Merge tag 'block-5.17-2022-01-21' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:
 "Various little minor fixes that should go into this release:

   - Fix issue with cloned bios and IO accounting (Christoph)

   - Remove redundant assignments (Colin, GuoYong)

   - Fix an issue with the mq-deadline async_depth sysfs interface (me)

   - Fix brd module loading race (Tetsuo)

   - Shared tag map wakeup fix (Laibin)

   - End of bdev read fix (OGAWA)

   - srcu leak fix (Ming)"

* tag 'block-5.17-2022-01-21' of git://git.kernel.dk/linux-block:
  block: fix async_depth sysfs interface for mq-deadline
  block: Fix wrong offset in bio_truncate()
  block: assign bi_bdev for cloned bios in blk_rq_prep_clone
  block: cleanup q->srcu
  block: Remove unnecessary variable assignment
  brd: remove brd_devices_mutex mutex
  aoe: remove redundant assignment on variable n
  loop: remove redundant initialization of pointer node
  blk-mq: fix tag_get wait task can't be awakened

block/bio.c

@@ -569,7 +569,8 @@ static void bio_truncate(struct bio *bio, unsigned new_size)
 				offset = new_size - done;
 			else
 				offset = 0;
-			zero_user(bv.bv_page, offset, bv.bv_len - offset);
+			zero_user(bv.bv_page, bv.bv_offset + offset,
+				  bv.bv_len - offset);
 			truncated = true;
 		}
 		done += bv.bv_len;

block/blk-mq-tag.c

@@ -16,6 +16,21 @@
 #include "blk-mq-sched.h"
 #include "blk-mq-tag.h"
 
+/*
+ * Recalculate wakeup batch when tag is shared by hctx.
+ */
+static void blk_mq_update_wake_batch(struct blk_mq_tags *tags,
+		unsigned int users)
+{
+	if (!users)
+		return;
+
+	sbitmap_queue_recalculate_wake_batch(&tags->bitmap_tags,
+			users);
+	sbitmap_queue_recalculate_wake_batch(&tags->breserved_tags,
+			users);
+}
+
 /*
  * If a previously inactive queue goes active, bump the active user count.
  * We need to do this before try to allocate driver tag, then even if fail
@@ -24,18 +39,26 @@
  */
 bool __blk_mq_tag_busy(struct blk_mq_hw_ctx *hctx)
 {
+	unsigned int users;
+
 	if (blk_mq_is_shared_tags(hctx->flags)) {
 		struct request_queue *q = hctx->queue;
 
-		if (!test_bit(QUEUE_FLAG_HCTX_ACTIVE, &q->queue_flags) &&
-		    !test_and_set_bit(QUEUE_FLAG_HCTX_ACTIVE, &q->queue_flags))
-			atomic_inc(&hctx->tags->active_queues);
+		if (test_bit(QUEUE_FLAG_HCTX_ACTIVE, &q->queue_flags) ||
+		    test_and_set_bit(QUEUE_FLAG_HCTX_ACTIVE, &q->queue_flags)) {
+			return true;
+		}
 	} else {
-		if (!test_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state) &&
-		    !test_and_set_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state))
-			atomic_inc(&hctx->tags->active_queues);
+		if (test_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state) ||
+		    test_and_set_bit(BLK_MQ_S_TAG_ACTIVE, &hctx->state)) {
+			return true;
+		}
 	}
 
+	users = atomic_inc_return(&hctx->tags->active_queues);
+
+	blk_mq_update_wake_batch(hctx->tags, users);
+
 	return true;
 }
 
@@ -56,6 +79,7 @@ void blk_mq_tag_wakeup_all(struct blk_mq_tags *tags, bool include_reserve)
 void __blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx)
 {
 	struct blk_mq_tags *tags = hctx->tags;
+	unsigned int users;
 
 	if (blk_mq_is_shared_tags(hctx->flags)) {
 		struct request_queue *q = hctx->queue;
@@ -68,7 +92,9 @@ void __blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx)
 			return;
 	}
 
