block-6.5-2023-07-03

-----BEGIN PGP SIGNATURE-----
 
 iQJEBAABCAAuFiEEwPw5LcreJtl1+l5K99NY+ylx4KYFAmSjJ2IQHGF4Ym9lQGtl
 cm5lbC5kawAKCRD301j7KXHgpsQMEACQiUBw81tXvetYhz3P/4KrrjvUobgqMU0w
 jtrxqMgPee9FbqCShpj76c+La5wu23DnlrCXoHZxFQuiQLnsX5xFV66NYVi+W1CN
 k5MHP7f2e9V0T7qJ9UoHFRV1k22LF4X6T8njEZimxsm/uXfpav/knkhI7nUDnB1K
 wxlu9akD2Bo/X9O2NTS+X6qjoawZ6rDWN15THMXlC45VzJPLmIcs07Ev+mvw21KE
 XqasoZrxEO0S8dWxmJgJGqnRIOQptTS5U+0OPBZT8H220Qp/1q0pQHPw6iLXNrkc
 w1a2W1Bge012gjJt7gCMkdDnZb76sKiyGuMbFME7DoRbLCQeaOtoSfmg7NoRI2gp
 74TCSr7dPWZUVUy5Tmsy0DCv0552vIbnlQ69W6Xwx8YkplM3FPiMpWrQ5JWEHdvv
 Zl84mLP6Yyo54JVuk9zi8q/2L0HfyfMDj4UM/mNs8hwmcUSbPO2TKdIWDaq8xPuS
 Ed+D+kg6XFux8tLnCSDLNbaD5JE+ak9gTVhNdRa/zFE04o/OeidscKEqRSYTkdXL
 2p34qtw5kEQocO4Pa3eUGO6KJCDTR36Rms5p6ZFybL4O2oZYrAbRi1TGDxaG2Hag
 GCr2vaFbmz1zbGuMpFhLha5B7HeDLs+PHOn+B1iUNjEr9RC0EOHV7moJKqjxlnCh
 4mBkK/Nlyg==
 =kSeX
 -----END PGP SIGNATURE-----

Merge tag 'block-6.5-2023-07-03' of git://git.kernel.dk/linux

Pull more block updates from Jens Axboe:
 "Mostly items that came in a bit late for the initial pull request,
  wanted to make sure they had the appropriate amount of linux-next soak
  before going upstream.

  Outside of stragglers, just generic fixes for either merge window
  items, or longer standing bugs"

* tag 'block-6.5-2023-07-03' of git://git.kernel.dk/linux: (25 commits)
  md/raid0: add discard support for the 'original' layout
  nvme: disable controller on reset state failure
  nvme: sync timeout work on failed reset
  nvme: ensure unquiesce on teardown
  cdrom/gdrom: Fix build error
  nvme: improved uring polling
  block: add request polling helper
  nvme-mpath: fix I/O failure with EAGAIN when failing over I/O
  nvme: host: fix command name spelling
  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
  ...

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

@@ -49,17 +49,8 @@ static void blk_mq_request_bypass_insert(struct request *rq,
 		blk_insert_t flags);
 static void blk_mq_try_issue_list_directly(struct blk_mq_hw_ctx *hctx,
 		struct list_head *list);
-
+static int blk_hctx_poll(struct request_queue *q, struct blk_mq_hw_ctx *hctx,
-static inline struct blk_mq_hw_ctx *blk_qc_to_hctx(struct request_queue *q,
+			 struct io_comp_batch *iob, unsigned int flags);
-		blk_qc_t qc)
-{
-	return xa_load(&q->hctx_table, qc);
-}
-
-static inline blk_qc_t blk_rq_to_qc(struct request *rq)
-{
-	return rq->mq_hctx->queue_num;
-}
 
 /*
  * Check if any of the ctx, dispatch list or elevator
@@ -1248,7 +1239,7 @@ void blk_mq_start_request(struct request *rq)
 		q->integrity.profile->prepare_fn(rq);
 #endif
 	if (rq->bio && rq->bio->bi_opf & REQ_POLLED)
-	        WRITE_ONCE(rq->bio->bi_cookie, blk_rq_to_qc(rq));
+	        WRITE_ONCE(rq->bio->bi_cookie, rq->mq_hctx->queue_num);
 }
 EXPORT_SYMBOL(blk_mq_start_request);
 
