Merge branch 'for-6.5/block-late' into block-6.5

* for-6.5/block-late:
  blk-sysfs: add a new attr_group for blk_mq
  blk-iocost: move wbt_enable/disable_default() out of spinlock
  blk-wbt: cleanup rwb_enabled() and wbt_disabled()
  blk-wbt: remove dead code to handle wbt enable/disable with io inflight
  blk-wbt: don't create wbt sysfs entry if CONFIG_BLK_WBT is disabled
  blk-mq: fix two misuses on RQF_USE_SCHED
  blk-throttle: Fix io statistics for cgroup v1
  bcache: Fix bcache device claiming
  bcache: Alloc holder object before async registration
  raid10: avoid spin_lock from fastpath from raid10_unplug()
  md: fix 'delete_mutex' deadlock
  md: use mddev->external to select holder in export_rdev()
  md/raid1-10: fix casting from randomized structure in raid1_submit_write()
  md/raid10: fix the condition to call bio_end_io_acct()

block/blk-cgroup.c

@@ -2086,6 +2086,9 @@ void blk_cgroup_bio_start(struct bio *bio)
 	struct blkg_iostat_set *bis;
 	unsigned long flags;
 
+	if (!cgroup_subsys_on_dfl(io_cgrp_subsys))
+		return;
+
 	/* Root-level stats are sourced from system-wide IO stats */
 	if (!cgroup_parent(blkcg->css.cgroup))
 		return;
@@ -2116,8 +2119,7 @@ void blk_cgroup_bio_start(struct bio *bio)
 	}
 
 	u64_stats_update_end_irqrestore(&bis->sync, flags);
-	if (cgroup_subsys_on_dfl(io_cgrp_subsys))
+	cgroup_rstat_updated(blkcg->css.cgroup, cpu);
-		cgroup_rstat_updated(blkcg->css.cgroup, cpu);
 	put_cpu();
 }
 

block/blk-iocost.c

@@ -3301,11 +3301,9 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
 		blk_stat_enable_accounting(disk->queue);
 		blk_queue_flag_set(QUEUE_FLAG_RQ_ALLOC_TIME, disk->queue);
 		ioc->enabled = true;
-		wbt_disable_default(disk);
 	} else {
 		blk_queue_flag_clear(QUEUE_FLAG_RQ_ALLOC_TIME, disk->queue);
 		ioc->enabled = false;
-		wbt_enable_default(disk);
 	}
 
 	if (user) {
@@ -3318,6 +3316,11 @@ static ssize_t ioc_qos_write(struct kernfs_open_file *of, char *input,
 	ioc_refresh_params(ioc, true);
 	spin_unlock_irq(&ioc->lock);
 
+	if (enable)
+		wbt_disable_default(disk);
+	else
+		wbt_enable_default(disk);
+
 	blk_mq_unquiesce_queue(disk->queue);
 	blk_mq_unfreeze_queue(disk->queue);
 

block/blk-mq.c

@@ -1280,7 +1280,11 @@ static void blk_add_rq_to_plug(struct blk_plug *plug, struct request *rq)
 
 	if (!plug->multiple_queues && last && last->q != rq->q)
 		plug->multiple_queues = true;
-	if (!plug->has_elevator && (rq->rq_flags & RQF_USE_SCHED))
+	/*
+	 * Any request allocated from sched tags can't be issued to
+	 * ->queue_rqs() directly
+	 */
+	if (!plug->has_elevator && (rq->rq_flags & RQF_SCHED_TAGS))
 		plug->has_elevator = true;
 	rq->rq_next = NULL;
 	rq_list_add(&plug->mq_list, rq);

block/blk-sysfs.c

@@ -47,19 +47,6 @@ queue_var_store(unsigned long *var, const char *page, size_t count)
 	return count;
 }
 
-static ssize_t queue_var_store64(s64 *var, const char *page)
-{
-	int err;
-	s64 v;
-
-	err = kstrtos64(page, 10, &v);
-	if (err < 0)
-		return err;
-
-	*var = v;
-	return 0;
-}
-
 static ssize_t queue_requests_show(struct request_queue *q, char *page)
 {
 	return queue_var_show(q->nr_requests, page);
@@ -451,61 +438,6 @@ static ssize_t queue_io_timeout_store(struct request_queue *q, const char *page,
 	return count;
 }
 
