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

Pull block fixes from Jens Axboe:
 "A selection of fixes/changes that should make it into this series.
  This contains:

   - NVMe, two merges, containing:
        - pci-e, rdma, and fc fixes
        - Device quirks

   - Fix for a badblocks leak in null_blk

   - bcache fix from Rui Hua for a race condition regression where
     -EINTR was returned to upper layers that didn't expect it.

   - Regression fix for blktrace for a bug introduced in this series.

   - blktrace cleanup for cgroup id.

   - bdi registration error handling.

   - Small series with cleanups for blk-wbt.

   - Various little fixes for typos and the like.

  Nothing earth shattering, most important are the NVMe and bcache fixes"

* 'for-linus' of git://git.kernel.dk/linux-block: (34 commits)
  nvme-pci: fix NULL pointer dereference in nvme_free_host_mem()
  nvme-rdma: fix memory leak during queue allocation
  blktrace: fix trace mutex deadlock
  nvme-rdma: Use mr pool
  nvme-rdma: Check remotely invalidated rkey matches our expected rkey
  nvme-rdma: wait for local invalidation before completing a request
  nvme-rdma: don't complete requests before a send work request has completed
  nvme-rdma: don't suppress send completions
  bcache: check return value of register_shrinker
  bcache: recover data from backing when data is clean
  bcache: Fix building error on MIPS
  bcache: add a comment in journal bucket reading
  nvme-fc: don't use bit masks for set/test_bit() numbers
  blk-wbt: fix comments typo
  blk-wbt: move wbt_clear_stat to common place in wbt_done
  blk-sysfs: remove NULL pointer checking in queue_wb_lat_store
  blk-wbt: remove duplicated setting in wbt_init
  nvme-pci: add quirk for delay before CHK RDY for WDC SN200
  block: remove useless assignment in bio_split
  null_blk: fix dev->badblocks leak
  ...

block/bio.c

@@ -1819,7 +1819,7 @@ EXPORT_SYMBOL(bio_endio);
 struct bio *bio_split(struct bio *bio, int sectors,
 		      gfp_t gfp, struct bio_set *bs)
 {
-	struct bio *split = NULL;
+	struct bio *split;
 
 	BUG_ON(sectors <= 0);
 	BUG_ON(sectors >= bio_sectors(bio));

block/blk-sysfs.c

@@ -450,12 +450,9 @@ static ssize_t queue_wb_lat_store(struct request_queue *q, const char *page,
 		ret = wbt_init(q);
 		if (ret)
 			return ret;
-
-		rwb = q->rq_wb;
-		if (!rwb)
-			return -EINVAL;
 	}
 
+	rwb = q->rq_wb;
 	if (val == -1)
 		rwb->min_lat_nsec = wbt_default_latency_nsec(q);
 	else if (val >= 0)

block/blk-wbt.c

@@ -178,12 +178,11 @@ void wbt_done(struct rq_wb *rwb, struct blk_issue_stat *stat)
 
 		if (wbt_is_read(stat))
 			wb_timestamp(rwb, &rwb->last_comp);
-		wbt_clear_state(stat);
 	} else {
 		WARN_ON_ONCE(stat == rwb->sync_cookie);
 		__wbt_done(rwb, wbt_stat_to_mask(stat));
-		wbt_clear_state(stat);
 	}
+	wbt_clear_state(stat);
 }
 
 /*
@@ -482,7 +481,7 @@ static inline unsigned int get_limit(struct rq_wb *rwb, unsigned long rw)
 
 	/*
 	 * At this point we know it's a buffered write. If this is
-	 * kswapd trying to free memory, or REQ_SYNC is set, set, then
+	 * kswapd trying to free memory, or REQ_SYNC is set, then
 	 * it's WB_SYNC_ALL writeback, and we'll use the max limit for
 	 * that. If the write is marked as a background write, then use
 	 * the idle limit, or go to normal if we haven't had competing
@@ -723,8 +722,6 @@ int wbt_init(struct request_queue *q)
 		init_waitqueue_head(&rwb->rq_wait[i].wait);
 	}
 