-	atomic_dec(&tags->active_queues);
+	users = atomic_dec_return(&tags->active_queues);
+
+	blk_mq_update_wake_batch(tags, users);
 
 	blk_mq_tag_wakeup_all(tags, false);
 }

block/blk-mq.c

@@ -2976,6 +2976,7 @@ int blk_rq_prep_clone(struct request *rq, struct request *rq_src,
 		bio = bio_clone_fast(bio_src, gfp_mask, bs);
 		if (!bio)
 			goto free_and_out;
+		bio->bi_bdev = rq->q->disk->part0;
 
 		if (bio_ctr && bio_ctr(bio, bio_src, data))
 			goto free_and_out;

block/blk-sysfs.c

@@ -811,6 +811,9 @@ static void blk_release_queue(struct kobject *kobj)
 
 	bioset_exit(&q->bio_split);
 
+	if (blk_queue_has_srcu(q))
+		cleanup_srcu_struct(q->srcu);
+
 	ida_simple_remove(&blk_queue_ida, q->id);
 	call_rcu(&q->rcu_head, blk_free_queue_rcu);
 }
@@ -887,7 +890,6 @@ int blk_register_queue(struct gendisk *disk)
 		kobject_uevent(&q->elevator->kobj, KOBJ_ADD);
 	mutex_unlock(&q->sysfs_lock);
 
-	ret = 0;
 unlock:
 	mutex_unlock(&q->sysfs_dir_lock);
 

block/mq-deadline.c

@@ -865,7 +865,7 @@ SHOW_JIFFIES(deadline_write_expire_show, dd->fifo_expire[DD_WRITE]);
 SHOW_JIFFIES(deadline_prio_aging_expire_show, dd->prio_aging_expire);
 SHOW_INT(deadline_writes_starved_show, dd->writes_starved);
 SHOW_INT(deadline_front_merges_show, dd->front_merges);
-SHOW_INT(deadline_async_depth_show, dd->front_merges);
+SHOW_INT(deadline_async_depth_show, dd->async_depth);
 SHOW_INT(deadline_fifo_batch_show, dd->fifo_batch);
 #undef SHOW_INT
 #undef SHOW_JIFFIES
@@ -895,7 +895,7 @@ STORE_JIFFIES(deadline_write_expire_store, &dd->fifo_expire[DD_WRITE], 0, INT_MA
 STORE_JIFFIES(deadline_prio_aging_expire_store, &dd->prio_aging_expire, 0, INT_MAX);
 STORE_INT(deadline_writes_starved_store, &dd->writes_starved, INT_MIN, INT_MAX);
 STORE_INT(deadline_front_merges_store, &dd->front_merges, 0, 1);
-STORE_INT(deadline_async_depth_store, &dd->front_merges, 1, INT_MAX);
+STORE_INT(deadline_async_depth_store, &dd->async_depth, 1, INT_MAX);
 STORE_INT(deadline_fifo_batch_store, &dd->fifo_batch, 0, INT_MAX);
 #undef STORE_FUNCTION
 #undef STORE_INT

drivers/block/aoe/aoecmd.c

@@ -122,7 +122,7 @@ newtag(struct aoedev *d)
 	register ulong n;
 
 	n = jiffies & 0xffff;
-	return n |= (++d->lasttag & 0x7fff) << 16;
+	return n | (++d->lasttag & 0x7fff) << 16;
 }
 
 static u32

drivers/block/brd.c

@@ -362,7 +362,6 @@ __setup("ramdisk_size=", ramdisk_size);
  * (should share code eventually).
  */
 static LIST_HEAD(brd_devices);
-static DEFINE_MUTEX(brd_devices_mutex);
 static struct dentry *brd_debugfs_dir;
 
 static int brd_alloc(int i)
@@ -372,21 +371,14 @@ static int brd_alloc(int i)
 	char buf[DISK_NAME_LEN];
 	int err = -ENOMEM;
 