-static ssize_t queue_wb_lat_show(struct request_queue *q, char *page)
-{
-	if (!wbt_rq_qos(q))
-		return -EINVAL;
-
-	if (wbt_disabled(q))
-		return sprintf(page, "0\n");
-
-	return sprintf(page, "%llu\n", div_u64(wbt_get_min_lat(q), 1000));
-}
-
-static ssize_t queue_wb_lat_store(struct request_queue *q, const char *page,
-				  size_t count)
-{
-	struct rq_qos *rqos;
-	ssize_t ret;
-	s64 val;
-
-	ret = queue_var_store64(&val, page);
-	if (ret < 0)
-		return ret;
-	if (val < -1)
-		return -EINVAL;
-
-	rqos = wbt_rq_qos(q);
-	if (!rqos) {
-		ret = wbt_init(q->disk);
-		if (ret)
-			return ret;
-	}
-
-	if (val == -1)
-		val = wbt_default_latency_nsec(q);
-	else if (val >= 0)
-		val *= 1000ULL;
-
-	if (wbt_get_min_lat(q) == val)
-		return count;
-
-	/*
-	 * Ensure that the queue is idled, in case the latency update
-	 * ends up either enabling or disabling wbt completely. We can't
-	 * have IO inflight if that happens.
-	 */
-	blk_mq_freeze_queue(q);
-	blk_mq_quiesce_queue(q);
-
-	wbt_set_min_lat(q, val);
-
-	blk_mq_unquiesce_queue(q);
-	blk_mq_unfreeze_queue(q);
-
-	return count;
-}
-
 static ssize_t queue_wc_show(struct request_queue *q, char *page)
 {
 	if (test_bit(QUEUE_FLAG_WC, &q->queue_flags))
@@ -598,7 +530,6 @@ QUEUE_RW_ENTRY(queue_wc, "write_cache");
 QUEUE_RO_ENTRY(queue_fua, "fua");
 QUEUE_RO_ENTRY(queue_dax, "dax");
 QUEUE_RW_ENTRY(queue_io_timeout, "io_timeout");
-QUEUE_RW_ENTRY(queue_wb_lat, "wbt_lat_usec");
 QUEUE_RO_ENTRY(queue_virt_boundary_mask, "virt_boundary_mask");
 QUEUE_RO_ENTRY(queue_dma_alignment, "dma_alignment");
 
@@ -617,8 +548,79 @@ QUEUE_RW_ENTRY(queue_iostats, "iostats");
 QUEUE_RW_ENTRY(queue_random, "add_random");
 QUEUE_RW_ENTRY(queue_stable_writes, "stable_writes");
 
