Merge branch 'md-next' of...

Merge branch 'md-next' of git://git.kernel.org/pub/scm/linux/kernel/git/song/md into for-5.8/drivers

Pull MD changes from Song.

* 'md-next' of git://git.kernel.org/pub/scm/linux/kernel/git/song/md:
  md/raid1: Replace zero-length array with flexible-array
  md: add a newline when printing parameter 'start_ro' by sysfs
  md: stop using ->queuedata
  md/raid1: release pending accounting for an I/O only after write-behind is also finished
  md: remove redundant memalloc scope API usage
  raid5: update code comment of scribble_alloc()
  raid5: remove gfp flags from scribble_alloc()
  md: use memalloc scope APIs in mddev_suspend()/mddev_resume()
  md: remove the extra line for ->hot_add_disk
  md: flush md_rdev_misc_wq for HOT_ADD_DISK case
  md: don't flush workqueue unconditionally in md_open
  md: add new workqueue for delete rdev
  md: add checkings before flush md_misc_wq

drivers/md/md-linear.h

@@ -12,6 +12,6 @@ struct linear_conf
 	struct rcu_head		rcu;
 	sector_t		array_sectors;
 	int			raid_disks; /* a copy of mddev->raid_disks */
-	struct dev_info		disks[0];
+	struct dev_info		disks[];
 };
 #endif

drivers/md/md.c

@@ -89,6 +89,7 @@ static struct module *md_cluster_mod;
 static DECLARE_WAIT_QUEUE_HEAD(resync_wait);
 static struct workqueue_struct *md_wq;
 static struct workqueue_struct *md_misc_wq;
+static struct workqueue_struct *md_rdev_misc_wq;
 
 static int remove_and_add_spares(struct mddev *mddev,
 				 struct md_rdev *this);
@@ -227,13 +228,13 @@ void mddev_create_serial_pool(struct mddev *mddev, struct md_rdev *rdev,
 		goto abort;
 
 	if (mddev->serial_info_pool == NULL) {
-		unsigned int noio_flag;
-
-		noio_flag = memalloc_noio_save();
+		/*
+		 * already in memalloc noio context by
+		 * mddev_suspend()
+		 */
 		mddev->serial_info_pool =
 			mempool_create_kmalloc_pool(NR_SERIAL_INFOS,
 						sizeof(struct serial_info));
-		memalloc_noio_restore(noio_flag);
 		if (!mddev->serial_info_pool) {
 			rdevs_uninit_serial(mddev);
 			pr_err("can't alloc memory pool for serialization\n");
@@ -466,7 +467,7 @@ static blk_qc_t md_make_request(struct request_queue *q, struct bio *bio)
 {
 	const int rw = bio_data_dir(bio);
 	const int sgrp = op_stat_group(bio_op(bio));
-	struct mddev *mddev = q->queuedata;
+	struct mddev *mddev = bio->bi_disk->private_data;
 	unsigned int sectors;
 
 	if (unlikely(test_bit(MD_BROKEN, &mddev->flags)) && (rw == WRITE)) {
@@ -527,11 +528,15 @@ void mddev_suspend(struct mddev *mddev)
 	wait_event(mddev->sb_wait, !test_bit(MD_UPDATING_SB, &mddev->flags));
 
 	del_timer_sync(&mddev->safemode_timer);
+	/* restrict memory reclaim I/O during raid array is suspend */
+	mddev->noio_flag = memalloc_noio_save();
 }
 EXPORT_SYMBOL_GPL(mddev_suspend);
 
 void mddev_resume(struct mddev *mddev)
 {
+	/* entred the memalloc scope from mddev_suspend() */
+	memalloc_noio_restore(mddev->noio_flag);
 	lockdep_assert_held(&mddev->reconfig_mutex);
 	if (--mddev->suspended)
 		return;
@@ -2454,7 +2459,7 @@ static int bind_rdev_to_array(struct md_rdev *rdev, struct mddev *mddev)
 	return err;
 }
 
