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

* 'for-linus' of git://neil.brown.name/md:
  md: don't retry recovery of raid1 that fails due to error on source drive.
  md: Allow md devices to be created by name.
  md: make devices disappear when they are no longer needed.
  md: centralise all freeing of an 'mddev' in 'md_free'
  md: move allocation of ->queue from mddev_find to md_probe
  md: need another print_sb for mdp_superblock_1
  md: use list_for_each_entry macro directly
  md: raid0: make hash_spacing and preshift sector-based.
  md: raid0: Represent the size of strip zones in sectors.
  md: raid0 create_strip_zones(): Add KERN_INFO/KERN_ERR to printk's.
  md: raid0 create_strip_zones(): Make two local variables sector-based.
  md: raid0: Represent zone->zone_offset in sectors.
  md: raid0: Represent device offset in sectors.
  md: raid0_make_request(): Replace local variable block by sector.
  md: raid0_make_request(): Remove local variable chunk_size.
  md: raid0_make_request(): Replace chunksize_bits by chunksect_bits.
  md: use sysfs_notify_dirent to notify changes to md/sync_action.
  md: fix bitmap-on-external-file bug.

drivers/md/bitmap.c

@@ -215,7 +215,6 @@ static struct page *read_sb_page(mddev_t *mddev, long offset,
 	/* choose a good rdev and read the page from there */
 
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 	sector_t target;
 
 	if (!page)
@@ -223,7 +222,7 @@ static struct page *read_sb_page(mddev_t *mddev, long offset,
 	if (!page)
 		return ERR_PTR(-ENOMEM);
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		if (! test_bit(In_sync, &rdev->flags)
 		    || test_bit(Faulty, &rdev->flags))
 			continue;
@@ -964,7 +963,9 @@ static int bitmap_init_from_disk(struct bitmap *bitmap, sector_t start)
 				 */
 				page = bitmap->sb_page;
 				offset = sizeof(bitmap_super_t);
-				read_sb_page(bitmap->mddev, bitmap->offset,
+				if (!file)
+					read_sb_page(bitmap->mddev,
+						     bitmap->offset,
 						     page,
 						     index, count);
 			} else if (file) {

drivers/md/faulty.c

@@ -283,7 +283,6 @@ static int reconfig(mddev_t *mddev, int layout, int chunk_size)
 static int run(mddev_t *mddev)
 {
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 	int i;
 
 	conf_t *conf = kmalloc(sizeof(*conf), GFP_KERNEL);
@@ -296,7 +295,7 @@ static int run(mddev_t *mddev)
 	}
 	conf->nfaults = 0;
 
-	rdev_for_each(rdev, tmp, mddev)
+	list_for_each_entry(rdev, &mddev->disks, same_set)
 		conf->rdev = rdev;
 
 	mddev->array_sectors = mddev->size * 2;

drivers/md/linear.c

@@ -105,7 +105,6 @@ static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks)
 	int i, nb_zone, cnt;
 	sector_t min_sectors;
 	sector_t curr_sector;
-	struct list_head *tmp;
 
 	conf = kzalloc (sizeof (*conf) + raid_disks*sizeof(dev_info_t),
 			GFP_KERNEL);
@@ -115,7 +114,7 @@ static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks)
 	cnt = 0;
 	conf->array_sectors = 0;
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		int j = rdev->raid_disk;
 		dev_info_t *disk = conf->disks + j;
 

drivers/md/md.c

@@ -214,19 +214,32 @@ static inline mddev_t *mddev_get(mddev_t *mddev)
 	return mddev;
 }
 
+static void mddev_delayed_delete(struct work_struct *ws)
+{
+	mddev_t *mddev = container_of(ws, mddev_t, del_work);
+	kobject_del(&mddev->kobj);
+	kobject_put(&mddev->kobj);
+}
+
 static void mddev_put(mddev_t *mddev)
 {
 	if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock))
 		return;
-	if (!mddev->raid_disks && list_empty(&mddev->disks)) {
+	if (!mddev->raid_disks && list_empty(&mddev->disks) &&
+	    !mddev->hold_active) {
 		list_del(&mddev->all_mddevs);
-		spin_unlock(&all_mddevs_lock);
-		blk_cleanup_queue(mddev->queue);
-		if (mddev->sysfs_state)
-			sysfs_put(mddev->sysfs_state);
-		mddev->sysfs_state = NULL;
-		kobject_put(&mddev->kobj);
+		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.
+			 */
+			INIT_WORK(&mddev->del_work, mddev_delayed_delete);
+			schedule_work(&mddev->del_work);
 		} else
+			kfree(mddev);
+	}
 	spin_unlock(&all_mddevs_lock);
 }
 