+#ifdef CONFIG_BLK_WBT
+static ssize_t queue_var_store64(s64 *var, const char *page)
+{
+	int err;
+	s64 v;
+
+	err = kstrtos64(page, 10, &v);
+	if (err < 0)
+		return err;
+
+	*var = v;
+	return 0;
+}
+
+static ssize_t queue_wb_lat_show(struct request_queue *q, char *page)
+{
+	if (!wbt_rq_qos(q))
+		return -EINVAL;
+
+	if (wbt_disabled(q))
+		return sprintf(page, "0\n");
+
+	return sprintf(page, "%llu\n", div_u64(wbt_get_min_lat(q), 1000));
+}
+
+static ssize_t queue_wb_lat_store(struct request_queue *q, const char *page,
+				  size_t count)
+{
+	struct rq_qos *rqos;
+	ssize_t ret;
+	s64 val;
+
+	ret = queue_var_store64(&val, page);
+	if (ret < 0)
+		return ret;
+	if (val < -1)
+		return -EINVAL;
+
+	rqos = wbt_rq_qos(q);
+	if (!rqos) {
+		ret = wbt_init(q->disk);
+		if (ret)
+			return ret;
+	}
+
+	if (val == -1)
+		val = wbt_default_latency_nsec(q);
+	else if (val >= 0)
+		val *= 1000ULL;
+
+	if (wbt_get_min_lat(q) == val)
+		return count;
+
+	/*
+	 * Ensure that the queue is idled, in case the latency update
+	 * ends up either enabling or disabling wbt completely. We can't
+	 * have IO inflight if that happens.
+	 */
+	blk_mq_freeze_queue(q);
+	blk_mq_quiesce_queue(q);
+
+	wbt_set_min_lat(q, val);
+
+	blk_mq_unquiesce_queue(q);
+	blk_mq_unfreeze_queue(q);
+
+	return count;
+}
+
+QUEUE_RW_ENTRY(queue_wb_lat, "wbt_lat_usec");
+#endif
+
 static struct attribute *queue_attrs[] = {
-	&queue_requests_entry.attr,
 	&queue_ra_entry.attr,
 	&queue_max_hw_sectors_entry.attr,
 	&queue_max_sectors_entry.attr,
@@ -626,7 +628,6 @@ static struct attribute *queue_attrs[] = {
 	&queue_max_discard_segments_entry.attr,
 	&queue_max_integrity_segments_entry.attr,
 	&queue_max_segment_size_entry.attr,
-	&elv_iosched_entry.attr,
 	&queue_hw_sector_size_entry.attr,
 	&queue_logical_block_size_entry.attr,
 	&queue_physical_block_size_entry.attr,
@@ -647,7 +648,6 @@ static struct attribute *queue_attrs[] = {
 	&queue_max_open_zones_entry.attr,
 	&queue_max_active_zones_entry.attr,
 	&queue_nomerges_entry.attr,
-	&queue_rq_affinity_entry.attr,
 	&queue_iostats_entry.attr,
 	&queue_stable_writes_entry.attr,
 	&queue_random_entry.attr,
@@ -655,9 +655,7 @@ static struct attribute *queue_attrs[] = {
 	&queue_wc_entry.attr,
 	&queue_fua_entry.attr,
 	&queue_dax_entry.attr,
-	&queue_wb_lat_entry.attr,
 	&queue_poll_delay_entry.attr,
-	&queue_io_timeout_entry.attr,
 #ifdef CONFIG_BLK_DEV_THROTTLING_LOW
 	&blk_throtl_sample_time_entry.attr,
 #endif
@@ -666,16 +664,23 @@ static struct attribute *queue_attrs[] = {
 	NULL,
 };
 
+static struct attribute *blk_mq_queue_attrs[] = {
+	&queue_requests_entry.attr,
+	&elv_iosched_entry.attr,
+	&queue_rq_affinity_entry.attr,
+	&queue_io_timeout_entry.attr,
+#ifdef CONFIG_BLK_WBT
+	&queue_wb_lat_entry.attr,
+#endif
+	NULL,
+};
+
 static umode_t queue_attr_visible(struct kobject *kobj, struct attribute *attr,
 				int n)
 {
 	struct gendisk *disk = container_of(kobj, struct gendisk, queue_kobj);
 	struct request_queue *q = disk->queue;
 
-	if (attr == &queue_io_timeout_entry.attr &&
-		(!q->mq_ops || !q->mq_ops->timeout))
-			return 0;
-
 	if ((attr == &queue_max_open_zones_entry.attr ||
 	     attr == &queue_max_active_zones_entry.attr) &&
 	    !blk_queue_is_zoned(q))
@@ -684,11 +689,30 @@ static umode_t queue_attr_visible(struct kobject *kobj, struct attribute *attr,
 	return attr->mode;
 }
 
+static umode_t blk_mq_queue_attr_visible(struct kobject *kobj,
+					 struct attribute *attr, int n)
+{
+	struct gendisk *disk = container_of(kobj, struct gendisk, queue_kobj);
+	struct request_queue *q = disk->queue;
+
+	if (!queue_is_mq(q))
+		return 0;
+
+	if (attr == &queue_io_timeout_entry.attr && !q->mq_ops->timeout)
+		return 0;
+
+	return attr->mode;
+}
+
 static struct attribute_group queue_attr_group = {
 	.attrs = queue_attrs,
 	.is_visible = queue_attr_visible,
 };
 
+static struct attribute_group blk_mq_queue_attr_group = {
+	.attrs = blk_mq_queue_attrs,
+	.is_visible = blk_mq_queue_attr_visible,
+};
 
 #define to_queue(atr) container_of((atr), struct queue_sysfs_entry, attr)
 
