Merge branch 'for-linus' of git://neil.brown.name/md

* 'for-linus' of git://neil.brown.name/md:
  md: tidy up device searches in read_balance.
  md/raid1: fix some typos in comments.
  md/raid1: discard unused variable.
  md: unplug writes to external bitmaps.
  md: use separate bio pool for each md device.
  md: change type of first arg to sync_page_io.
  md/raid1: perform mem allocation before disabling writes during resync.
  md: use bio_kmalloc rather than bio_alloc when failure is acceptable.
  md: Fix possible deadlock with multiple mempool allocations.
  md: fix and update workqueue usage
  md: use sector_t in bitmap_get_counter
  md: remove md_mutex locking.
  md: Fix regression with raid1 arrays without persistent metadata.

drivers/md/bitmap.c

@@ -212,7 +212,7 @@ static struct page *read_sb_page(mddev_t *mddev, loff_t offset,
 
 		target = rdev->sb_start + offset + index * (PAGE_SIZE/512);
 
-		if (sync_page_io(rdev->bdev, target,
+		if (sync_page_io(rdev, target,
 				 roundup(size, bdev_logical_block_size(rdev->bdev)),
 				 page, READ)) {
 			page->index = index;
@@ -343,7 +343,7 @@ static void write_page(struct bitmap *bitmap, struct page *page, int wait)
 			atomic_inc(&bitmap->pending_writes);
 			set_buffer_locked(bh);
 			set_buffer_mapped(bh);
-			submit_bh(WRITE, bh);
+			submit_bh(WRITE | REQ_UNPLUG | REQ_SYNC, bh);
 			bh = bh->b_this_page;
 		}
 
@@ -1101,7 +1101,7 @@ static void bitmap_count_page(struct bitmap *bitmap, sector_t offset, int inc)
 	bitmap_checkfree(bitmap, page);
 }
 static bitmap_counter_t *bitmap_get_counter(struct bitmap *bitmap,
-					    sector_t offset, int *blocks,
+					    sector_t offset, sector_t *blocks,
 					    int create);
 
 /*
@@ -1115,7 +1115,7 @@ void bitmap_daemon_work(mddev_t *mddev)
 	unsigned long j;
 	unsigned long flags;
 	struct page *page = NULL, *lastpage = NULL;
-	int blocks;
+	sector_t blocks;
 	void *paddr;
 	struct dm_dirty_log *log = mddev->bitmap_info.log;
 
@@ -1258,7 +1258,7 @@ void bitmap_daemon_work(mddev_t *mddev)
 }
 
 static bitmap_counter_t *bitmap_get_counter(struct bitmap *bitmap,
-					    sector_t offset, int *blocks,
+					    sector_t offset, sector_t *blocks,
 					    int create)
 __releases(bitmap->lock)
 __acquires(bitmap->lock)
@@ -1316,7 +1316,7 @@ int bitmap_startwrite(struct bitmap *bitmap, sector_t offset, unsigned long sect
 	}
 
 	while (sectors) {
-		int blocks;
+		sector_t blocks;
 		bitmap_counter_t *bmc;
 
 		spin_lock_irq(&bitmap->lock);
@@ -1381,7 +1381,7 @@ void bitmap_endwrite(struct bitmap *bitmap, sector_t offset, unsigned long secto
 		success = 0;
 
 	while (sectors) {
-		int blocks;
+		sector_t blocks;
 		unsigned long flags;
 		bitmap_counter_t *bmc;
 
@@ -1423,7 +1423,7 @@ void bitmap_endwrite(struct bitmap *bitmap, sector_t offset, unsigned long secto
 }
 EXPORT_SYMBOL(bitmap_endwrite);
 
-static int __bitmap_start_sync(struct bitmap *bitmap, sector_t offset, int *blocks,
+static int __bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
 			       int degraded)
 {
 	bitmap_counter_t *bmc;
@@ -1452,7 +1452,7 @@ static int __bitmap_start_sync(struct bitmap *bitmap, sector_t offset, int *bloc
 	return rv;
 }
 
-int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, int *blocks,
+int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks,
 		      int degraded)
 {
 	/* bitmap_start_sync must always report on multiples of whole
@@ -1463,7 +1463,7 @@ int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, int *blocks,
 	 * Return the 'or' of the result.
 	 */
 	int rv = 0;
-	int blocks1;
+	sector_t blocks1;
 