-static void md_delayed_delete(struct work_struct *ws)
+static void rdev_delayed_delete(struct work_struct *ws)
 {
 	struct md_rdev *rdev = container_of(ws, struct md_rdev, del_work);
 	kobject_del(&rdev->kobj);
@@ -2479,9 +2484,9 @@ static void unbind_rdev_from_array(struct md_rdev *rdev)
 	 * to delay it due to rcu usage.
 	 */
 	synchronize_rcu();
-	INIT_WORK(&rdev->del_work, md_delayed_delete);
+	INIT_WORK(&rdev->del_work, rdev_delayed_delete);
 	kobject_get(&rdev->kobj);
-	queue_work(md_misc_wq, &rdev->del_work);
+	queue_work(md_rdev_misc_wq, &rdev->del_work);
 }
 
 /*
@@ -3191,8 +3196,7 @@ slot_store(struct md_rdev *rdev, const char *buf, size_t len)
 			rdev->saved_raid_disk = -1;
 		clear_bit(In_sync, &rdev->flags);
 		clear_bit(Bitmap_sync, &rdev->flags);
-		err = rdev->mddev->pers->
-			hot_add_disk(rdev->mddev, rdev);
+		err = rdev->mddev->pers->hot_add_disk(rdev->mddev, rdev);
 		if (err) {
 			rdev->raid_disk = -1;
 			return err;
@@ -4514,6 +4518,20 @@ null_show(struct mddev *mddev, char *page)
 	return -EINVAL;
 }
 
+/* need to ensure rdev_delayed_delete() has completed */
+static void flush_rdev_wq(struct mddev *mddev)
+{
+	struct md_rdev *rdev;
+
+	rcu_read_lock();
+	rdev_for_each_rcu(rdev, mddev)
+		if (work_pending(&rdev->del_work)) {
+			flush_workqueue(md_rdev_misc_wq);
+			break;
+		}
+	rcu_read_unlock();
+}
+
 static ssize_t
 new_dev_store(struct mddev *mddev, const char *buf, size_t len)
 {
@@ -4541,8 +4559,7 @@ new_dev_store(struct mddev *mddev, const char *buf, size_t len)
 	    minor != MINOR(dev))
 		return -EOVERFLOW;
 
-	flush_workqueue(md_misc_wq);
-
+	flush_rdev_wq(mddev);
 	err = mddev_lock(mddev);
 	if (err)
 		return err;
@@ -4780,7 +4797,8 @@ action_store(struct mddev *mddev, const char *page, size_t len)
 			clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
 		if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery) &&
 		    mddev_lock(mddev) == 0) {
-			flush_workqueue(md_misc_wq);
+			if (work_pending(&mddev->del_work))
+				flush_workqueue(md_misc_wq);
 			if (mddev->sync_thread) {
 				set_bit(MD_RECOVERY_INTR, &mddev->recovery);
 				md_reap_sync_thread(mddev);
@@ -5626,7 +5644,6 @@ static int md_alloc(dev_t dev, char *name)
 	mddev->queue = blk_alloc_queue(md_make_request, NUMA_NO_NODE);
 	if (!mddev->queue)
 		goto abort;
-	mddev->queue->queuedata = mddev;
 
 	blk_set_stacking_limits(&mddev->queue->limits);
 
@@ -6147,7 +6164,8 @@ static void md_clean(struct mddev *mddev)
 static void __md_stop_writes(struct mddev *mddev)
 {
 	set_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
-	flush_workqueue(md_misc_wq);
+	if (work_pending(&mddev->del_work))
+		flush_workqueue(md_misc_wq);
 	if (mddev->sync_thread) {
 		set_bit(MD_RECOVERY_INTR, &mddev->recovery);
 		md_reap_sync_thread(mddev);
@@ -6200,7 +6218,8 @@ static void __md_stop(struct mddev *mddev)
 	md_bitmap_destroy(mddev);
 	mddev_detach(mddev);
 	/* Ensure ->event_work is done */
-	flush_workqueue(md_misc_wq);
+	if (mddev->event_work.func)
+		flush_workqueue(md_misc_wq);
 	spin_lock(&mddev->lock);
 	mddev->pers = NULL;
 	spin_unlock(&mddev->lock);
@@ -7495,9 +7514,8 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 
 	}
 