@@ -733,6 +757,7 @@ static const struct sysfs_ops queue_sysfs_ops = {
 
 static const struct attribute_group *blk_queue_attr_groups[] = {
 	&queue_attr_group,
+	&blk_mq_queue_attr_group,
 	NULL
 };
 

block/blk-throttle.c

@@ -2178,12 +2178,6 @@ bool __blk_throtl_bio(struct bio *bio)
 
 	rcu_read_lock();
 
-	if (!cgroup_subsys_on_dfl(io_cgrp_subsys)) {
-		blkg_rwstat_add(&tg->stat_bytes, bio->bi_opf,
-				bio->bi_iter.bi_size);
-		blkg_rwstat_add(&tg->stat_ios, bio->bi_opf, 1);
-	}
-
 	spin_lock_irq(&q->queue_lock);
 
 	throtl_update_latency_buckets(td);

block/blk-throttle.h

@@ -185,6 +185,15 @@ static inline bool blk_should_throtl(struct bio *bio)
 	struct throtl_grp *tg = blkg_to_tg(bio->bi_blkg);
 	int rw = bio_data_dir(bio);
 
+	if (!cgroup_subsys_on_dfl(io_cgrp_subsys)) {
+		if (!bio_flagged(bio, BIO_CGROUP_ACCT)) {
+			bio_set_flag(bio, BIO_CGROUP_ACCT);
+			blkg_rwstat_add(&tg->stat_bytes, bio->bi_opf,
+					bio->bi_iter.bi_size);
+		}
+		blkg_rwstat_add(&tg->stat_ios, bio->bi_opf, 1);
+	}
+
 	/* iops limit is always counted */
 	if (tg->has_rules_iops[rw])
 		return true;

block/blk-wbt.c

@@ -146,7 +146,7 @@ enum {
 static inline bool rwb_enabled(struct rq_wb *rwb)
 {
 	return rwb && rwb->enable_state != WBT_STATE_OFF_DEFAULT &&
-		      rwb->wb_normal != 0;
+		      rwb->enable_state != WBT_STATE_OFF_MANUAL;
 }
 
 static void wb_timestamp(struct rq_wb *rwb, unsigned long *var)
@@ -200,15 +200,6 @@ static void wbt_rqw_done(struct rq_wb *rwb, struct rq_wait *rqw,
 
 	inflight = atomic_dec_return(&rqw->inflight);
 
-	/*
-	 * wbt got disabled with IO in flight. Wake up any potential
-	 * waiters, we don't have to do more than that.
-	 */
-	if (unlikely(!rwb_enabled(rwb))) {
-		rwb_wake_all(rwb);
-		return;
-	}
-
 	/*
 	 * For discards, our limit is always the background. For writes, if
 	 * the device does write back caching, drop further down before we
@@ -503,8 +494,7 @@ bool wbt_disabled(struct request_queue *q)
 {
 	struct rq_qos *rqos = wbt_rq_qos(q);
 
-	return !rqos || RQWB(rqos)->enable_state == WBT_STATE_OFF_DEFAULT ||
+	return !rqos || !rwb_enabled(RQWB(rqos));
-	       RQWB(rqos)->enable_state == WBT_STATE_OFF_MANUAL;
 }
 
 u64 wbt_get_min_lat(struct request_queue *q)
@@ -545,13 +535,6 @@ static inline unsigned int get_limit(struct rq_wb *rwb, blk_opf_t opf)
 {
 	unsigned int limit;
 
-	/*
-	 * If we got disabled, just return UINT_MAX. This ensures that
-	 * we'll properly inc a new IO, and dec+wakeup at the end.
-	 */
-	if (!rwb_enabled(rwb))
-		return UINT_MAX;
-
 	if ((opf & REQ_OP_MASK) == REQ_OP_DISCARD)
 		return rwb->wb_background;
 

block/blk-wbt.h

@@ -18,10 +18,6 @@ u64 wbt_default_latency_nsec(struct request_queue *);
 
 #else
 
-static inline int wbt_init(struct gendisk *disk)
-{
-	return -EINVAL;
-}
 static inline void wbt_disable_default(struct gendisk *disk)
 {
 }
@@ -31,21 +27,6 @@ static inline void wbt_enable_default(struct gendisk *disk)
 static inline void wbt_set_write_cache(struct request_queue *q, bool wc)
 {
 }
-static inline u64 wbt_get_min_lat(struct request_queue *q)
-{
-	return 0;
-}
-static inline void wbt_set_min_lat(struct request_queue *q, u64 val)
-{
-}
-static inline u64 wbt_default_latency_nsec(struct request_queue *q)
-{
-	return 0;
-}
-static inline bool wbt_disabled(struct request_queue *q)
-{
-	return true;
-}
 
