);
 	return 0;
+
+out_cleanup_blk_queue:
+	blk_cleanup_queue(nullb->q);
+out_cleanup_tags:
+	if (queue_mode == NULL_Q_MQ)
+		blk_mq_free_tag_set(&nullb->tag_set);
+out_cleanup_queues:
+	cleanup_queues(nullb);
+out_free_nullb:
+	kfree(nullb);
+out:
+	return -ENOMEM;
 }
 
 static int __init null_init(void)
diff --git a/drivers/block/paride/pcd.c b/drivers/block/paride/pcd.c
index e76bdc074dbe..719cb1bc1640 100644
--- a/drivers/block/paride/pcd.c
+++ b/drivers/block/paride/pcd.c
@@ -747,7 +747,7 @@ static void do_pcd_request(struct request_queue * q)
 			pcd_current = cd;
 			pcd_sector = blk_rq_pos(pcd_req);
 			pcd_count = blk_rq_cur_sectors(pcd_req);
-			pcd_buf = pcd_req->buffer;
+			pcd_buf = bio_data(pcd_req->bio);
 			pcd_busy = 1;
 			ps_set_intr(do_pcd_read, NULL, 0, nice);
 			return;
diff --git a/drivers/block/paride/pd.c b/drivers/block/paride/pd.c
index 19ad8f0c83ef..fea7e76a00de 100644
--- a/drivers/block/paride/pd.c
+++ b/drivers/block/paride/pd.c
@@ -454,7 +454,7 @@ static enum action do_pd_io_start(void)
 		if (pd_block + pd_count > get_capacity(pd_req->rq_disk))
 			return Fail;
 		pd_run = blk_rq_sectors(pd_req);
-		pd_buf = pd_req->buffer;
+		pd_buf = bio_data(pd_req->bio);
 		pd_retries = 0;
 		if (pd_cmd == READ)
 			return do_pd_read_start();
@@ -485,7 +485,7 @@ static int pd_next_buf(void)
 	spin_lock_irqsave(&pd_lock, saved_flags);
 	__blk_end_request_cur(pd_req, 0);
 	pd_count = blk_rq_cur_sectors(pd_req);
-	pd_buf = pd_req->buffer;
+	pd_buf = bio_data(pd_req->bio);
 	spin_unlock_irqrestore(&pd_lock, saved_flags);
 	return 0;
 }
diff --git a/drivers/block/paride/pf.c b/drivers/block/paride/pf.c
index f5c86d523ba0..9a15fd3c9349 100644
--- a/drivers/block/paride/pf.c
+++ b/drivers/block/paride/pf.c
@@ -795,7 +795,7 @@ repeat:
 	}
 
 	pf_cmd = rq_data_dir(pf_req);
-	pf_buf = pf_req->buffer;
+	pf_buf = bio_data(pf_req->bio);
 	pf_retries = 0;
 
 	pf_busy = 1;
@@ -827,7 +827,7 @@ static int pf_next_buf(void)
 		if (!pf_req)
 			return 1;
 		pf_count = blk_rq_cur_sectors(pf_req);
-		pf_buf = pf_req->buffer;
+		pf_buf = bio_data(pf_req->bio);
 	}
 	return 0;
 }
diff --git a/drivers/block/skd_main.c b/drivers/block/skd_main.c
index a69dd93d1bd5..c48d9084c965 100644
--- a/drivers/block/skd_main.c
+++ b/drivers/block/skd_main.c
@@ -563,7 +563,6 @@ skd_prep_discard_cdb(struct skd_scsi_request *scsi_req,
 
 	req = skreq->req;
 	blk_add_request_payload(req, page, len);
-	req->buffer = buf;
 }
 
 static void skd_request_fn_not_online(struct request_queue *q);
@@ -744,6 +743,7 @@ static void skd_request_fn(struct request_queue *q)
 				break;
 			}
 			skreq->discard_page = 1;
+			req->completion_data = page;
 			skd_prep_discard_cdb(scsi_req, skreq, page, lba, count);
 