-	mutex_lock(&brd_devices_mutex);
-	list_for_each_entry(brd, &brd_devices, brd_list) {
-		if (brd->brd_number == i) {
-			mutex_unlock(&brd_devices_mutex);
+	list_for_each_entry(brd, &brd_devices, brd_list)
+		if (brd->brd_number == i)
 			return -EEXIST;
-		}
-	}
 	brd = kzalloc(sizeof(*brd), GFP_KERNEL);
-	if (!brd) {
-		mutex_unlock(&brd_devices_mutex);
+	if (!brd)
 		return -ENOMEM;
-	}
 	brd->brd_number		= i;
 	list_add_tail(&brd->brd_list, &brd_devices);
-	mutex_unlock(&brd_devices_mutex);
 
 	spin_lock_init(&brd->brd_lock);
 	INIT_RADIX_TREE(&brd->brd_pages, GFP_ATOMIC);
@@ -429,9 +421,7 @@ static int brd_alloc(int i)
 out_cleanup_disk:
 	blk_cleanup_disk(disk);
 out_free_dev:
-	mutex_lock(&brd_devices_mutex);
 	list_del(&brd->brd_list);
-	mutex_unlock(&brd_devices_mutex);
 	kfree(brd);
 	return err;
 }
@@ -441,15 +431,19 @@ static void brd_probe(dev_t dev)
 	brd_alloc(MINOR(dev) / max_part);
 }
 
-static void brd_del_one(struct brd_device *brd)
+static void brd_cleanup(void)
 {
-	del_gendisk(brd->brd_disk);
-	blk_cleanup_disk(brd->brd_disk);
-	brd_free_pages(brd);
-	mutex_lock(&brd_devices_mutex);
-	list_del(&brd->brd_list);
-	mutex_unlock(&brd_devices_mutex);
-	kfree(brd);
+	struct brd_device *brd, *next;
+
+	debugfs_remove_recursive(brd_debugfs_dir);
+
+	list_for_each_entry_safe(brd, next, &brd_devices, brd_list) {
+		del_gendisk(brd->brd_disk);
+		blk_cleanup_disk(brd->brd_disk);
+		brd_free_pages(brd);
+		list_del(&brd->brd_list);
+		kfree(brd);
+	}
 }
 
 static inline void brd_check_and_reset_par(void)
@@ -473,9 +467,18 @@ static inline void brd_check_and_reset_par(void)
 
 static int __init brd_init(void)
 {
-	struct brd_device *brd, *next;
 	int err, i;
 
+	brd_check_and_reset_par();
+
+	brd_debugfs_dir = debugfs_create_dir("ramdisk_pages", NULL);
+
+	for (i = 0; i < rd_nr; i++) {
+		err = brd_alloc(i);
+		if (err)
+			goto out_free;
+	}
+
 	/*
 	 * brd module now has a feature to instantiate underlying device
 	 * structure on-demand, provided that there is an access dev node.
@@ -491,42 +494,26 @@ static int __init brd_init(void)
 	 *	dynamically.
 	 */
 
-	if (__register_blkdev(RAMDISK_MAJOR, "ramdisk", brd_probe))
-		return -EIO;
-
-	brd_check_and_reset_par();
-
-	brd_debugfs_dir = debugfs_create_dir("ramdisk_pages", NULL);
-
-	for (i = 0; i < rd_nr; i++) {
-		err = brd_alloc(i);
-		if (err)
-			goto out_free;
+	if (__register_blkdev(RAMDISK_MAJOR, "ramdisk", brd_probe)) {
+		err = -EIO;
+		goto out_free;
 	}
 
 	pr_info("brd: module loaded\n");
 	return 0;
 