@@ -236,6 +249,8 @@ static mddev_t * mddev_find(dev_t unit)
 
  retry:
 	spin_lock(&all_mddevs_lock);
+
+	if (unit) {
 		list_for_each_entry(mddev, &all_mddevs, all_mddevs)
 			if (mddev->unit == unit) {
 				mddev_get(mddev);
@@ -245,6 +260,39 @@ static mddev_t * mddev_find(dev_t unit)
 			}
 
 		if (new) {
+			list_add(&new->all_mddevs, &all_mddevs);
+			spin_unlock(&all_mddevs_lock);
+			new->hold_active = UNTIL_IOCTL;
+			return new;
+		}
+	} else if (new) {
+		/* find an unused unit number */
+		static int next_minor = 512;
+		int start = next_minor;
+		int is_free = 0;
+		int dev = 0;
+		while (!is_free) {
+			dev = MKDEV(MD_MAJOR, next_minor);
+			next_minor++;
+			if (next_minor > MINORMASK)
+				next_minor = 0;
+			if (next_minor == start) {
+				/* Oh dear, all in use. */
+				spin_unlock(&all_mddevs_lock);
+				kfree(new);
+				return NULL;
+			}
+				
+			is_free = 1;
+			list_for_each_entry(mddev, &all_mddevs, all_mddevs)
+				if (mddev->unit == dev) {
+					is_free = 0;
+					break;
+				}
+		}
+		new->unit = dev;
+		new->md_minor = MINOR(dev);
+		new->hold_active = UNTIL_STOP;
 		list_add(&new->all_mddevs, &all_mddevs);
 		spin_unlock(&all_mddevs_lock);
 		return new;
@@ -275,16 +323,6 @@ static mddev_t * mddev_find(dev_t unit)
 	new->resync_max = MaxSector;
 	new->level = LEVEL_NONE;
 
-	new->queue = blk_alloc_queue(GFP_KERNEL);
-	if (!new->queue) {
-		kfree(new);
-		return NULL;
-	}
-	/* Can be unlocked because the queue is new: no concurrency */
-	queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, new->queue);
-
-	blk_queue_make_request(new->queue, md_fail_request);
-
 	goto retry;
 }
 
@@ -307,25 +345,23 @@ static inline void mddev_unlock(mddev_t * mddev)
 
 static mdk_rdev_t * find_rdev_nr(mddev_t *mddev, int nr)
 {
-	mdk_rdev_t * rdev;
-	struct list_head *tmp;
+	mdk_rdev_t *rdev;
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set)
 		if (rdev->desc_nr == nr)
 			return rdev;
-	}
+
 	return NULL;
 }
 
 static mdk_rdev_t * find_rdev(mddev_t * mddev, dev_t dev)
 {
-	struct list_head *tmp;
 	mdk_rdev_t *rdev;
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set)
 		if (rdev->bdev->bd_dev == dev)
 			return rdev;
-	}
+
 	return NULL;
 }
 
@@ -861,7 +897,6 @@ static int super_90_validate(mddev_t *mddev, mdk_rdev_t *rdev)
 static void super_90_sync(mddev_t *mddev, mdk_rdev_t *rdev)
 {
 	mdp_super_t *sb;
-	struct list_head *tmp;
 	mdk_rdev_t *rdev2;
 	int next_spare = mddev->raid_disks;
 