 		} else if (flush == SKD_FLUSH_ZERO_SIZE_FIRST) {
@@ -858,8 +858,7 @@ static void skd_end_request(struct skd_device *skdev,
 		(skreq->discard_page == 1)) {
 		pr_debug("%s:%s:%d, free the page!",
 			 skdev->name, __func__, __LINE__);
-		free_page((unsigned long)req->buffer);
-		req->buffer = NULL;
+		__free_page(req->completion_data);
 	}
 
 	if (unlikely(error)) {
diff --git a/drivers/block/swim.c b/drivers/block/swim.c
index b02d53a399f3..6b44bbe528b7 100644
--- a/drivers/block/swim.c
+++ b/drivers/block/swim.c
@@ -549,7 +549,7 @@ static void redo_fd_request(struct request_queue *q)
 		case READ:
 			err = floppy_read_sectors(fs, blk_rq_pos(req),
 						  blk_rq_cur_sectors(req),
-						  req->buffer);
+						  bio_data(req->bio));
 			break;
 		}
 	done:
diff --git a/drivers/block/swim3.c b/drivers/block/swim3.c
index c74f7b56e7c4..523ee8fd4c15 100644
--- a/drivers/block/swim3.c
+++ b/drivers/block/swim3.c
@@ -342,7 +342,7 @@ static void start_request(struct floppy_state *fs)
 		swim3_dbg("do_fd_req: dev=%s cmd=%d sec=%ld nr_sec=%u buf=%p\n",
 			  req->rq_disk->disk_name, req->cmd,
 			  (long)blk_rq_pos(req), blk_rq_sectors(req),
-			  req->buffer);
+			  bio_data(req->bio));
 		swim3_dbg("           errors=%d current_nr_sectors=%u\n",
 			  req->errors, blk_rq_cur_sectors(req));
 #endif
@@ -479,11 +479,11 @@ static inline void setup_transfer(struct floppy_state *fs)
 		/* Set up 3 dma commands: write preamble, data, postamble */
 		init_dma(cp, OUTPUT_MORE, write_preamble, sizeof(write_preamble));
 		++cp;
-		init_dma(cp, OUTPUT_MORE, req->buffer, 512);
+		init_dma(cp, OUTPUT_MORE, bio_data(req->bio), 512);
 		++cp;
 		init_dma(cp, OUTPUT_LAST, write_postamble, sizeof(write_postamble));
 	} else {
-		init_dma(cp, INPUT_LAST, req->buffer, n * 512);
+		init_dma(cp, INPUT_LAST, bio_data(req->bio), n * 512);
 	}
 	++cp;
 	out_le16(&cp->command, DBDMA_STOP);
diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c
index cb9b1f8326c3..c8f286e8d80f 100644
--- a/drivers/block/virtio_blk.c
+++ b/drivers/block/virtio_blk.c
@@ -30,6 +30,9 @@ struct virtio_blk
 	/* The disk structure for the kernel. */
 	struct gendisk *disk;
 
+	/* Block layer tags. */
+	struct blk_mq_tag_set tag_set;
+
 	/* Process context for config space updates */
 	struct work_struct config_work;
 
@@ -112,7 +115,7 @@ static int __virtblk_add_req(struct virtqueue *vq,
 
 static inline void virtblk_request_done(struct request *req)
 {
-	struct virtblk_req *vbr = req->special;
+	struct virtblk_req *vbr = blk_mq_rq_to_pdu(req);
 	int error = virtblk_result(vbr);
 
 	if (req->cmd_type == REQ_TYPE_BLOCK_PC) {
@@ -147,14 +150,14 @@ static void virtblk_done(struct virtqueue *vq)
 
 	/* In case queue is stopped waiting for more buffers. */
 	if (req_done)
-		blk_mq_start_stopped_hw_queues(vblk->disk->queue);
+		blk_mq_start_stopped_hw_queues(vblk->disk->queue, true);
 	spin_unlock_irqrestore(&vblk->vq_lock, flags);
 }
 
 static int virtio_queue_rq(struct blk_mq_hw_ctx *hctx, struct request *req)
 {
 	struct virtio_blk *vblk = hctx->queue->queuedata;
-	struct virtblk_req *vbr = req->special;
+	struct virtblk_req *vbr = blk_mq_rq_to_pdu(req);
 	unsigned long flags;
 	unsigned int num;
 	const bool last = (req->cmd_flags & REQ_END) != 0;
@@ -480,33 +483,27 @@ static const struct device_attribute dev_attr_cache_type_rw =
 	__ATTR(cache_type, S_IRUGO|S_IWUSR,
 	       virtblk_cache_type_show, virtblk_cache_type_store);
 