@@ -1280,7 +1271,11 @@ 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);
@@ -1350,7 +1345,7 @@ EXPORT_SYMBOL_GPL(blk_rq_is_poll);
 static void blk_rq_poll_completion(struct request *rq, struct completion *wait)
 {
 	do {
-		blk_mq_poll(rq->q, blk_rq_to_qc(rq), NULL, 0);
+		blk_hctx_poll(rq->q, rq->mq_hctx, NULL, 0);
 		cond_resched();
 	} while (!completion_done(wait));
 }
@@ -4745,10 +4740,9 @@ void blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, int nr_hw_queues)
 }
 EXPORT_SYMBOL_GPL(blk_mq_update_nr_hw_queues);
 
-int blk_mq_poll(struct request_queue *q, blk_qc_t cookie, struct io_comp_batch *iob,
+static int blk_hctx_poll(struct request_queue *q, struct blk_mq_hw_ctx *hctx,
-		unsigned int flags)
+			 struct io_comp_batch *iob, unsigned int flags)
 {
-	struct blk_mq_hw_ctx *hctx = blk_qc_to_hctx(q, cookie);
 	long state = get_current_state();
 	int ret;
 
@@ -4773,6 +4767,32 @@ int blk_mq_poll(struct request_queue *q, blk_qc_t cookie, struct io_comp_batch *
 	return 0;
 }
 
+int blk_mq_poll(struct request_queue *q, blk_qc_t cookie,
+		struct io_comp_batch *iob, unsigned int flags)
+{
+	struct blk_mq_hw_ctx *hctx = xa_load(&q->hctx_table, cookie);
+
+	return blk_hctx_poll(q, hctx, iob, flags);
+}
+
+int blk_rq_poll(struct request *rq, struct io_comp_batch *iob,
+		unsigned int poll_flags)
+{
+	struct request_queue *q = rq->q;
+	int ret;
+
+	if (!blk_rq_is_poll(rq))
+		return 0;
+	if (!percpu_ref_tryget(&q->q_usage_counter))
+		return 0;
+
+	ret = blk_hctx_poll(q, rq->mq_hctx, iob, poll_flags);
+	blk_queue_exit(q);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(blk_rq_poll);
+
 unsigned int blk_mq_rq_cpu(struct request *rq)
 {
 	return rq->mq_ctx->cpu;

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/cdrom/gdrom.c

@@ -481,7 +481,7 @@ static int gdrom_bdops_open(struct gendisk *disk, blk_mode_t mode)
 	disk_check_media_change(disk);
 
 	mutex_lock(&gdrom_mutex);
-	ret = cdrom_open(gd.cd_info);
+	ret = cdrom_open(gd.cd_info, mode);
 	mutex_unlock(&gdrom_mutex);
 	return ret;
 }
@@ -489,7 +489,7 @@ static int gdrom_bdops_open(struct gendisk *disk, blk_mode_t mode)
 static void gdrom_bdops_release(struct gendisk *disk)
 {
 	mutex_lock(&gdrom_mutex);
-	cdrom_release(gd.cd_info, mode);
+	cdrom_release(gd.cd_info);
 	mutex_unlock(&gdrom_mutex);
 }
 

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 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 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/raid0.c

@@ -270,6 +270,18 @@ static int create_strip_zones(struct mddev *mddev, struct r0conf **private_conf)
 		goto abort;
 	}
 
+	if (conf->layout == RAID0_ORIG_LAYOUT) {
+		for (i = 1; i < conf->nr_strip_zones; i++) {
+			sector_t first_sector = conf->strip_zone[i-1].zone_end;
+
+			sector_div(first_sector, mddev->chunk_sectors);
+			zone = conf->strip_zone + i;
+			/* disk_shift is first disk index used in the zone */
+			zone->disk_shift = sector_div(first_sector,
+						      zone->nb_dev);
+		}
+	}
+
 	pr_debug("md/raid0:%s: done.\n", mdname(mddev));
 	*private_conf = conf;
 