 #endif /* CONFIG_BLK_WBT */
 

drivers/md/bcache/super.c

@@ -1369,7 +1369,7 @@ static void cached_dev_free(struct closure *cl)
 		put_page(virt_to_page(dc->sb_disk));
 
 	if (!IS_ERR_OR_NULL(dc->bdev))
-		blkdev_put(dc->bdev, bcache_kobj);
+		blkdev_put(dc->bdev, dc);
 
 	wake_up(&unregister_wait);
 
@@ -1453,7 +1453,6 @@ static int register_bdev(struct cache_sb *sb, struct cache_sb_disk *sb_disk,
 
 	memcpy(&dc->sb, sb, sizeof(struct cache_sb));
 	dc->bdev = bdev;
-	dc->bdev->bd_holder = dc;
 	dc->sb_disk = sb_disk;
 
 	if (cached_dev_init(dc, sb->block_size << 9))
@@ -2218,7 +2217,7 @@ void bch_cache_release(struct kobject *kobj)
 		put_page(virt_to_page(ca->sb_disk));
 
 	if (!IS_ERR_OR_NULL(ca->bdev))
-		blkdev_put(ca->bdev, bcache_kobj);
+		blkdev_put(ca->bdev, ca);
 
 	kfree(ca);
 	module_put(THIS_MODULE);
@@ -2345,7 +2344,6 @@ static int register_cache(struct cache_sb *sb, struct cache_sb_disk *sb_disk,
 
 	memcpy(&ca->sb, sb, sizeof(struct cache_sb));
 	ca->bdev = bdev;
-	ca->bdev->bd_holder = ca;
 	ca->sb_disk = sb_disk;
 
 	if (bdev_max_discard_sectors((bdev)))
@@ -2359,7 +2357,7 @@ static int register_cache(struct cache_sb *sb, struct cache_sb_disk *sb_disk,
 		 * call blkdev_put() to bdev in bch_cache_release(). So we
 		 * explicitly call blkdev_put() here.
 		 */
-		blkdev_put(bdev, bcache_kobj);
+		blkdev_put(bdev, ca);
 		if (ret == -ENOMEM)
 			err = "cache_alloc(): -ENOMEM";
 		else if (ret == -EPERM)
@@ -2448,6 +2446,7 @@ struct async_reg_args {
 	struct cache_sb *sb;
 	struct cache_sb_disk *sb_disk;
 	struct block_device *bdev;
+	void *holder;
 };
 
 static void register_bdev_worker(struct work_struct *work)
@@ -2455,22 +2454,13 @@ static void register_bdev_worker(struct work_struct *work)
 	int fail = false;
 	struct async_reg_args *args =
 		container_of(work, struct async_reg_args, reg_work.work);
-	struct cached_dev *dc;
-
-	dc = kzalloc(sizeof(*dc), GFP_KERNEL);
-	if (!dc) {
-		fail = true;
-		put_page(virt_to_page(args->sb_disk));
-		blkdev_put(args->bdev, bcache_kobj);
-		goto out;
-	}
 
 	mutex_lock(&bch_register_lock);
-	if (register_bdev(args->sb, args->sb_disk, args->bdev, dc) < 0)
+	if (register_bdev(args->sb, args->sb_disk, args->bdev, args->holder)
+	    < 0)
 		fail = true;
 	mutex_unlock(&bch_register_lock);
 
-out:
 	if (fail)
 		pr_info("error %s: fail to register backing device\n",
 			args->path);
@@ -2485,21 +2475,11 @@ static void register_cache_worker(struct work_struct *work)
 	int fail = false;
 	struct async_reg_args *args =
 		container_of(work, struct async_reg_args, reg_work.work);
-	struct cache *ca;
-
-	ca = kzalloc(sizeof(*ca), GFP_KERNEL);
-	if (!ca) {
-		fail = true;
-		put_page(virt_to_page(args->sb_disk));
-		blkdev_put(args->bdev, bcache_kobj);
-		goto out;
-	}
 