@@ -933,7 +968,7 @@ static void super_90_sync(mddev_t *mddev, mdk_rdev_t *rdev)
 		sb->state |= (1<<MD_SB_BITMAP_PRESENT);
 
 	sb->disks[0].state = (1<<MD_DISK_REMOVED);
-	rdev_for_each(rdev2, tmp, mddev) {
+	list_for_each_entry(rdev2, &mddev->disks, same_set) {
 		mdp_disk_t *d;
 		int desc_nr;
 		if (rdev2->raid_disk >= 0 && test_bit(In_sync, &rdev2->flags)
@@ -1259,7 +1294,6 @@ static int super_1_validate(mddev_t *mddev, mdk_rdev_t *rdev)
 static void super_1_sync(mddev_t *mddev, mdk_rdev_t *rdev)
 {
 	struct mdp_superblock_1 *sb;
-	struct list_head *tmp;
 	mdk_rdev_t *rdev2;
 	int max_dev, i;
 	/* make rdev->sb match mddev and rdev data. */
@@ -1307,7 +1341,7 @@ static void super_1_sync(mddev_t *mddev, mdk_rdev_t *rdev)
 	}
 
 	max_dev = 0;
-	rdev_for_each(rdev2, tmp, mddev)
+	list_for_each_entry(rdev2, &mddev->disks, same_set)
 		if (rdev2->desc_nr+1 > max_dev)
 			max_dev = rdev2->desc_nr+1;
 
@@ -1316,7 +1350,7 @@ static void super_1_sync(mddev_t *mddev, mdk_rdev_t *rdev)
 	for (i=0; i<max_dev;i++)
 		sb->dev_roles[i] = cpu_to_le16(0xfffe);
 	
-	rdev_for_each(rdev2, tmp, mddev) {
+	list_for_each_entry(rdev2, &mddev->disks, same_set) {
 		i = rdev2->desc_nr;
 		if (test_bit(Faulty, &rdev2->flags))
 			sb->dev_roles[i] = cpu_to_le16(0xfffe);
@@ -1466,6 +1500,9 @@ static int bind_rdev_to_array(mdk_rdev_t * rdev, mddev_t * mddev)
 
 	list_add_rcu(&rdev->same_set, &mddev->disks);
 	bd_claim_by_disk(rdev->bdev, rdev->bdev->bd_holder, mddev->gendisk);
+
+	/* May as well allow recovery to be retried once */
+	mddev->recovery_disabled = 0;
 	return 0;
 
  fail:
@@ -1571,8 +1608,7 @@ static void kick_rdev_from_array(mdk_rdev_t * rdev)
 
 static void export_array(mddev_t *mddev)
 {
-	struct list_head *tmp;
-	mdk_rdev_t *rdev;
+	mdk_rdev_t *rdev, *tmp;
 
 	rdev_for_each(rdev, tmp, mddev) {
 		if (!rdev->mddev) {
@@ -1593,7 +1629,7 @@ static void print_desc(mdp_disk_t *desc)
 		desc->major,desc->minor,desc->raid_disk,desc->state);
 }
 
-static void print_sb(mdp_super_t *sb)
+static void print_sb_90(mdp_super_t *sb)
 {
 	int i;
 
@@ -1624,10 +1660,57 @@ static void print_sb(mdp_super_t *sb)
 	}
 	printk(KERN_INFO "md:     THIS: ");
 	print_desc(&sb->this_disk);
-
 }
 