-	rwb->wc = 1;
-	rwb->queue_depth = RWB_DEF_DEPTH;
 	rwb->last_comp = rwb->last_issue = jiffies;
 	rwb->queue = q;
 	rwb->win_nsec = RWB_WINDOW_NSEC;

block/genhd.c

@@ -671,10 +671,13 @@ void device_add_disk(struct device *parent, struct gendisk *disk)
 		disk->flags |= GENHD_FL_SUPPRESS_PARTITION_INFO;
 		disk->flags |= GENHD_FL_NO_PART_SCAN;
 	} else {
+		int ret;
+
 		/* Register BDI before referencing it from bdev */
 		disk_to_dev(disk)->devt = devt;
-		bdi_register_owner(disk->queue->backing_dev_info,
-				disk_to_dev(disk));
+		ret = bdi_register_owner(disk->queue->backing_dev_info,
+						disk_to_dev(disk));
+		WARN_ON(ret);
 		blk_register_region(disk_devt(disk), disk->minors, NULL,
 				    exact_match, exact_lock, disk);
 	}
@@ -1389,7 +1392,7 @@ struct gendisk *__alloc_disk_node(int minors, int node_id)
 
 	if (minors > DISK_MAX_PARTS) {
 		printk(KERN_ERR
-			"block: can't allocated more than %d partitions\n",
+			"block: can't allocate more than %d partitions\n",
 			DISK_MAX_PARTS);
 		minors = DISK_MAX_PARTS;
 	}

drivers/block/null_blk.c

@@ -471,7 +471,6 @@ static void nullb_device_release(struct config_item *item)
 {
 	struct nullb_device *dev = to_nullb_device(item);
 
-	badblocks_exit(&dev->badblocks);
 	null_free_device_storage(dev, false);
 	null_free_dev(dev);
 }
@@ -582,6 +581,10 @@ static struct nullb_device *null_alloc_dev(void)
 
 static void null_free_dev(struct nullb_device *dev)
 {
+	if (!dev)
+		return;
+
+	badblocks_exit(&dev->badblocks);
 	kfree(dev);
 }
 

drivers/md/bcache/alloc.c

@@ -490,7 +490,7 @@ int __bch_bucket_alloc_set(struct cache_set *c, unsigned reserve,
 		if (b == -1)
 			goto err;
 
-		k->ptr[i] = PTR(ca->buckets[b].gen,
+		k->ptr[i] = MAKE_PTR(ca->buckets[b].gen,
 				bucket_to_sector(c, b),
 				ca->sb.nr_this_dev);
 

drivers/md/bcache/btree.c

@@ -807,7 +807,10 @@ int bch_btree_cache_alloc(struct cache_set *c)
 	c->shrink.scan_objects = bch_mca_scan;
 	c->shrink.seeks = 4;
 	c->shrink.batch = c->btree_pages * 2;
-	register_shrinker(&c->shrink);
+
+	if (register_shrinker(&c->shrink))
+		pr_warn("bcache: %s: could not register shrinker",
+				__func__);
 
 	return 0;
 }

drivers/md/bcache/extents.c

@@ -585,7 +585,7 @@ static bool bch_extent_merge(struct btree_keys *bk, struct bkey *l, struct bkey
 		return false;
 
 	for (i = 0; i < KEY_PTRS(l); i++)
-		if (l->ptr[i] + PTR(0, KEY_SIZE(l), 0) != r->ptr[i] ||
+		if (l->ptr[i] + MAKE_PTR(0, KEY_SIZE(l), 0) != r->ptr[i] ||
 		    PTR_BUCKET_NR(b->c, l, i) != PTR_BUCKET_NR(b->c, r, i))
 			return false;
 

drivers/md/bcache/journal.c

@@ -170,6 +170,11 @@ int bch_journal_read(struct cache_set *c, struct list_head *list)
 		 * find a sequence of buckets with valid journal entries
 		 */
 		for (i = 0; i < ca->sb.njournal_buckets; i++) {
+			/*
+			 * We must try the index l with ZERO first for
+			 * correctness due to the scenario that the journal
+			 * bucket is circular buffer which might have wrapped
+			 */
 			l = (i * 2654435769U) % ca->sb.njournal_buckets;
 
 			if (test_bit(l, bitmap))
@@ -507,7 +512,7 @@ static void journal_reclaim(struct cache_set *c)
 			continue;
 
 		ja->cur_idx = next;
-		k->ptr[n++] = PTR(0,
+		k->ptr[n++] = MAKE_PTR(0,
 				  bucket_to_sector(c, ca->sb.d[ja->cur_idx]),
 				  ca->sb.nr_this_dev);
 	}

drivers/md/bcache/request.c

@@ -708,16 +708,15 @@ static void cached_dev_read_error(struct closure *cl)
 {
 	struct search *s = container_of(cl, struct search, cl);
 	struct bio *bio = &s->bio.bio;
-	struct cached_dev *dc = container_of(s->d, struct cached_dev, disk);
 