 	/* blkdev_put() will be called in bch_cache_release() */
-	if (register_cache(args->sb, args->sb_disk, args->bdev, ca) != 0)
+	if (register_cache(args->sb, args->sb_disk, args->bdev, args->holder))
 		fail = true;
 
-out:
 	if (fail)
 		pr_info("error %s: fail to register cache device\n",
 			args->path);
@@ -2520,6 +2500,13 @@ static void register_device_async(struct async_reg_args *args)
 	queue_delayed_work(system_wq, &args->reg_work, 10);
 }
 
+static void *alloc_holder_object(struct cache_sb *sb)
+{
+	if (SB_IS_BDEV(sb))
+		return kzalloc(sizeof(struct cached_dev), GFP_KERNEL);
+	return kzalloc(sizeof(struct cache), GFP_KERNEL);
+}
+
 static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 			       const char *buffer, size_t size)
 {
@@ -2527,9 +2514,11 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 	char *path = NULL;
 	struct cache_sb *sb;
 	struct cache_sb_disk *sb_disk;
-	struct block_device *bdev;
+	struct block_device *bdev, *bdev2;
+	void *holder = NULL;
 	ssize_t ret;
 	bool async_registration = false;
+	bool quiet = false;
 
 #ifdef CONFIG_BCACHE_ASYNC_REGISTRATION
 	async_registration = true;
@@ -2558,10 +2547,34 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 
 	ret = -EINVAL;
 	err = "failed to open device";
-	bdev = blkdev_get_by_path(strim(path), BLK_OPEN_READ | BLK_OPEN_WRITE,
+	bdev = blkdev_get_by_path(strim(path), BLK_OPEN_READ, NULL, NULL);
-				  bcache_kobj, NULL);
+	if (IS_ERR(bdev))
+		goto out_free_sb;
+
+	err = "failed to set blocksize";
+	if (set_blocksize(bdev, 4096))
+		goto out_blkdev_put;
+
+	err = read_super(sb, bdev, &sb_disk);
+	if (err)
+		goto out_blkdev_put;
+
+	holder = alloc_holder_object(sb);
+	if (!holder) {
+		ret = -ENOMEM;
+		err = "cannot allocate memory";
+		goto out_put_sb_page;
+	}
+
+	/* Now reopen in exclusive mode with proper holder */
+	bdev2 = blkdev_get_by_dev(bdev->bd_dev, BLK_OPEN_READ | BLK_OPEN_WRITE,
+				  holder, NULL);
+	blkdev_put(bdev, NULL);
+	bdev = bdev2;
 	if (IS_ERR(bdev)) {
-		if (bdev == ERR_PTR(-EBUSY)) {
+		ret = PTR_ERR(bdev);
+		bdev = NULL;
+		if (ret == -EBUSY) {
 			dev_t dev;
 
 			mutex_lock(&bch_register_lock);
@@ -2571,20 +2584,14 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 			else
 				err = "device busy";
 			mutex_unlock(&bch_register_lock);
-			if (attr == &ksysfs_register_quiet)
+			if (attr == &ksysfs_register_quiet) {
-				goto done;
+				quiet = true;
+				ret = size;
+			}
 		}
-		goto out_free_sb;
+		goto out_free_holder;
 	}
 
-	err = "failed to set blocksize";
-	if (set_blocksize(bdev, 4096))
-		goto out_blkdev_put;
-
-	err = read_super(sb, bdev, &sb_disk);
-	if (err)
-		goto out_blkdev_put;
-
 	err = "failed to register device";
 
 	if (async_registration) {
@@ -2595,59 +2602,46 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 		if (!args) {
 			ret = -ENOMEM;
 			err = "cannot allocate memory";
-			goto out_put_sb_page;
+			goto out_free_holder;
 		}
 
 		args->path	= path;
 		args->sb	= sb;
 		args->sb_disk	= sb_disk;
 		args->bdev	= bdev;
+		args->holder	= holder;
 		register_device_async(args);
 		/* No wait and returns to user space */
 		goto async_done;
 	}
 