-static void print_rdev(mdk_rdev_t *rdev)
+static void print_sb_1(struct mdp_superblock_1 *sb)
+{
+	__u8 *uuid;
+
+	uuid = sb->set_uuid;
+	printk(KERN_INFO "md:  SB: (V:%u) (F:0x%08x) Array-ID:<%02x%02x%02x%02x"
+			":%02x%02x:%02x%02x:%02x%02x:%02x%02x%02x%02x%02x%02x>\n"
+	       KERN_INFO "md:    Name: \"%s\" CT:%llu\n",
+		le32_to_cpu(sb->major_version),
+		le32_to_cpu(sb->feature_map),
+		uuid[0], uuid[1], uuid[2], uuid[3],
+		uuid[4], uuid[5], uuid[6], uuid[7],
+		uuid[8], uuid[9], uuid[10], uuid[11],
+		uuid[12], uuid[13], uuid[14], uuid[15],
+		sb->set_name,
+		(unsigned long long)le64_to_cpu(sb->ctime)
+		       & MD_SUPERBLOCK_1_TIME_SEC_MASK);
+
+	uuid = sb->device_uuid;
+	printk(KERN_INFO "md:       L%u SZ%llu RD:%u LO:%u CS:%u DO:%llu DS:%llu SO:%llu"
+			" RO:%llu\n"
+	       KERN_INFO "md:     Dev:%08x UUID: %02x%02x%02x%02x:%02x%02x:%02x%02x:%02x%02x"
+			":%02x%02x%02x%02x%02x%02x\n"
+	       KERN_INFO "md:       (F:0x%08x) UT:%llu Events:%llu ResyncOffset:%llu CSUM:0x%08x\n"
+	       KERN_INFO "md:         (MaxDev:%u) \n",
+		le32_to_cpu(sb->level),
+		(unsigned long long)le64_to_cpu(sb->size),
+		le32_to_cpu(sb->raid_disks),
+		le32_to_cpu(sb->layout),
+		le32_to_cpu(sb->chunksize),
+		(unsigned long long)le64_to_cpu(sb->data_offset),
+		(unsigned long long)le64_to_cpu(sb->data_size),
+		(unsigned long long)le64_to_cpu(sb->super_offset),
+		(unsigned long long)le64_to_cpu(sb->recovery_offset),
+		le32_to_cpu(sb->dev_number),
+		uuid[0], uuid[1], uuid[2], uuid[3],
+		uuid[4], uuid[5], uuid[6], uuid[7],
+		uuid[8], uuid[9], uuid[10], uuid[11],
+		uuid[12], uuid[13], uuid[14], uuid[15],
+		sb->devflags,
+		(unsigned long long)le64_to_cpu(sb->utime) & MD_SUPERBLOCK_1_TIME_SEC_MASK,
+		(unsigned long long)le64_to_cpu(sb->events),
+		(unsigned long long)le64_to_cpu(sb->resync_offset),
+		le32_to_cpu(sb->sb_csum),
+		le32_to_cpu(sb->max_dev)
+		);
+}
+
+static void print_rdev(mdk_rdev_t *rdev, int major_version)
 {
 	char b[BDEVNAME_SIZE];
 	printk(KERN_INFO "md: rdev %s, SZ:%08llu F:%d S:%d DN:%u\n",
@@ -1635,15 +1718,22 @@ static void print_rdev(mdk_rdev_t *rdev)
 	        test_bit(Faulty, &rdev->flags), test_bit(In_sync, &rdev->flags),
 	        rdev->desc_nr);
 	if (rdev->sb_loaded) {
-		printk(KERN_INFO "md: rdev superblock:\n");
-		print_sb((mdp_super_t*)page_address(rdev->sb_page));
+		printk(KERN_INFO "md: rdev superblock (MJ:%d):\n", major_version);
+		switch (major_version) {
+		case 0:
+			print_sb_90((mdp_super_t*)page_address(rdev->sb_page));
+			break;
+		case 1:
+			print_sb_1((struct mdp_superblock_1 *)page_address(rdev->sb_page));
+			break;
+		}
 	} else
 		printk(KERN_INFO "md: no rdev superblock!\n");
 }
 
 static void md_print_devices(void)
 {
-	struct list_head *tmp, *tmp2;
+	struct list_head *tmp;
 	mdk_rdev_t *rdev;
 	mddev_t *mddev;
 	char b[BDEVNAME_SIZE];
@@ -1658,12 +1748,12 @@ static void md_print_devices(void)
 			bitmap_print_sb(mddev->bitmap);
 		else
 			printk("%s: ", mdname(mddev));
-		rdev_for_each(rdev, tmp2, mddev)
+		list_for_each_entry(rdev, &mddev->disks, same_set)
 			printk("<%s>", bdevname(rdev->bdev,b));
 		printk("\n");
 
-		rdev_for_each(rdev, tmp2, mddev)
-			print_rdev(rdev);
+		list_for_each_entry(rdev, &mddev->disks, same_set)
+			print_rdev(rdev, mddev->major_version);
 	}
 	printk("md:	**********************************\n");
 	printk("\n");
@@ -1679,9 +1769,8 @@ static void sync_sbs(mddev_t * mddev, int nospares)
 	 * with the rest of the array)
 	 */
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		if (rdev->sb_events == mddev->events ||
 		    (nospares &&
 		     rdev->raid_disk < 0 &&
@@ -1699,7 +1788,6 @@ static void sync_sbs(mddev_t * mddev, int nospares)
 