@@ -431,6 +443,20 @@ static int raid0_run(struct mddev *mddev)
 	return ret;
 }
 
+/*
+ * Convert disk_index to the disk order in which it is read/written.
+ *  For example, if we have 4 disks, they are numbered 0,1,2,3. If we
+ *  write the disks starting at disk 3, then the read/write order would
+ *  be disk 3, then 0, then 1, and then disk 2 and we want map_disk_shift()
+ *  to map the disks as follows 0,1,2,3 => 1,2,3,0. So disk 0 would map
+ *  to 1, 1 to 2, 2 to 3, and 3 to 0. That way we can compare disks in
+ *  that 'output' space to understand the read/write disk ordering.
+ */
+static int map_disk_shift(int disk_index, int num_disks, int disk_shift)
+{
+	return ((disk_index + num_disks - disk_shift) % num_disks);
+}
+
 static void raid0_handle_discard(struct mddev *mddev, struct bio *bio)
 {
 	struct r0conf *conf = mddev->private;
@@ -444,7 +470,9 @@ static void raid0_handle_discard(struct mddev *mddev, struct bio *bio)
 	sector_t end_disk_offset;
 	unsigned int end_disk_index;
 	unsigned int disk;
+	sector_t orig_start, orig_end;
 
+	orig_start = start;
 	zone = find_zone(conf, &start);
 
 	if (bio_end_sector(bio) > zone->zone_end) {
@@ -458,6 +486,7 @@ static void raid0_handle_discard(struct mddev *mddev, struct bio *bio)
 	} else
 		end = bio_end_sector(bio);
 
+	orig_end = end;
 	if (zone != conf->strip_zone)
 		end = end - zone[-1].zone_end;
 
@@ -469,13 +498,26 @@ static void raid0_handle_discard(struct mddev *mddev, struct bio *bio)
 	last_stripe_index = end;
 	sector_div(last_stripe_index, stripe_size);
 
-	start_disk_index = (int)(start - first_stripe_index * stripe_size) /
+	/* In the first zone the original and alternate layouts are the same */
-		mddev->chunk_sectors;
+	if ((conf->layout == RAID0_ORIG_LAYOUT) && (zone != conf->strip_zone)) {
+		sector_div(orig_start, mddev->chunk_sectors);
+		start_disk_index = sector_div(orig_start, zone->nb_dev);
+		start_disk_index = map_disk_shift(start_disk_index,
+						  zone->nb_dev,
+						  zone->disk_shift);
+		sector_div(orig_end, mddev->chunk_sectors);
+		end_disk_index = sector_div(orig_end, zone->nb_dev);
+		end_disk_index = map_disk_shift(end_disk_index,
+						zone->nb_dev, zone->disk_shift);
+	} else {
+		start_disk_index = (int)(start - first_stripe_index * stripe_size) /
+			mddev->chunk_sectors;
+		end_disk_index = (int)(end - last_stripe_index * stripe_size) /
+			mddev->chunk_sectors;
+	}
 	start_disk_offset = ((int)(start - first_stripe_index * stripe_size) %
 		mddev->chunk_sectors) +
 		first_stripe_index * mddev->chunk_sectors;
-	end_disk_index = (int)(end - last_stripe_index * stripe_size) /
-		mddev->chunk_sectors;
 	end_disk_offset = ((int)(end - last_stripe_index * stripe_size) %
 		mddev->chunk_sectors) +
 		last_stripe_index * mddev->chunk_sectors;
@@ -483,18 +525,22 @@ static void raid0_handle_discard(struct mddev *mddev, struct bio *bio)
 	for (disk = 0; disk < zone->nb_dev; disk++) {
 		sector_t dev_start, dev_end;
 		struct md_rdev *rdev;
+		int compare_disk;
 
-		if (disk < start_disk_index)
+		compare_disk = map_disk_shift(disk, zone->nb_dev,
+					      zone->disk_shift);
+
+		if (compare_disk < start_disk_index)
 			dev_start = (first_stripe_index + 1) *
 				mddev->chunk_sectors;
-		else if (disk > start_disk_index)
+		else if (compare_disk > start_disk_index)
 			dev_start = first_stripe_index * mddev->chunk_sectors;
 		else
 			dev_start = start_disk_offset;
 
-		if (disk < end_disk_index)
+		if (compare_disk < end_disk_index)
 			dev_end = (last_stripe_index + 1) * mddev->chunk_sectors;
-		else if (disk > end_disk_index)
+		else if (compare_disk > end_disk_index)
 			dev_end = last_stripe_index * mddev->chunk_sectors;
 		else
 			dev_end = end_disk_offset;

drivers/md/raid0.h

@@ -6,6 +6,7 @@ struct strip_zone {
 	sector_t zone_end;	/* Start of the next zone (in sectors) */
 	sector_t dev_start;	/* Zone offset in real dev (in sectors) */
 	int	 nb_dev;	/* # of devices attached to the zone */
+	int	 disk_shift;	/* start disk for the original layout */
 };
 