 	*blocks = 0;
 	while (*blocks < (PAGE_SIZE>>9)) {
@@ -1476,7 +1476,7 @@ int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, int *blocks,
 }
 EXPORT_SYMBOL(bitmap_start_sync);
 
-void bitmap_end_sync(struct bitmap *bitmap, sector_t offset, int *blocks, int aborted)
+void bitmap_end_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks, int aborted)
 {
 	bitmap_counter_t *bmc;
 	unsigned long flags;
@@ -1515,7 +1515,7 @@ void bitmap_close_sync(struct bitmap *bitmap)
 	 * RESYNC bit wherever it is still on
 	 */
 	sector_t sector = 0;
-	int blocks;
+	sector_t blocks;
 	if (!bitmap)
 		return;
 	while (sector < bitmap->mddev->resync_max_sectors) {
@@ -1528,7 +1528,7 @@ EXPORT_SYMBOL(bitmap_close_sync);
 void bitmap_cond_end_sync(struct bitmap *bitmap, sector_t sector)
 {
 	sector_t s = 0;
-	int blocks;
+	sector_t blocks;
 
 	if (!bitmap)
 		return;
@@ -1562,7 +1562,7 @@ static void bitmap_set_memory_bits(struct bitmap *bitmap, sector_t offset, int n
 	 * be 0 at this point
 	 */
 
-	int secs;
+	sector_t secs;
 	bitmap_counter_t *bmc;
 	spin_lock_irq(&bitmap->lock);
 	bmc = bitmap_get_counter(bitmap, offset, &secs, 1);
@@ -1790,7 +1790,7 @@ int bitmap_load(mddev_t *mddev)
 	 * All chunks should be clean, but some might need_sync.
 	 */
 	while (sector < mddev->resync_max_sectors) {
-		int blocks;
+		sector_t blocks;
 		bitmap_start_sync(bitmap, sector, &blocks, 0);
 		sector += blocks;
 	}

drivers/md/bitmap.h

@@ -271,8 +271,8 @@ int bitmap_startwrite(struct bitmap *bitmap, sector_t offset,
 			unsigned long sectors, int behind);
 void bitmap_endwrite(struct bitmap *bitmap, sector_t offset,
 			unsigned long sectors, int success, int behind);
-int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, int *blocks, int degraded);
-void bitmap_end_sync(struct bitmap *bitmap, sector_t offset, int *blocks, int aborted);
+int bitmap_start_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks, int degraded);
+void bitmap_end_sync(struct bitmap *bitmap, sector_t offset, sector_t *blocks, int aborted);
 void bitmap_close_sync(struct bitmap *bitmap);
 void bitmap_cond_end_sync(struct bitmap *bitmap, sector_t sector);
 

drivers/md/faulty.c

@@ -210,7 +210,7 @@ static int make_request(mddev_t *mddev, struct bio *bio)
 		}
 	}
 	if (failit) {
-		struct bio *b = bio_clone(bio, GFP_NOIO);
+		struct bio *b = bio_clone_mddev(bio, GFP_NOIO, mddev);
 		b->bi_bdev = conf->rdev->bdev;
 		b->bi_private = bio;
 		b->bi_end_io = faulty_fail;

drivers/md/md.c

@@ -57,8 +57,6 @@
 #define DEBUG 0
 #define dprintk(x...) ((void)(DEBUG && printk(x)))
 
-static DEFINE_MUTEX(md_mutex);
-
 #ifndef MODULE
 static void autostart_arrays(int part);
 #endif
@@ -69,6 +67,8 @@ static DEFINE_SPINLOCK(pers_lock);
 static void md_print_devices(void);
 
 static DECLARE_WAIT_QUEUE_HEAD(resync_wait);
+static struct workqueue_struct *md_wq;
+static struct workqueue_struct *md_misc_wq;
 
 #define MD_BUG(x...) { printk("md: bug in file %s, line %d\n", __FILE__, __LINE__); md_print_devices(); }
 
@@ -149,6 +149,72 @@ static const struct block_device_operations md_fops;
 
 static int start_readonly;
 
