Documentation/workqueue.txt:
	If there is dependency among multiple work items used
	during memory reclaim, they should be queued to separate
	wq each with WQ_MEM_RECLAIM.

Loop devices can be stacked, so we have to convert to per-device
workqueue. One example is Fedora live CD.

Fixes: b5dd2f6047ca108001328aac0e8588edd15f1778
Cc: stable@vger.kernel.org (v4.0)
Cc: Justin M. Forbes <jforbes@fedoraproject.org>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
---
 drivers/block/loop.c | 30 ++++++++++++++----------------
 drivers/block/loop.h |  1 +
 2 files changed, 15 insertions(+), 16 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index ae3fcb4..3dc1598 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -86,8 +86,6 @@ static DEFINE_MUTEX(loop_index_mutex);
 static int max_part;
 static int part_shift;
 
-static struct workqueue_struct *loop_wq;
-
 /*
  * Transfer functions
  */
@@ -725,6 +723,12 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
 	size = get_loop_size(lo, file);
 	if ((loff_t)(sector_t)size != size)
 		goto out_putf;
+	error = -ENOMEM;
+	lo->wq = alloc_workqueue("kloopd%d",
+			WQ_MEM_RECLAIM | WQ_HIGHPRI | WQ_UNBOUND, 0,
+			lo->lo_number);
+	if (!lo->wq)
+		goto out_putf;
 
 	error = 0;
 
@@ -872,6 +876,8 @@ static int loop_clr_fd(struct loop_device *lo)
 	lo->lo_flags = 0;
 	if (!part_shift)
 		lo->lo_disk->flags |= GENHD_FL_NO_PART_SCAN;
+	destroy_workqueue(lo->wq);
+	lo->wq = NULL;
 	mutex_unlock(&lo->lo_ctl_mutex);
 	/*
 	 * Need not hold lo_ctl_mutex to fput backing file.
@@ -1425,9 +1431,13 @@ static int loop_queue_rq(struct blk_mq_hw_ctx *hctx,
 		const struct blk_mq_queue_data *bd)
 {
 	struct loop_cmd *cmd = blk_mq_rq_to_pdu(bd->rq);
+	struct loop_device *lo = cmd->rq->q->queuedata;
 
 	blk_mq_start_request(bd->rq);
 
+	if (lo->lo_state != Lo_bound)
+		return -EIO;
+
 	if (cmd->rq->cmd_flags & REQ_WRITE) {
 		struct loop_device *lo = cmd->rq->q->queuedata;
 		bool need_sched = true;
@@ -1441,9 +1451,9 @@ static int loop_queue_rq(struct blk_mq_hw_ctx *hctx,
 		spin_unlock_irq(&lo->lo_lock);
 
 		if (need_sched)
-			queue_work(loop_wq, &lo->write_work);
+			queue_work(lo->wq, &lo->write_work);
 	} else {
-		queue_work(loop_wq, &cmd->read_work);
+		queue_work(lo->wq, &cmd->read_work);
 	}
 
 	return BLK_MQ_RQ_QUEUE_OK;
@@ -1455,9 +1465,6 @@ static void loop_handle_cmd(struct loop_cmd *cmd)
 	struct loop_device *lo = cmd->rq->q->queuedata;
 	int ret = -EIO;
 
-	if (lo->lo_state != Lo_bound)
-		goto failed;
-
 	if (write && (lo->lo_flags & LO_FLAGS_READ_ONLY))
 		goto failed;
 
@@ -1806,13 +1813,6 @@ static int __init loop_init(void)
 		goto misc_out;
 	}
 
-	loop_wq = alloc_workqueue("kloopd",
-			WQ_MEM_RECLAIM | WQ_HIGHPRI | WQ_UNBOUND, 0);
-	if (!loop_wq) {
-		err = -ENOMEM;
-		goto misc_out;
-	}
-
 	blk_register_region(MKDEV(LOOP_MAJOR, 0), range,
 				  THIS_MODULE, loop_probe, NULL, NULL);
 
@@ -1850,8 +1850,6 @@ static void __exit loop_exit(void)
 	blk_unregister_region(MKDEV(LOOP_MAJOR, 0), range);
 	unregister_blkdev(LOOP_MAJOR, "loop");
 
-	destroy_workqueue(loop_wq);
-
 	misc_deregister(&loop_misc);
 }
 
diff --git a/drivers/block/loop.h b/drivers/block/loop.h
index 301c27f..49564ed 100644
--- a/drivers/block/loop.h
+++ b/drivers/block/loop.h
@@ -54,6 +54,7 @@ struct loop_device {
 	gfp_t		old_gfp_mask;
 
 	spinlock_t		lo_lock;
+	struct workqueue_struct *wq;
 	struct list_head	write_cmd_head;
 	struct work_struct	write_work;
 	bool			write_started;
-- 
1.9.1
If there are too many pending per work I/O, too many
high priority work thread can be generated so that
system performance can be effected.

This patch limits the max_active parameter of workqueue as 16.

This patch fixes Fedora 22 live booting performance
regression when it is booted from squashfs over dm
based on loop, and looks the following reasons are
related with the problem:

- not like other filesyststems(such as ext4), squashfs
is a bit special, and I observed that increasing I/O jobs
to access file in squashfs only improve I/O performance a
little, but it can make big difference for ext4

- nested loop: both squashfs.img and ext3fs.img are mounted
as loop block, and ext3fs.img is inside the squashfs

- during booting, lots of tasks may run concurrently

Fixes: b5dd2f6047ca108001328aac0e8588edd15f1778
Cc: stable@vger.kernel.org (v4.0)
Cc: Justin M. Forbes <jforbes@fedoraproject.org>
Signed-off-by: Ming Lei <ming.lei@canonical.com>
---
 drivers/block/loop.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 3dc1598..1bee523 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -725,7 +725,7 @@ static int loop_set_fd(struct loop_device *lo, fmode_t mode,
 		goto out_putf;
 	error = -ENOMEM;
 	lo->wq = alloc_workqueue("kloopd%d",
-			WQ_MEM_RECLAIM | WQ_HIGHPRI | WQ_UNBOUND, 0,
+			WQ_MEM_RECLAIM | WQ_HIGHPRI | WQ_UNBOUND, 16,
 			lo->lo_number);
 	if (!lo->wq)
 		goto out_putf;
-- 
1.9.1