 	/*
-	 * If cache device is dirty (dc->has_dirty is non-zero), then
-	 * recovery a failed read request from cached device may get a
-	 * stale data back. So read failure recovery is only permitted
-	 * when cache device is clean.
+	 * If read request hit dirty data (s->read_dirty_data is true),
+	 * then recovery a failed read request from cached device may
+	 * get a stale data back. So read failure recovery is only
+	 * permitted when read request hit clean data in cache device,
+	 * or when cache read race happened.
 	 */
-	if (s->recoverable &&
-	    (dc && !atomic_read(&dc->has_dirty))) {
+	if (s->recoverable && !s->read_dirty_data) {
 		/* Retry from the backing device: */
 		trace_bcache_read_retry(s->orig_bio);
 

drivers/nvme/host/core.c

@@ -1449,19 +1449,19 @@ static int nvme_pr_command(struct block_device *bdev, u32 cdw10,
 	int srcu_idx, ret;
 	u8 data[16] = { 0, };
 
+	ns = nvme_get_ns_from_disk(bdev->bd_disk, &head, &srcu_idx);
+	if (unlikely(!ns))
+		return -EWOULDBLOCK;
+
 	put_unaligned_le64(key, &data[0]);
 	put_unaligned_le64(sa_key, &data[8]);
 
 	memset(&c, 0, sizeof(c));
 	c.common.opcode = op;
-	c.common.nsid = cpu_to_le32(head->ns_id);
+	c.common.nsid = cpu_to_le32(ns->head->ns_id);
 	c.common.cdw10[0] = cpu_to_le32(cdw10);
 
-	ns = nvme_get_ns_from_disk(bdev->bd_disk, &head, &srcu_idx);
-	if (unlikely(!ns))
-		ret = -EWOULDBLOCK;
-	else
-		ret = nvme_submit_sync_cmd(ns->queue, &c, data, 16);
+	ret = nvme_submit_sync_cmd(ns->queue, &c, data, 16);
 	nvme_put_ns_from_disk(head, srcu_idx);
 	return ret;
 }
@@ -2961,8 +2961,6 @@ static void nvme_alloc_ns(struct nvme_ctrl *ctrl, unsigned nsid)
 
 static void nvme_ns_remove(struct nvme_ns *ns)
 {
-	struct nvme_ns_head *head = ns->head;
-
 	if (test_and_set_bit(NVME_NS_REMOVING, &ns->flags))
 		return;
 
@@ -2980,15 +2978,14 @@ static void nvme_ns_remove(struct nvme_ns *ns)
 
 	mutex_lock(&ns->ctrl->subsys->lock);
 	nvme_mpath_clear_current_path(ns);
-	if (head)
-		list_del_rcu(&ns->siblings);
+	list_del_rcu(&ns->siblings);
 	mutex_unlock(&ns->ctrl->subsys->lock);
 
 	mutex_lock(&ns->ctrl->namespaces_mutex);
 	list_del_init(&ns->list);
 	mutex_unlock(&ns->ctrl->namespaces_mutex);
 
-	synchronize_srcu(&head->srcu);
+	synchronize_srcu(&ns->head->srcu);
 	nvme_put_ns(ns);
 }
 

drivers/nvme/host/fabrics.h

@@ -156,4 +156,34 @@ void nvmf_free_options(struct nvmf_ctrl_options *opts);
 int nvmf_get_address(struct nvme_ctrl *ctrl, char *buf, int size);
 bool nvmf_should_reconnect(struct nvme_ctrl *ctrl);
 
+static inline blk_status_t nvmf_check_init_req(struct nvme_ctrl *ctrl,
+		struct request *rq)
+{
+	struct nvme_command *cmd = nvme_req(rq)->cmd;
+
+	/*
+	 * We cannot accept any other command until the connect command has
+	 * completed, so only allow connect to pass.
+	 */
+	if (!blk_rq_is_passthrough(rq) ||
+	    cmd->common.opcode != nvme_fabrics_command ||
+	    cmd->fabrics.fctype != nvme_fabrics_type_connect) {
+		/*
+		 * Reconnecting state means transport disruption, which can take
+		 * a long time and even might fail permanently, fail fast to
+		 * give upper layers a chance to failover.
+		 * Deleting state means that the ctrl will never accept commands
+		 * again, fail it permanently.
+		 */
+		if (ctrl->state == NVME_CTRL_RECONNECTING ||
+		    ctrl->state == NVME_CTRL_DELETING) {
+			nvme_req(rq)->status = NVME_SC_ABORT_REQ;
+			return BLK_STS_IOERR;
+		}
+		return BLK_STS_RESOURCE; /* try again later */
+	}
+
+	return BLK_STS_OK;
+}
+
 #endif /* _NVME_FABRICS_H */