+/* bio_clone_mddev
+ * like bio_clone, but with a local bio set
+ */
+
+static void mddev_bio_destructor(struct bio *bio)
+{
+	mddev_t *mddev, **mddevp;
+
+	mddevp = (void*)bio;
+	mddev = mddevp[-1];
+
+	bio_free(bio, mddev->bio_set);
+}
+
+struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
+			    mddev_t *mddev)
+{
+	struct bio *b;
+	mddev_t **mddevp;
+
+	if (!mddev || !mddev->bio_set)
+		return bio_alloc(gfp_mask, nr_iovecs);
+
+	b = bio_alloc_bioset(gfp_mask, nr_iovecs,
+			     mddev->bio_set);
+	if (!b)
+		return NULL;
+	mddevp = (void*)b;
+	mddevp[-1] = mddev;
+	b->bi_destructor = mddev_bio_destructor;
+	return b;
+}
+EXPORT_SYMBOL_GPL(bio_alloc_mddev);
+
+struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask,
+			    mddev_t *mddev)
+{
+	struct bio *b;
+	mddev_t **mddevp;
+
+	if (!mddev || !mddev->bio_set)
+		return bio_clone(bio, gfp_mask);
+
+	b = bio_alloc_bioset(gfp_mask, bio->bi_max_vecs,
+			     mddev->bio_set);
+	if (!b)
+		return NULL;
+	mddevp = (void*)b;
+	mddevp[-1] = mddev;
+	b->bi_destructor = mddev_bio_destructor;
+	__bio_clone(b, bio);
+	if (bio_integrity(bio)) {
+		int ret;
+
+		ret = bio_integrity_clone(b, bio, gfp_mask, mddev->bio_set);
+
+		if (ret < 0) {
+			bio_put(b);
+			return NULL;
+		}
+	}
+
+	return b;
+}
+EXPORT_SYMBOL_GPL(bio_clone_mddev);
+
 /*
  * We have a system wide 'event count' that is incremented
  * on any 'interesting' event, and readers of /proc/mdstat
@@ -300,7 +366,7 @@ static void md_end_flush(struct bio *bio, int err)
 
 	if (atomic_dec_and_test(&mddev->flush_pending)) {
 		/* The pre-request flush has finished */
-		schedule_work(&mddev->flush_work);
+		queue_work(md_wq, &mddev->flush_work);
 	}
 	bio_put(bio);
 }
@@ -321,7 +387,7 @@ static void submit_flushes(mddev_t *mddev)
 			atomic_inc(&rdev->nr_pending);
 			atomic_inc(&rdev->nr_pending);
 			rcu_read_unlock();
-			bi = bio_alloc(GFP_KERNEL, 0);
+			bi = bio_alloc_mddev(GFP_KERNEL, 0, mddev);
 			bi->bi_end_io = md_end_flush;
 			bi->bi_private = rdev;
 			bi->bi_bdev = rdev->bdev;
@@ -369,7 +435,7 @@ void md_flush_request(mddev_t *mddev, struct bio *bio)
 	submit_flushes(mddev);
 
 	if (atomic_dec_and_test(&mddev->flush_pending))
-		schedule_work(&mddev->flush_work);
+		queue_work(md_wq, &mddev->flush_work);
 }
 EXPORT_SYMBOL(md_flush_request);
 
@@ -428,6 +494,8 @@ static void mddev_delayed_delete(struct work_struct *ws);
 
 static void mddev_put(mddev_t *mddev)
 {
+	struct bio_set *bs = NULL;
+
 	if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock))
 		return;
 	if (!mddev->raid_disks && list_empty(&mddev->disks) &&
@@ -435,19 +503,22 @@ static void mddev_put(mddev_t *mddev)
 		/* Array is not configured at all, and not held active,
 		 * so destroy it */
 		list_del(&mddev->all_mddevs);
+		bs = mddev->bio_set;
+		mddev->bio_set = NULL;
 		if (mddev->gendisk) {
-			/* we did a probe so need to clean up.
-			 * Call schedule_work inside the spinlock
-			 * so that flush_scheduled_work() after
-			 * mddev_find will succeed in waiting for the
-			 * work to be done.
+			/* We did a probe so need to clean up.  Call
+			 * queue_work inside the spinlock so that
+			 * flush_workqueue() after mddev_find will
+			 * succeed in waiting for the work to be done.
 			 */
 			INIT_WORK(&mddev->del_work, mddev_delayed_delete);
-			schedule_work(&mddev->del_work);
+			queue_work(md_misc_wq, &mddev->del_work);
 		} else
 			kfree(mddev);
 	}
 	spin_unlock(&all_mddevs_lock);
+	if (bs)
+		bioset_free(bs);
 }
 