 	if (SB_IS_BDEV(sb)) {
-		struct cached_dev *dc = kzalloc(sizeof(*dc), GFP_KERNEL);
-
-		if (!dc) {
-			ret = -ENOMEM;
-			err = "cannot allocate memory";
-			goto out_put_sb_page;
-		}
-
 		mutex_lock(&bch_register_lock);
-		ret = register_bdev(sb, sb_disk, bdev, dc);
+		ret = register_bdev(sb, sb_disk, bdev, holder);
 		mutex_unlock(&bch_register_lock);
 		/* blkdev_put() will be called in cached_dev_free() */
 		if (ret < 0)
 			goto out_free_sb;
 	} else {
-		struct cache *ca = kzalloc(sizeof(*ca), GFP_KERNEL);
-
-		if (!ca) {
-			ret = -ENOMEM;
-			err = "cannot allocate memory";
-			goto out_put_sb_page;
-		}
-
 		/* blkdev_put() will be called in bch_cache_release() */
-		ret = register_cache(sb, sb_disk, bdev, ca);
+		ret = register_cache(sb, sb_disk, bdev, holder);
 		if (ret)
 			goto out_free_sb;
 	}
 
-done:
 	kfree(sb);
 	kfree(path);
 	module_put(THIS_MODULE);
 async_done:
 	return size;
 
+out_free_holder:
+	kfree(holder);
 out_put_sb_page:
 	put_page(virt_to_page(sb_disk));
 out_blkdev_put:
-	blkdev_put(bdev, register_bcache);
+	if (bdev)
+		blkdev_put(bdev, holder);
 out_free_sb:
 	kfree(sb);
 out_free_path:
@@ -2656,7 +2650,8 @@ out_free_path:
 out_module_put:
 	module_put(THIS_MODULE);
 out:
-	pr_info("error %s: %s\n", path?path:"", err);
+	if (!quiet)
+		pr_info("error %s: %s\n", path?path:"", err);
 	return ret;
 }
 

drivers/md/md.c

@@ -643,7 +643,6 @@ void mddev_init(struct mddev *mddev)
 {
 	mutex_init(&mddev->open_mutex);
 	mutex_init(&mddev->reconfig_mutex);
-	mutex_init(&mddev->delete_mutex);
 	mutex_init(&mddev->bitmap_info.mutex);
 	INIT_LIST_HEAD(&mddev->disks);
 	INIT_LIST_HEAD(&mddev->all_mddevs);
@@ -749,26 +748,15 @@ static void mddev_free(struct mddev *mddev)
 
 static const struct attribute_group md_redundancy_group;
 
-static void md_free_rdev(struct mddev *mddev)
+void mddev_unlock(struct mddev *mddev)
 {
 	struct md_rdev *rdev;
 	struct md_rdev *tmp;
+	LIST_HEAD(delete);
 
-	mutex_lock(&mddev->delete_mutex);
+	if (!list_empty(&mddev->deleting))
-	if (list_empty(&mddev->deleting))
+		list_splice_init(&mddev->deleting, &delete);
-		goto out;
 
-	list_for_each_entry_safe(rdev, tmp, &mddev->deleting, same_set) {
-		list_del_init(&rdev->same_set);
-		kobject_del(&rdev->kobj);
-		export_rdev(rdev, mddev);
-	}
-out:
-	mutex_unlock(&mddev->delete_mutex);
-}
-
-void mddev_unlock(struct mddev *mddev)
-{
 	if (mddev->to_remove) {
 		/* These cannot be removed under reconfig_mutex as
 		 * an access to the files will try to take reconfig_mutex
@@ -808,7 +796,11 @@ void mddev_unlock(struct mddev *mddev)
 	} else
 		mutex_unlock(&mddev->reconfig_mutex);
 
-	md_free_rdev(mddev);
+	list_for_each_entry_safe(rdev, tmp, &delete, same_set) {
+		list_del_init(&rdev->same_set);
+		kobject_del(&rdev->kobj);
+		export_rdev(rdev, mddev);
+	}
 