 out_free:
-	unregister_blkdev(RAMDISK_MAJOR, "ramdisk");
-	debugfs_remove_recursive(brd_debugfs_dir);
-
-	list_for_each_entry_safe(brd, next, &brd_devices, brd_list)
-		brd_del_one(brd);
+	brd_cleanup();
 
 	pr_info("brd: module NOT loaded !!!\n");
 	return err;
 }
 
 static void __exit brd_exit(void)
 {
-	struct brd_device *brd, *next;
 
 	unregister_blkdev(RAMDISK_MAJOR, "ramdisk");
-	debugfs_remove_recursive(brd_debugfs_dir);
-
-	list_for_each_entry_safe(brd, next, &brd_devices, brd_list)
-		brd_del_one(brd);
+	brd_cleanup();
 
 	pr_info("brd: module unloaded\n");
 }

drivers/block/loop.c

@@ -820,7 +820,7 @@ static inline int queue_on_root_worker(struct cgroup_subsys_state *css)
 
 static void loop_queue_work(struct loop_device *lo, struct loop_cmd *cmd)
 {
-	struct rb_node **node = &(lo->worker_tree.rb_node), *parent = NULL;
+	struct rb_node **node, *parent = NULL;
 	struct loop_worker *cur_worker, *worker = NULL;
 	struct work_struct *work;
 	struct list_head *cmd_list;

include/linux/sbitmap.h

@@ -415,6 +415,17 @@ static inline void sbitmap_queue_free(struct sbitmap_queue *sbq)
 	sbitmap_free(&sbq->sb);
 }
 
+/**
+ * sbitmap_queue_recalculate_wake_batch() - Recalculate wake batch
+ * @sbq: Bitmap queue to recalculate wake batch.
+ * @users: Number of shares.
+ *
+ * Like sbitmap_queue_update_wake_batch(), this will calculate wake batch
+ * by depth. This interface is for HCTX shared tags or queue shared tags.
+ */
+void sbitmap_queue_recalculate_wake_batch(struct sbitmap_queue *sbq,
+					    unsigned int users);
+
 /**
  * sbitmap_queue_resize() - Resize a &struct sbitmap_queue.
  * @sbq: Bitmap queue to resize.

lib/sbitmap.c

@@ -457,10 +457,9 @@ int sbitmap_queue_init_node(struct sbitmap_queue *sbq, unsigned int depth,
 }
 EXPORT_SYMBOL_GPL(sbitmap_queue_init_node);
 
-static void sbitmap_queue_update_wake_batch(struct sbitmap_queue *sbq,
-					    unsigned int depth)
+static inline void __sbitmap_queue_update_wake_batch(struct sbitmap_queue *sbq,
+					    unsigned int wake_batch)
 {
-	unsigned int wake_batch = sbq_calc_wake_batch(sbq, depth);
 	int i;
 
 	if (sbq->wake_batch != wake_batch) {
@@ -476,6 +475,26 @@ static void sbitmap_queue_update_wake_batch(struct sbitmap_queue *sbq,
 	}
 }
 
+static void sbitmap_queue_update_wake_batch(struct sbitmap_queue *sbq,
+					    unsigned int depth)
+{
+	unsigned int wake_batch;
+
+	wake_batch = sbq_calc_wake_batch(sbq, depth);
+	__sbitmap_queue_update_wake_batch(sbq, wake_batch);
+}
+
+void sbitmap_queue_recalculate_wake_batch(struct sbitmap_queue *sbq,
+					    unsigned int users)
+{
+	unsigned int wake_batch;
+
+	wake_batch = clamp_val((sbq->sb.depth + users - 1) /
+			users, 4, SBQ_WAKE_BATCH);
+	__sbitmap_queue_update_wake_batch(sbq, wake_batch);
+}
+EXPORT_SYMBOL_GPL(sbitmap_queue_recalculate_wake_batch);
+
 void sbitmap_queue_resize(struct sbitmap_queue *sbq, unsigned int depth)
 {
 	sbitmap_queue_update_wake_batch(sbq, depth);