 /* Linux 3.14 (20d0189b101) made an unintended change to

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;

drivers/nvme/host/constants.c

@@ -12,7 +12,7 @@ static const char * const nvme_ops[] = {
 	[nvme_cmd_read] = "Read",
 	[nvme_cmd_write_uncor] = "Write Uncorrectable",
 	[nvme_cmd_compare] = "Compare",
-	[nvme_cmd_write_zeroes] = "Write Zeros",
+	[nvme_cmd_write_zeroes] = "Write Zeroes",
 	[nvme_cmd_dsm] = "Dataset Management",
 	[nvme_cmd_verify] = "Verify",
 	[nvme_cmd_resv_register] = "Reservation Register",

drivers/nvme/host/core.c

@@ -1134,8 +1134,11 @@ void nvme_passthru_end(struct nvme_ctrl *ctrl, struct nvme_ns *ns, u32 effects,
 		mutex_unlock(&ctrl->scan_lock);
 	}
 	if (effects & NVME_CMD_EFFECTS_CCC) {
-		dev_info(ctrl->device,
+		if (!test_and_set_bit(NVME_CTRL_DIRTY_CAPABILITY,
+				      &ctrl->flags)) {
+			dev_info(ctrl->device,
 "controller capabilities changed, reset may be required to take effect.\n");
+		}
 	}
 	if (effects & (NVME_CMD_EFFECTS_NIC | NVME_CMD_EFFECTS_NCC)) {
 		nvme_queue_scan(ctrl);
@@ -3177,6 +3180,7 @@ int nvme_init_ctrl_finish(struct nvme_ctrl *ctrl, bool was_suspended)
 			return ret;
 	}
 
+	clear_bit(NVME_CTRL_DIRTY_CAPABILITY, &ctrl->flags);
 	ctrl->identified = true;
 
 	return 0;

drivers/nvme/host/ioctl.c

@@ -505,7 +505,6 @@ static enum rq_end_io_ret nvme_uring_cmd_end_io(struct request *req,
 {
 	struct io_uring_cmd *ioucmd = req->end_io_data;
 	struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd);
-	void *cookie = READ_ONCE(ioucmd->cookie);
 
 	req->bio = pdu->bio;
 	if (nvme_req(req)->flags & NVME_REQ_CANCELLED)
@@ -518,10 +517,12 @@ static enum rq_end_io_ret nvme_uring_cmd_end_io(struct request *req,
 	 * For iopoll, complete it directly.
 	 * Otherwise, move the completion to task work.
 	 */
-	if (cookie != NULL && blk_rq_is_poll(req))
+	if (blk_rq_is_poll(req)) {
+		WRITE_ONCE(ioucmd->cookie, NULL);
 		nvme_uring_task_cb(ioucmd, IO_URING_F_UNLOCKED);
-	else
+	} else {
 		io_uring_cmd_do_in_task_lazy(ioucmd, nvme_uring_task_cb);
+	}
 
 	return RQ_END_IO_FREE;
 }
@@ -531,7 +532,6 @@ static enum rq_end_io_ret nvme_uring_cmd_end_io_meta(struct request *req,
 {
 	struct io_uring_cmd *ioucmd = req->end_io_data;
 	struct nvme_uring_cmd_pdu *pdu = nvme_uring_cmd_pdu(ioucmd);
-	void *cookie = READ_ONCE(ioucmd->cookie);
 
 	req->bio = pdu->bio;
 	pdu->req = req;
@@ -540,10 +540,12 @@ static enum rq_end_io_ret nvme_uring_cmd_end_io_meta(struct request *req,
 	 * For iopoll, complete it directly.
 	 * Otherwise, move the completion to task work.
 	 */
-	if (cookie != NULL && blk_rq_is_poll(req))
+	if (blk_rq_is_poll(req)) {
+		WRITE_ONCE(ioucmd->cookie, NULL);
 		nvme_uring_task_meta_cb(ioucmd, IO_URING_F_UNLOCKED);
-	else
+	} else {
 		io_uring_cmd_do_in_task_lazy(ioucmd, nvme_uring_task_meta_cb);
+	}
 
 	return RQ_END_IO_NONE;
 }
@@ -599,7 +601,6 @@ static int nvme_uring_cmd_io(struct nvme_ctrl *ctrl, struct nvme_ns *ns,
 	if (issue_flags & IO_URING_F_IOPOLL)
 		rq_flags |= REQ_POLLED;
 