-	if (cmd == ADD_NEW_DISK)
-		/* need to ensure md_delayed_delete() has completed */
-		flush_workqueue(md_misc_wq);
+	if (cmd == ADD_NEW_DISK || cmd == HOT_ADD_DISK)
+		flush_rdev_wq(mddev);
 
 	if (cmd == HOT_REMOVE_DISK)
 		/* need to ensure recovery thread has run */
@@ -7752,7 +7770,8 @@ static int md_open(struct block_device *bdev, fmode_t mode)
 		 */
 		mddev_put(mddev);
 		/* Wait until bdev->bd_disk is definitely gone */
-		flush_workqueue(md_misc_wq);
+		if (work_pending(&mddev->del_work))
+			flush_workqueue(md_misc_wq);
 		/* Then retry the open from the top */
 		return -ERESTARTSYS;
 	}
@@ -9040,8 +9059,7 @@ static int remove_and_add_spares(struct mddev *mddev,
 
 			rdev->recovery_offset = 0;
 		}
-		if (mddev->pers->
-		    hot_add_disk(mddev, rdev) == 0) {
+		if (mddev->pers->hot_add_disk(mddev, rdev) == 0) {
 			if (sysfs_link_rdev(mddev, rdev))
 				/* failure here is OK */;
 			if (!test_bit(Journal, &rdev->flags))
@@ -9469,6 +9487,10 @@ static int __init md_init(void)
 	if (!md_misc_wq)
 		goto err_misc_wq;
 
+	md_rdev_misc_wq = alloc_workqueue("md_rdev_misc", 0, 0);
+	if (!md_misc_wq)
+		goto err_rdev_misc_wq;
+
 	if ((ret = register_blkdev(MD_MAJOR, "md")) < 0)
 		goto err_md;
 
@@ -9490,6 +9512,8 @@ static int __init md_init(void)
 err_mdp:
 	unregister_blkdev(MD_MAJOR, "md");
 err_md:
+	destroy_workqueue(md_rdev_misc_wq);
+err_rdev_misc_wq:
 	destroy_workqueue(md_misc_wq);
 err_misc_wq:
 	destroy_workqueue(md_wq);
@@ -9776,6 +9800,7 @@ static __exit void md_exit(void)
 		 * destroy_workqueue() below will wait for that to complete.
 		 */
 	}
+	destroy_workqueue(md_rdev_misc_wq);
 	destroy_workqueue(md_misc_wq);
 	destroy_workqueue(md_wq);
 }
@@ -9785,7 +9810,7 @@ module_exit(md_exit)
 
 static int get_ro(char *buffer, const struct kernel_param *kp)
 {
-	return sprintf(buffer, "%d", start_readonly);
+	return sprintf(buffer, "%d\n", start_readonly);
 }
 static int set_ro(const char *val, const struct kernel_param *kp)
 {

drivers/md/md.h

@@ -497,6 +497,7 @@ struct mddev {
 	void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev);
 	struct md_cluster_info		*cluster_info;
 	unsigned int			good_device_nr;	/* good device num within cluster raid */
+	unsigned int			noio_flag; /* for memalloc scope API */
 
 	bool	has_superblocks:1;
 	bool	fail_last_dev:1;

drivers/md/raid1.c

@@ -296,22 +296,17 @@ static void reschedule_retry(struct r1bio *r1_bio)
 static void call_bio_endio(struct r1bio *r1_bio)
 {
 	struct bio *bio = r1_bio->master_bio;
-	struct r1conf *conf = r1_bio->mddev->private;
 
 	if (!test_bit(R1BIO_Uptodate, &r1_bio->state))
 		bio->bi_status = BLK_STS_IOERR;
 
 	bio_endio(bio);
-	/*
-	 * Wake up any possible resync thread that waits for the device
-	 * to go idle.
-	 */
-	allow_barrier(conf, r1_bio->sector);
 }
 
 static void raid_end_bio_io(struct r1bio *r1_bio)
 {
 	struct bio *bio = r1_bio->master_bio;
+	struct r1conf *conf = r1_bio->mddev->private;
 