 void mddev_init(mddev_t *mddev)
@@ -691,7 +762,7 @@ void md_super_write(mddev_t *mddev, mdk_rdev_t *rdev,
 	 * if zero is reached.
 	 * If an error occurred, call md_error
 	 */
-	struct bio *bio = bio_alloc(GFP_NOIO, 1);
+	struct bio *bio = bio_alloc_mddev(GFP_NOIO, 1, mddev);
 
 	bio->bi_bdev = rdev->bdev;
 	bio->bi_sector = sector;
@@ -722,16 +793,16 @@ static void bi_complete(struct bio *bio, int error)
 	complete((struct completion*)bio->bi_private);
 }
 
-int sync_page_io(struct block_device *bdev, sector_t sector, int size,
-		   struct page *page, int rw)
+int sync_page_io(mdk_rdev_t *rdev, sector_t sector, int size,
+		 struct page *page, int rw)
 {
-	struct bio *bio = bio_alloc(GFP_NOIO, 1);
+	struct bio *bio = bio_alloc_mddev(GFP_NOIO, 1, rdev->mddev);
 	struct completion event;
 	int ret;
 
 	rw |= REQ_SYNC | REQ_UNPLUG;
 
-	bio->bi_bdev = bdev;
+	bio->bi_bdev = rdev->bdev;
 	bio->bi_sector = sector;
 	bio_add_page(bio, page, size, 0);
 	init_completion(&event);
@@ -757,7 +828,7 @@ static int read_disk_sb(mdk_rdev_t * rdev, int size)
 		return 0;
 
 
-	if (!sync_page_io(rdev->bdev, rdev->sb_start, size, rdev->sb_page, READ))
+	if (!sync_page_io(rdev, rdev->sb_start, size, rdev->sb_page, READ))
 		goto fail;
 	rdev->sb_loaded = 1;
 	return 0;
@@ -1850,7 +1921,7 @@ static void unbind_rdev_from_array(mdk_rdev_t * rdev)
 	synchronize_rcu();
 	INIT_WORK(&rdev->del_work, md_delayed_delete);
 	kobject_get(&rdev->kobj);
-	schedule_work(&rdev->del_work);
+	queue_work(md_misc_wq, &rdev->del_work);
 }
 
 /*
@@ -2108,6 +2179,8 @@ static void md_update_sb(mddev_t * mddev, int force_change)
 	if (!mddev->persistent) {
 		clear_bit(MD_CHANGE_CLEAN, &mddev->flags);
 		clear_bit(MD_CHANGE_DEVS, &mddev->flags);
+		if (!mddev->external)
+			clear_bit(MD_CHANGE_PENDING, &mddev->flags);
 		wake_up(&mddev->sb_wait);
 		return;
 	}
@@ -4192,10 +4265,10 @@ static int md_alloc(dev_t dev, char *name)
 	shift = partitioned ? MdpMinorShift : 0;
 	unit = MINOR(mddev->unit) >> shift;
 
-	/* wait for any previous instance if this device
-	 * to be completed removed (mddev_delayed_delete).
+	/* wait for any previous instance of this device to be
+	 * completely removed (mddev_delayed_delete).
 	 */
-	flush_scheduled_work();
+	flush_workqueue(md_misc_wq);
 
 	mutex_lock(&disks_mutex);
 	error = -EEXIST;
@@ -4378,6 +4451,9 @@ int md_run(mddev_t *mddev)
 		sysfs_notify_dirent_safe(rdev->sysfs_state);
 	}
 
+	if (mddev->bio_set == NULL)
+		mddev->bio_set = bioset_create(BIO_POOL_SIZE, sizeof(mddev));
+
 	spin_lock(&pers_lock);
 	pers = find_pers(mddev->level, mddev->clevel);
 	if (!pers || !try_module_get(pers->owner)) {
@@ -5885,16 +5961,14 @@ static int md_open(struct block_device *bdev, fmode_t mode)
 	mddev_t *mddev = mddev_find(bdev->bd_dev);
 	int err;
 