-static struct blk_mq_ops virtio_mq_ops = {
-	.queue_rq	= virtio_queue_rq,
-	.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 = {
-	.ops		= &virtio_mq_ops,
-	.nr_hw_queues	= 1,
-	.queue_depth	= 0, /* Set in virtblk_probe */
-	.numa_node	= NUMA_NO_NODE,
-	.flags		= BLK_MQ_F_SHOULD_MERGE,
-};
-module_param_named(queue_depth, virtio_mq_reg.queue_depth, uint, 0444);
-
-static int virtblk_init_vbr(void *data, struct blk_mq_hw_ctx *hctx,
-			     struct request *rq, unsigned int nr)
+static int virtblk_init_request(void *data, struct request *rq,
+		unsigned int hctx_idx, unsigned int request_idx,
+		unsigned int numa_node)
 {
 	struct virtio_blk *vblk = data;
-	struct virtblk_req *vbr = rq->special;
+	struct virtblk_req *vbr = blk_mq_rq_to_pdu(rq);
 
 	sg_init_table(vbr->sg, vblk->sg_elems);
 	return 0;
 }
 
+static struct blk_mq_ops virtio_mq_ops = {
+	.queue_rq	= virtio_queue_rq,
+	.map_queue	= blk_mq_map_queue,
+	.complete	= virtblk_request_done,
+	.init_request	= virtblk_init_request,
+};
+
+static unsigned int virtblk_queue_depth;
+module_param_named(queue_depth, virtblk_queue_depth, uint, 0444);
+
 static int virtblk_probe(struct virtio_device *vdev)
 {
 	struct virtio_blk *vblk;
@@ -561,24 +558,34 @@ static int virtblk_probe(struct virtio_device *vdev)
 	}
 
 	/* Default queue sizing is to fill the ring. */
-	if (!virtio_mq_reg.queue_depth) {
-		virtio_mq_reg.queue_depth = vblk->vq->num_free;
+	if (!virtblk_queue_depth) {
+		virtblk_queue_depth = vblk->vq->num_free;
 		/* ... but without indirect descs, we use 2 descs per req */
 		if (!virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC))
-			virtio_mq_reg.queue_depth /= 2;
+			virtblk_queue_depth /= 2;
 	}
-	virtio_mq_reg.cmd_size =
+
+	memset(&vblk->tag_set, 0, sizeof(vblk->tag_set));
+	vblk->tag_set.ops = &virtio_mq_ops;
+	vblk->tag_set.nr_hw_queues = 1;
+	vblk->tag_set.queue_depth = virtblk_queue_depth;
+	vblk->tag_set.numa_node = NUMA_NO_NODE;
+	vblk->tag_set.flags = BLK_MQ_F_SHOULD_MERGE;
+	vblk->tag_set.cmd_size =
 		sizeof(struct virtblk_req) +
 		sizeof(struct scatterlist) * sg_elems;
+	vblk->tag_set.driver_data = vblk;
 
-	q = vblk->disk->queue = blk_mq_init_queue(&virtio_mq_reg, vblk);
+	err = blk_mq_alloc_tag_set(&vblk->tag_set);
+	if (err)
+		goto out_put_disk;
+
+	q = vblk->disk->queue = blk_mq_init_queue(&vblk->tag_set);
 	if (!q) {
 		err = -ENOMEM;
-		goto out_put_disk;
+		goto out_free_tags;
 	}
 