drivers/nvme/host/fc.c

@@ -31,7 +31,8 @@
 
 
 enum nvme_fc_queue_flags {
-	NVME_FC_Q_CONNECTED = (1 << 0),
+	NVME_FC_Q_CONNECTED = 0,
+	NVME_FC_Q_LIVE,
 };
 
 #define NVMEFC_QUEUE_DELAY	3		/* ms units */
@@ -1927,14 +1928,14 @@ nvme_fc_free_queue(struct nvme_fc_queue *queue)
 	if (!test_and_clear_bit(NVME_FC_Q_CONNECTED, &queue->flags))
 		return;
 
+	clear_bit(NVME_FC_Q_LIVE, &queue->flags);
 	/*
 	 * Current implementation never disconnects a single queue.
 	 * It always terminates a whole association. So there is never
 	 * a disconnect(queue) LS sent to the target.
 	 */
 
 	queue->connection_id = 0;
-	clear_bit(NVME_FC_Q_CONNECTED, &queue->flags);
 }
 
 static void
@@ -2013,6 +2014,8 @@ nvme_fc_connect_io_queues(struct nvme_fc_ctrl *ctrl, u16 qsize)
 		ret = nvmf_connect_io_queue(&ctrl->ctrl, i);
 		if (ret)
 			break;
+
+		set_bit(NVME_FC_Q_LIVE, &ctrl->queues[i].flags);
 	}
 
 	return ret;
@@ -2320,6 +2323,14 @@ nvme_fc_start_fcp_op(struct nvme_fc_ctrl *ctrl, struct nvme_fc_queue *queue,
 	return BLK_STS_RESOURCE;
 }
 
+static inline blk_status_t nvme_fc_is_ready(struct nvme_fc_queue *queue,
+		struct request *rq)
+{
+	if (unlikely(!test_bit(NVME_FC_Q_LIVE, &queue->flags)))
+		return nvmf_check_init_req(&queue->ctrl->ctrl, rq);
+	return BLK_STS_OK;
+}
+
 static blk_status_t
 nvme_fc_queue_rq(struct blk_mq_hw_ctx *hctx,
 			const struct blk_mq_queue_data *bd)
@@ -2335,6 +2346,10 @@ nvme_fc_queue_rq(struct blk_mq_hw_ctx *hctx,
 	u32 data_len;
 	blk_status_t ret;
 
+	ret = nvme_fc_is_ready(queue, rq);
+	if (unlikely(ret))
+		return ret;
+
 	ret = nvme_setup_cmd(ns, rq, sqe);
 	if (ret)
 		return ret;
@@ -2727,6 +2742,8 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl)
 	if (ret)
 		goto out_disconnect_admin_queue;
 
+	set_bit(NVME_FC_Q_LIVE, &ctrl->queues[0].flags);
+
 	/*
 	 * Check controller capabilities
 	 *

drivers/nvme/host/multipath.c

@@ -131,7 +131,7 @@ static blk_qc_t nvme_ns_head_make_request(struct request_queue *q,
 		bio->bi_opf |= REQ_NVME_MPATH;
 		ret = direct_make_request(bio);
 	} else if (!list_empty_careful(&head->list)) {
-		dev_warn_ratelimited(dev, "no path available - requeing I/O\n");
+		dev_warn_ratelimited(dev, "no path available - requeuing I/O\n");
 
 		spin_lock_irq(&head->requeue_lock);
 		bio_list_add(&head->requeue_list, bio);

drivers/nvme/host/nvme.h

@@ -114,7 +114,7 @@ static inline struct nvme_request *nvme_req(struct request *req)
  * NVME_QUIRK_DELAY_BEFORE_CHK_RDY quirk enabled. The value (in ms) was
  * found empirically.
  */
-#define NVME_QUIRK_DELAY_AMOUNT		2000
+#define NVME_QUIRK_DELAY_AMOUNT		2300
 
 enum nvme_ctrl_state {
 	NVME_CTRL_NEW,