-	mutex_lock(&md_mutex);
 	if (mddev->gendisk != bdev->bd_disk) {
 		/* we are racing with mddev_put which is discarding this
 		 * bd_disk.
 		 */
 		mddev_put(mddev);
 		/* Wait until bdev->bd_disk is definitely gone */
-		flush_scheduled_work();
+		flush_workqueue(md_misc_wq);
 		/* Then retry the open from the top */
-		mutex_unlock(&md_mutex);
 		return -ERESTARTSYS;
 	}
 	BUG_ON(mddev != bdev->bd_disk->private_data);
@@ -5908,7 +5982,6 @@ static int md_open(struct block_device *bdev, fmode_t mode)
 
 	check_disk_size_change(mddev->gendisk, bdev);
  out:
-	mutex_unlock(&md_mutex);
 	return err;
 }
 
@@ -5917,10 +5990,8 @@ static int md_release(struct gendisk *disk, fmode_t mode)
  	mddev_t *mddev = disk->private_data;
 
 	BUG_ON(!mddev);
-	mutex_lock(&md_mutex);
 	atomic_dec(&mddev->openers);
 	mddev_put(mddev);
-	mutex_unlock(&md_mutex);
 
 	return 0;
 }
@@ -6052,7 +6123,7 @@ void md_error(mddev_t *mddev, mdk_rdev_t *rdev)
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 	md_wakeup_thread(mddev->thread);
 	if (mddev->event_work.func)
-		schedule_work(&mddev->event_work);
+		queue_work(md_misc_wq, &mddev->event_work);
 	md_new_event_inintr(mddev);
 }
 
@@ -7212,12 +7283,23 @@ static void md_geninit(void)
 
 static int __init md_init(void)
 {
-	if (register_blkdev(MD_MAJOR, "md"))
-		return -1;
-	if ((mdp_major=register_blkdev(0, "mdp"))<=0) {
-		unregister_blkdev(MD_MAJOR, "md");
-		return -1;
-	}
+	int ret = -ENOMEM;
+
+	md_wq = alloc_workqueue("md", WQ_RESCUER, 0);
+	if (!md_wq)
+		goto err_wq;
+
+	md_misc_wq = alloc_workqueue("md_misc", 0, 0);
+	if (!md_misc_wq)
+		goto err_misc_wq;
+
+	if ((ret = register_blkdev(MD_MAJOR, "md")) < 0)
+		goto err_md;
+
+	if ((ret = register_blkdev(0, "mdp")) < 0)
+		goto err_mdp;
+	mdp_major = ret;
+
 	blk_register_region(MKDEV(MD_MAJOR, 0), 1UL<<MINORBITS, THIS_MODULE,
 			    md_probe, NULL, NULL);
 	blk_register_region(MKDEV(mdp_major, 0), 1UL<<MINORBITS, THIS_MODULE,
@@ -7228,8 +7310,16 @@ static int __init md_init(void)
 
 	md_geninit();
 	return 0;
-}
 
+err_mdp:
+	unregister_blkdev(MD_MAJOR, "md");
+err_md:
+	destroy_workqueue(md_misc_wq);
+err_misc_wq:
+	destroy_workqueue(md_wq);
+err_wq:
+	return ret;
+}
 
 #ifndef MODULE
 
@@ -7316,6 +7406,8 @@ static __exit void md_exit(void)
 		export_array(mddev);
 		mddev->hold_active = 0;
 	}
+	destroy_workqueue(md_misc_wq);
+	destroy_workqueue(md_wq);
 }
 
 subsys_initcall(md_init);

drivers/md/md.h

@@ -331,6 +331,8 @@ struct mddev_s
 	struct attribute_group		*to_remove;
 	struct plug_handle		*plug; /* if used by personality */
 