 	/* if nobody has done the final endio yet, do it now */
 	if (!test_and_set_bit(R1BIO_Returned, &r1_bio->state)) {
@@ -322,6 +317,12 @@ static void raid_end_bio_io(struct r1bio *r1_bio)
 
 		call_bio_endio(r1_bio);
 	}
+	/*
+	 * Wake up any possible resync thread that waits for the device
+	 * to go idle.  All I/Os, even write-behind writes, are done.
+	 */
+	allow_barrier(conf, r1_bio->sector);
+
 	free_r1bio(r1_bio);
 }
 

drivers/md/raid1.h

@@ -180,7 +180,7 @@ struct r1bio {
 	 * if the IO is in WRITE direction, then multiple bios are used.
 	 * We choose the number when they are allocated.
 	 */
-	struct bio		*bios[0];
+	struct bio		*bios[];
 	/* DO NOT PUT ANY NEW FIELDS HERE - bios array is contiguously alloced*/
 };
 

drivers/md/raid10.h

@@ -153,7 +153,7 @@ struct r10bio {
 		};
 		sector_t	addr;
 		int		devnum;
-	} devs[0];
+	} devs[];
 };
 
 /* bits for r10bio.state */

drivers/md/raid5.c

@@ -2215,10 +2215,13 @@ static int grow_stripes(struct r5conf *conf, int num)
 }
 
 /**
- * scribble_len - return the required size of the scribble region
+ * scribble_alloc - allocate percpu scribble buffer for required size
+ *		    of the scribble region
+ * @percpu - from for_each_present_cpu() of the caller
  * @num - total number of disks in the array
+ * @cnt - scribble objs count for required size of the scribble region
  *
- * The size must be enough to contain:
+ * The scribble buffer size must be enough to contain:
  * 1/ a struct page pointer for each device in the array +2
  * 2/ room to convert each entry in (1) to its corresponding dma
  *    (dma_map_page()) or page (page_address()) address.
@@ -2228,14 +2231,19 @@ static int grow_stripes(struct r5conf *conf, int num)
  * of the P and Q blocks.
  */
 static int scribble_alloc(struct raid5_percpu *percpu,
-			  int num, int cnt, gfp_t flags)
+			  int num, int cnt)
 {
 	size_t obj_size =
 		sizeof(struct page *) * (num+2) +
 		sizeof(addr_conv_t) * (num+2);
 	void *scribble;
 
-	scribble = kvmalloc_array(cnt, obj_size, flags);
+	/*
+	 * If here is in raid array suspend context, it is in memalloc noio
+	 * context as well, there is no potential recursive memory reclaim
+	 * I/Os with the GFP_KERNEL flag.
+	 */
+	scribble = kvmalloc_array(cnt, obj_size, GFP_KERNEL);
 	if (!scribble)
 		return -ENOMEM;
 
@@ -2267,8 +2275,7 @@ static int resize_chunks(struct r5conf *conf, int new_disks, int new_sectors)
 
 		percpu = per_cpu_ptr(conf->percpu, cpu);
 		err = scribble_alloc(percpu, new_disks,
-				     new_sectors / STRIPE_SECTORS,
-				     GFP_NOIO);
+				     new_sectors / STRIPE_SECTORS);
 		if (err)
 			break;
 	}
@@ -6759,8 +6766,7 @@ static int alloc_scratch_buffer(struct r5conf *conf, struct raid5_percpu *percpu
 			       conf->previous_raid_disks),
 			   max(conf->chunk_sectors,
 			       conf->prev_chunk_sectors)
-			   / STRIPE_SECTORS,
-			   GFP_KERNEL)) {
+			   / STRIPE_SECTORS)) {
 		free_scratch_buffer(conf, percpu);
 		return -ENOMEM;
 	}