-	blk_mq_init_commands(q, virtblk_init_vbr, vblk);
-
 	q->queuedata = vblk;
 
 	virtblk_name_format("vd", index, vblk->disk->disk_name, DISK_NAME_LEN);
@@ -679,6 +686,8 @@ static int virtblk_probe(struct virtio_device *vdev)
 out_del_disk:
 	del_gendisk(vblk->disk);
 	blk_cleanup_queue(vblk->disk->queue);
+out_free_tags:
+	blk_mq_free_tag_set(&vblk->tag_set);
 out_put_disk:
 	put_disk(vblk->disk);
 out_free_vq:
@@ -705,6 +714,8 @@ static void virtblk_remove(struct virtio_device *vdev)
 	del_gendisk(vblk->disk);
 	blk_cleanup_queue(vblk->disk->queue);
 
+	blk_mq_free_tag_set(&vblk->tag_set);
+
 	/* Stop all the virtqueues. */
 	vdev->config->reset(vdev);
 
@@ -749,7 +760,7 @@ static int virtblk_restore(struct virtio_device *vdev)
 	vblk->config_enable = true;
 	ret = init_vq(vdev->priv);
 	if (!ret)
-		blk_mq_start_stopped_hw_queues(vblk->disk->queue);
+		blk_mq_start_stopped_hw_queues(vblk->disk->queue, true);
 
 	return ret;
 }
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index efe1b4761735..283a30e88287 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -612,10 +612,10 @@ static void do_blkif_request(struct request_queue *rq)
 		}
 
 		pr_debug("do_blk_req %p: cmd %p, sec %lx, "
-			 "(%u/%u) buffer:%p [%s]\n",
+			 "(%u/%u) [%s]\n",
 			 req, req->cmd, (unsigned long)blk_rq_pos(req),
 			 blk_rq_cur_sectors(req), blk_rq_sectors(req),
-			 req->buffer, rq_data_dir(req) ? "write" : "read");
+			 rq_data_dir(req) ? "write" : "read");
 
 		if (blkif_queue_request(req)) {
 			blk_requeue_request(rq, req);
diff --git a/drivers/block/xsysace.c b/drivers/block/xsysace.c
index 1393b8871a28..ab3ea62e5dfc 100644
--- a/drivers/block/xsysace.c
+++ b/drivers/block/xsysace.c
@@ -661,7 +661,7 @@ static void ace_fsm_dostate(struct ace_device *ace)
 			rq_data_dir(req));
 
 		ace->req = req;
-		ace->data_ptr = req->buffer;
+		ace->data_ptr = bio_data(req->bio);
 		ace->data_count = blk_rq_cur_sectors(req) * ACE_BUF_PER_SECTOR;
 		ace_out32(ace, ACE_MPULBA, blk_rq_pos(req) & 0x0FFFFFFF);
 
@@ -733,7 +733,7 @@ static void ace_fsm_dostate(struct ace_device *ace)
 			 *      blk_rq_sectors(ace->req),
 			 *      blk_rq_cur_sectors(ace->req));
 			 */
-			ace->data_ptr = ace->req->buffer;
+			ace->data_ptr = bio_data(ace->req->bio);
 			ace->data_count = blk_rq_cur_sectors(ace->req) * 16;
 			ace_fsm_yieldirq(ace);
 			break;
diff --git a/drivers/block/z2ram.c b/drivers/block/z2ram.c
index 27de5046708a..968f9e52effa 100644
--- a/drivers/block/z2ram.c
+++ b/drivers/block/z2ram.c
@@ -87,13 +87,15 @@ static void do_z2_request(struct request_queue *q)
 		while (len) {
 			unsigned long addr = start & Z2RAM_CHUNKMASK;
 			unsigned long size = Z2RAM_CHUNKSIZE - addr;
+			void *buffer = bio_data(req->bio);
+
 			if (len < size)
 				size = len;
 			addr += z2ram_map[ start >> Z2RAM_CHUNKSHIFT ];
 			if (rq_data_dir(req) == READ)
-				memcpy(req->buffer, (char *)addr, size);
+				memcpy(buffer, (char *)addr, size);
 			else
-				memcpy((char *)addr, req->buffer, size);
+				memcpy((char *)addr, buffer, size);
 			start += size;
 			len -= size;
 		}