+	struct bio_set			*bio_set;
+
 	/* Generic flush handling.
 	 * The last to finish preflush schedules a worker to submit
 	 * the rest of the request (without the REQ_FLUSH flag).
@@ -495,7 +497,7 @@ extern void md_flush_request(mddev_t *mddev, struct bio *bio);
 extern void md_super_write(mddev_t *mddev, mdk_rdev_t *rdev,
 			   sector_t sector, int size, struct page *page);
 extern void md_super_wait(mddev_t *mddev);
-extern int sync_page_io(struct block_device *bdev, sector_t sector, int size,
+extern int sync_page_io(mdk_rdev_t *rdev, sector_t sector, int size,
 			struct page *page, int rw);
 extern void md_do_sync(mddev_t *mddev);
 extern void md_new_event(mddev_t *mddev);
@@ -517,4 +519,8 @@ extern void md_rdev_init(mdk_rdev_t *rdev);
 
 extern void mddev_suspend(mddev_t *mddev);
 extern void mddev_resume(mddev_t *mddev);
+extern struct bio *bio_clone_mddev(struct bio *bio, gfp_t gfp_mask,
+				   mddev_t *mddev);
+extern struct bio *bio_alloc_mddev(gfp_t gfp_mask, int nr_iovecs,
+				   mddev_t *mddev);
 #endif /* _MD_MD_H */

drivers/md/raid1.h

@@ -35,8 +35,6 @@ struct r1_private_data_s {
 	struct list_head	retry_list;
 	/* queue pending writes and submit them on unplug */
 	struct bio_list		pending_bio_list;
-	/* queue of writes that have been unplugged */
-	struct bio_list		flushing_bio_list;
 
 	/* for use when syncing mirrors: */
 

drivers/md/raid10.c

@@ -120,7 +120,7 @@ static void * r10buf_pool_alloc(gfp_t gfp_flags, void *data)
 	 * Allocate bios.
 	 */
 	for (j = nalloc ; j-- ; ) {
-		bio = bio_alloc(gfp_flags, RESYNC_PAGES);
+		bio = bio_kmalloc(gfp_flags, RESYNC_PAGES);
 		if (!bio)
 			goto out_free_bio;
 		r10_bio->devs[j].bio = bio;
@@ -801,7 +801,6 @@ static int make_request(mddev_t *mddev, struct bio * bio)
 	const int rw = bio_data_dir(bio);
 	const unsigned long do_sync = (bio->bi_rw & REQ_SYNC);
 	const unsigned long do_fua = (bio->bi_rw & REQ_FUA);
-	struct bio_list bl;
 	unsigned long flags;
 	mdk_rdev_t *blocked_rdev;
 
@@ -890,7 +889,7 @@ static int make_request(mddev_t *mddev, struct bio * bio)
 		}
 		mirror = conf->mirrors + disk;
 
-		read_bio = bio_clone(bio, GFP_NOIO);
+		read_bio = bio_clone_mddev(bio, GFP_NOIO, mddev);
 
 		r10_bio->devs[slot].bio = read_bio;
 
@@ -950,16 +949,16 @@ static int make_request(mddev_t *mddev, struct bio * bio)
 		goto retry_write;
 	}
 
-	atomic_set(&r10_bio->remaining, 0);
+	atomic_set(&r10_bio->remaining, 1);
+	bitmap_startwrite(mddev->bitmap, bio->bi_sector, r10_bio->sectors, 0);
 
-	bio_list_init(&bl);
 	for (i = 0; i < conf->copies; i++) {
 		struct bio *mbio;
 		int d = r10_bio->devs[i].devnum;
 		if (!r10_bio->devs[i].bio)
 			continue;
 
-		mbio = bio_clone(bio, GFP_NOIO);
+		mbio = bio_clone_mddev(bio, GFP_NOIO, mddev);
 		r10_bio->devs[i].bio = mbio;
 
 		mbio->bi_sector	= r10_bio->devs[i].addr+
@@ -970,22 +969,22 @@ static int make_request(mddev_t *mddev, struct bio * bio)
 		mbio->bi_private = r10_bio;
 
 		atomic_inc(&r10_bio->remaining);
-		bio_list_add(&bl, mbio);
+		spin_lock_irqsave(&conf->device_lock, flags);
+		bio_list_add(&conf->pending_bio_list, mbio);
+		blk_plug_device(mddev->queue);
+		spin_unlock_irqrestore(&conf->device_lock, flags);
 	}
 