 	md_wakeup_thread(mddev->thread);
 	wake_up(&mddev->sb_wait);
@@ -2458,7 +2450,7 @@ static void export_rdev(struct md_rdev *rdev, struct mddev *mddev)
 	if (test_bit(AutoDetected, &rdev->flags))
 		md_autodetect_dev(rdev->bdev->bd_dev);
 #endif
-	blkdev_put(rdev->bdev, mddev->major_version == -2 ? &claim_rdev : rdev);
+	blkdev_put(rdev->bdev, mddev->external ? &claim_rdev : rdev);
 	rdev->bdev = NULL;
 	kobject_put(&rdev->kobj);
 }
@@ -2488,9 +2480,7 @@ static void md_kick_rdev_from_array(struct md_rdev *rdev)
 	 * reconfig_mutex is held, hence it can't be called under
 	 * reconfig_mutex and it's delayed to mddev_unlock().
 	 */
-	mutex_lock(&mddev->delete_mutex);
 	list_add(&rdev->same_set, &mddev->deleting);
-	mutex_unlock(&mddev->delete_mutex);
 }
 
 static void export_array(struct mddev *mddev)
@@ -6140,7 +6130,7 @@ static void md_clean(struct mddev *mddev)
 	mddev->resync_min = 0;
 	mddev->resync_max = MaxSector;
 	mddev->reshape_position = MaxSector;
-	mddev->external = 0;
+	/* we still need mddev->external in export_rdev, do not clear it yet */
 	mddev->persistent = 0;
 	mddev->level = LEVEL_NONE;
 	mddev->clevel[0] = 0;

drivers/md/md.h

@@ -531,11 +531,9 @@ struct mddev {
 
 	/*
 	 * Temporarily store rdev that will be finally removed when
-	 * reconfig_mutex is unlocked.
+	 * reconfig_mutex is unlocked, protected by reconfig_mutex.
 	 */
 	struct list_head		deleting;
-	/* Protect the deleting list */
-	struct mutex			delete_mutex;
 
 	bool	has_superblocks:1;
 	bool	fail_last_dev:1;

drivers/md/raid1-10.c

@@ -116,7 +116,7 @@ static void md_bio_reset_resync_pages(struct bio *bio, struct resync_pages *rp,
 
 static inline void raid1_submit_write(struct bio *bio)
 {
-	struct md_rdev *rdev = (struct md_rdev *)bio->bi_bdev;
+	struct md_rdev *rdev = (void *)bio->bi_bdev;
 
 	bio->bi_next = NULL;
 	bio_set_dev(bio, rdev->bdev);

drivers/md/raid10.c

@@ -325,7 +325,7 @@ static void raid_end_bio_io(struct r10bio *r10_bio)
 	if (!test_bit(R10BIO_Uptodate, &r10_bio->state))
 		bio->bi_status = BLK_STS_IOERR;
 
-	if (blk_queue_io_stat(bio->bi_bdev->bd_disk->queue))
+	if (r10_bio->start_time)
 		bio_end_io_acct(bio, r10_bio->start_time);
 	bio_endio(bio);
 	/*
@@ -1118,7 +1118,7 @@ static void raid10_unplug(struct blk_plug_cb *cb, bool from_schedule)
 		spin_lock_irq(&conf->device_lock);
 		bio_list_merge(&conf->pending_bio_list, &plug->pending);
 		spin_unlock_irq(&conf->device_lock);
-		wake_up(&conf->wait_barrier);
+		wake_up_barrier(conf);
 		md_wakeup_thread(mddev->thread);
 		kfree(plug);
 		return;
@@ -1127,7 +1127,7 @@ static void raid10_unplug(struct blk_plug_cb *cb, bool from_schedule)
 	/* we aren't scheduling, so we can do the write-out directly. */
 	bio = bio_list_get(&plug->pending);
 	raid1_prepare_flush_writes(mddev->bitmap);
-	wake_up(&conf->wait_barrier);
+	wake_up_barrier(conf);
 
 	while (bio) { /* submit pending writes */
 		struct bio *next = bio->bi_next;

include/linux/blk-mq.h

@@ -852,7 +852,11 @@ static inline bool blk_mq_add_to_batch(struct request *req,
 				       struct io_comp_batch *iob, int ioerror,
 				       void (*complete)(struct io_comp_batch *))
 {
-	if (!iob || (req->rq_flags & RQF_USE_SCHED) || ioerror ||
+	/*
+	 * blk_mq_end_request_batch() can't end request allocated from
+	 * sched tags
+	 */
+	if (!iob || (req->rq_flags & RQF_SCHED_TAGS) || ioerror ||
 			(req->end_io && !blk_rq_is_passthrough(req)))
 		return false;