-retry:
 	req = nvme_alloc_user_request(q, &c, rq_flags, blk_flags);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
@@ -613,17 +614,11 @@ static int nvme_uring_cmd_io(struct nvme_ctrl *ctrl, struct nvme_ns *ns,
 			return ret;
 	}
 
-	if (issue_flags & IO_URING_F_IOPOLL && rq_flags & REQ_POLLED) {
+	if (blk_rq_is_poll(req)) {
-		if (unlikely(!req->bio)) {
+		ioucmd->flags |= IORING_URING_CMD_POLLED;
-			/* we can't poll this, so alloc regular req instead */
+		WRITE_ONCE(ioucmd->cookie, req);
-			blk_mq_free_request(req);
-			rq_flags &= ~REQ_POLLED;
-			goto retry;
-		} else {
-			WRITE_ONCE(ioucmd->cookie, req->bio);
-			req->bio->bi_opf |= REQ_POLLED;
-		}
 	}
+
 	/* to free bio on completion, as req->bio will be null at that time */
 	pdu->bio = req->bio;
 	pdu->meta_len = d.metadata_len;
@@ -785,18 +780,16 @@ int nvme_ns_chr_uring_cmd_iopoll(struct io_uring_cmd *ioucmd,
 				 struct io_comp_batch *iob,
 				 unsigned int poll_flags)
 {
-	struct bio *bio;
+	struct request *req;
 	int ret = 0;
-	struct nvme_ns *ns;
+
-	struct request_queue *q;
+	if (!(ioucmd->flags & IORING_URING_CMD_POLLED))
+		return 0;
 
 	rcu_read_lock();
-	bio = READ_ONCE(ioucmd->cookie);
+	req = READ_ONCE(ioucmd->cookie);
-	ns = container_of(file_inode(ioucmd->file)->i_cdev,
+	if (req && blk_rq_is_poll(req))
-			struct nvme_ns, cdev);
+		ret = blk_rq_poll(req, iob, poll_flags);
-	q = ns->queue;
-	if (test_bit(QUEUE_FLAG_POLL, &q->queue_flags) && bio && bio->bi_bdev)
-		ret = bio_poll(bio, iob, poll_flags);
 	rcu_read_unlock();
 	return ret;
 }
@@ -890,31 +883,6 @@ int nvme_ns_head_chr_uring_cmd(struct io_uring_cmd *ioucmd,
 	srcu_read_unlock(&head->srcu, srcu_idx);
 	return ret;
 }
-
-int nvme_ns_head_chr_uring_cmd_iopoll(struct io_uring_cmd *ioucmd,
-				      struct io_comp_batch *iob,
-				      unsigned int poll_flags)
-{
-	struct cdev *cdev = file_inode(ioucmd->file)->i_cdev;
-	struct nvme_ns_head *head = container_of(cdev, struct nvme_ns_head, cdev);
-	int srcu_idx = srcu_read_lock(&head->srcu);
-	struct nvme_ns *ns = nvme_find_path(head);
-	struct bio *bio;
-	int ret = 0;
-	struct request_queue *q;
-
-	if (ns) {
-		rcu_read_lock();
-		bio = READ_ONCE(ioucmd->cookie);
-		q = ns->queue;
-		if (test_bit(QUEUE_FLAG_POLL, &q->queue_flags) && bio
-				&& bio->bi_bdev)
-			ret = bio_poll(bio, iob, poll_flags);
-		rcu_read_unlock();
-	}
-	srcu_read_unlock(&head->srcu, srcu_idx);
-	return ret;
-}
 #endif /* CONFIG_NVME_MULTIPATH */
 
 int nvme_dev_uring_cmd(struct io_uring_cmd *ioucmd, unsigned int issue_flags)