-	if (unlikely(!atomic_read(&r10_bio->remaining))) {
-		/* the array is dead */
+	if (atomic_dec_and_test(&r10_bio->remaining)) {
+		/* This matches the end of raid10_end_write_request() */
+		bitmap_endwrite(r10_bio->mddev->bitmap, r10_bio->sector,
+				r10_bio->sectors,
+				!test_bit(R10BIO_Degraded, &r10_bio->state),
+				0);
 		md_write_end(mddev);
 		raid_end_bio_io(r10_bio);
-		return 0;
 	}
 
-	bitmap_startwrite(mddev->bitmap, bio->bi_sector, r10_bio->sectors, 0);
-	spin_lock_irqsave(&conf->device_lock, flags);
-	bio_list_merge(&conf->pending_bio_list, &bl);
-	blk_plug_device(mddev->queue);
-	spin_unlock_irqrestore(&conf->device_lock, flags);
-
 	/* In case raid10d snuck in to freeze_array */
 	wake_up(&conf->wait_barrier);
 
@@ -1558,7 +1557,7 @@ static void fix_read_error(conf_t *conf, mddev_t *mddev, r10bio_t *r10_bio)
 			    test_bit(In_sync, &rdev->flags)) {
 				atomic_inc(&rdev->nr_pending);
 				rcu_read_unlock();
-				success = sync_page_io(rdev->bdev,
+				success = sync_page_io(rdev,
 						       r10_bio->devs[sl].addr +
 						       sect + rdev->data_offset,
 						       s<<9,
@@ -1597,7 +1596,7 @@ static void fix_read_error(conf_t *conf, mddev_t *mddev, r10bio_t *r10_bio)
 				atomic_inc(&rdev->nr_pending);
 				rcu_read_unlock();
 				atomic_add(s, &rdev->corrected_errors);
-				if (sync_page_io(rdev->bdev,
+				if (sync_page_io(rdev,
 						 r10_bio->devs[sl].addr +
 						 sect + rdev->data_offset,
 						 s<<9, conf->tmppage, WRITE)
@@ -1634,7 +1633,7 @@ static void fix_read_error(conf_t *conf, mddev_t *mddev, r10bio_t *r10_bio)
 				char b[BDEVNAME_SIZE];
 				atomic_inc(&rdev->nr_pending);
 				rcu_read_unlock();
-				if (sync_page_io(rdev->bdev,
+				if (sync_page_io(rdev,
 						 r10_bio->devs[sl].addr +
 						 sect + rdev->data_offset,
 						 s<<9, conf->tmppage,
@@ -1747,7 +1746,8 @@ static void raid10d(mddev_t *mddev)
 					       mdname(mddev),
 					       bdevname(rdev->bdev,b),
 					       (unsigned long long)r10_bio->sector);
-				bio = bio_clone(r10_bio->master_bio, GFP_NOIO);
+				bio = bio_clone_mddev(r10_bio->master_bio,
+						      GFP_NOIO, mddev);
 				r10_bio->devs[r10_bio->read_slot].bio = bio;
 				bio->bi_sector = r10_bio->devs[r10_bio->read_slot].addr
 					+ rdev->data_offset;
@@ -1820,7 +1820,7 @@ static sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *skipped, i
 	int disk;
 	int i;
 	int max_sync;
-	int sync_blocks;
+	sector_t sync_blocks;
 
 	sector_t sectors_skipped = 0;
 	int chunks_skipped = 0;

drivers/md/raid5.c

@@ -3876,9 +3876,9 @@ static int chunk_aligned_read(mddev_t *mddev, struct bio * raid_bio)
 		return 0;
 	}
 	/*
-	 * use bio_clone to make a copy of the bio
+	 * use bio_clone_mddev to make a copy of the bio
 	 */
-	align_bi = bio_clone(raid_bio, GFP_NOIO);
+	align_bi = bio_clone_mddev(raid_bio, GFP_NOIO, mddev);
 	if (!align_bi)
 		return 0;
 	/*
@@ -4360,7 +4360,7 @@ static inline sector_t sync_request(mddev_t *mddev, sector_t sector_nr, int *ski
 	raid5_conf_t *conf = mddev->private;
 	struct stripe_head *sh;
 	sector_t max_sector = mddev->dev_sectors;
-	int sync_blocks;
+	sector_t sync_blocks;
 	int still_degraded = 0;
 	int i;