drivers/nvme/host/multipath.c

@@ -106,6 +106,14 @@ void nvme_failover_req(struct request *req)
 			bio->bi_opf &= ~REQ_POLLED;
 			bio->bi_cookie = BLK_QC_T_NONE;
 		}
+		/*
+		 * The alternate request queue that we may end up submitting
+		 * the bio to may be frozen temporarily, in this case REQ_NOWAIT
+		 * will fail the I/O immediately with EAGAIN to the issuer.
+		 * We are not in the issuer context which cannot block. Clear
+		 * the flag to avoid spurious EAGAIN I/O failures.
+		 */
+		bio->bi_opf &= ~REQ_NOWAIT;
 	}
 	blk_steal_bios(&ns->head->requeue_list, req);
 	spin_unlock_irqrestore(&ns->head->requeue_lock, flags);
@@ -470,7 +478,7 @@ static const struct file_operations nvme_ns_head_chr_fops = {
 	.unlocked_ioctl	= nvme_ns_head_chr_ioctl,
 	.compat_ioctl	= compat_ptr_ioctl,
 	.uring_cmd	= nvme_ns_head_chr_uring_cmd,
-	.uring_cmd_iopoll = nvme_ns_head_chr_uring_cmd_iopoll,
+	.uring_cmd_iopoll = nvme_ns_chr_uring_cmd_iopoll,
 };
 
 static int nvme_add_ns_head_cdev(struct nvme_ns_head *head)

drivers/nvme/host/nvme.h

@@ -250,6 +250,7 @@ enum nvme_ctrl_flags {
 	NVME_CTRL_STARTED_ONCE		= 2,
 	NVME_CTRL_STOPPED		= 3,
 	NVME_CTRL_SKIP_ID_CNS_CS	= 4,
+	NVME_CTRL_DIRTY_CAPABILITY	= 5,
 };
 
 struct nvme_ctrl {
@@ -856,8 +857,6 @@ long nvme_dev_ioctl(struct file *file, unsigned int cmd,
 		unsigned long arg);
 int nvme_ns_chr_uring_cmd_iopoll(struct io_uring_cmd *ioucmd,
 		struct io_comp_batch *iob, unsigned int poll_flags);
-int nvme_ns_head_chr_uring_cmd_iopoll(struct io_uring_cmd *ioucmd,
-		struct io_comp_batch *iob, unsigned int poll_flags);
 int nvme_ns_chr_uring_cmd(struct io_uring_cmd *ioucmd,
 		unsigned int issue_flags);
 int nvme_ns_head_chr_uring_cmd(struct io_uring_cmd *ioucmd,

drivers/nvme/host/pci.c

@@ -2690,7 +2690,8 @@ static void nvme_reset_work(struct work_struct *work)
 	if (dev->ctrl.state != NVME_CTRL_RESETTING) {
 		dev_warn(dev->ctrl.device, "ctrl state %d is not RESETTING\n",
 			 dev->ctrl.state);
-		return;
+		result = -ENODEV;
+		goto out;
 	}
 
 	/*
@@ -2777,7 +2778,9 @@ static void nvme_reset_work(struct work_struct *work)
 		 result);
 	nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_DELETING);
 	nvme_dev_disable(dev, true);
+	nvme_sync_queues(&dev->ctrl);
 	nvme_mark_namespaces_dead(&dev->ctrl);
+	nvme_unquiesce_io_queues(&dev->ctrl);
 	nvme_change_ctrl_state(&dev->ctrl, NVME_CTRL_DEAD);
 }
 

drivers/nvme/target/nvmet.h

@@ -79,8 +79,8 @@ struct nvmet_ns {
 	struct completion	disable_done;
 	mempool_t		*bvec_pool;
 
-	int			use_p2pmem;
 	struct pci_dev		*p2p_dev;
+	int			use_p2pmem;
 	int			pi_type;
 	int			metadata_size;
 	u8			csi;

include/linux/blk-mq.h

@@ -715,6 +715,8 @@ int blk_mq_alloc_sq_tag_set(struct blk_mq_tag_set *set,
 void blk_mq_free_tag_set(struct blk_mq_tag_set *set);
 
 void blk_mq_free_request(struct request *rq);
+int blk_rq_poll(struct request *rq, struct io_comp_batch *iob,
+		unsigned int poll_flags);
 
 bool blk_mq_queue_inflight(struct request_queue *q);
 
@@ -852,7 +854,11 @@ 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;
 

include/uapi/linux/io_uring.h

@@ -244,8 +244,10 @@ enum io_uring_op {
  * sqe->uring_cmd_flags
  * IORING_URING_CMD_FIXED	use registered buffer; pass this flag
  *				along with setting sqe->buf_index.
+ * IORING_URING_CMD_POLLED	driver use only
  */
 #define IORING_URING_CMD_FIXED	(1U << 0)
+#define IORING_URING_CMD_POLLED	(1U << 31)
 
 
 /*