 static void md_update_sb(mddev_t * mddev, int force_change)
 {
-	struct list_head *tmp;
 	mdk_rdev_t *rdev;
 	int sync_req;
 	int nospares = 0;
@@ -1790,7 +1878,7 @@ repeat:
 		mdname(mddev),mddev->in_sync);
 
 	bitmap_update_sb(mddev->bitmap);
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		char b[BDEVNAME_SIZE];
 		dprintk(KERN_INFO "md: ");
 		if (rdev->sb_loaded != 1)
@@ -1999,7 +2087,6 @@ slot_store(mdk_rdev_t *rdev, const char *buf, size_t len)
 		md_wakeup_thread(rdev->mddev->thread);
 	} else if (rdev->mddev->pers) {
 		mdk_rdev_t *rdev2;
-		struct list_head *tmp;
 		/* Activating a spare .. or possibly reactivating
 		 * if we every get bitmaps working here.
 		 */
@@ -2010,7 +2097,7 @@ slot_store(mdk_rdev_t *rdev, const char *buf, size_t len)
 		if (rdev->mddev->pers->hot_add_disk == NULL)
 			return -EINVAL;
 
-		rdev_for_each(rdev2, tmp, rdev->mddev)
+		list_for_each_entry(rdev2, &rdev->mddev->disks, same_set)
 			if (rdev2->raid_disk == slot)
 				return -EEXIST;
 
@@ -2125,14 +2212,14 @@ rdev_size_store(mdk_rdev_t *rdev, const char *buf, size_t len)
 		 */
 		mddev_t *mddev;
 		int overlap = 0;
-		struct list_head *tmp, *tmp2;
+		struct list_head *tmp;
 
 		mddev_unlock(my_mddev);
 		for_each_mddev(mddev, tmp) {
 			mdk_rdev_t *rdev2;
 
 			mddev_lock(mddev);
-			rdev_for_each(rdev2, tmp2, mddev)
+			list_for_each_entry(rdev2, &mddev->disks, same_set)
 				if (test_bit(AllReserved, &rdev2->flags) ||
 				    (rdev->bdev == rdev2->bdev &&
 				     rdev != rdev2 &&
@@ -2328,8 +2415,7 @@ abort_free:
 static void analyze_sbs(mddev_t * mddev)
 {
 	int i;
-	struct list_head *tmp;
-	mdk_rdev_t *rdev, *freshest;
+	mdk_rdev_t *rdev, *freshest, *tmp;
 	char b[BDEVNAME_SIZE];
 
 	freshest = NULL;
@@ -3046,7 +3132,7 @@ action_store(mddev_t *mddev, const char *page, size_t len)
 	}
 	set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
 	md_wakeup_thread(mddev->thread);
-	sysfs_notify(&mddev->kobj, NULL, "sync_action");
+	sysfs_notify_dirent(mddev->sysfs_action);
 	return len;
 }
 
@@ -3404,6 +3490,8 @@ md_attr_store(struct kobject *kobj, struct attribute *attr,
 	if (!capable(CAP_SYS_ADMIN))
 		return -EACCES;
 	rv = mddev_lock(mddev);
+	if (mddev->hold_active == UNTIL_IOCTL)
+		mddev->hold_active = 0;
 	if (!rv) {
 		rv = entry->store(mddev, page, length);
 		mddev_unlock(mddev);
@@ -3414,6 +3502,17 @@ md_attr_store(struct kobject *kobj, struct attribute *attr,
 static void md_free(struct kobject *ko)
 {
 	mddev_t *mddev = container_of(ko, mddev_t, kobj);
+
+	if (mddev->sysfs_state)
+		sysfs_put(mddev->sysfs_state);
+
+	if (mddev->gendisk) {
+		del_gendisk(mddev->gendisk);
+		put_disk(mddev->gendisk);
+	}
+	if (mddev->queue)
+		blk_cleanup_queue(mddev->queue);
+
 	kfree(mddev);
 }
 
@@ -3429,34 +3528,74 @@ static struct kobj_type md_ktype = {
 
 int mdp_major = 0;
 
-static struct kobject *md_probe(dev_t dev, int *part, void *data)
+static int md_alloc(dev_t dev, char *name)
 {
 	static DEFINE_MUTEX(disks_mutex);
 	mddev_t *mddev = mddev_find(dev);
 	struct gendisk *disk;
-	int partitioned = (MAJOR(dev) != MD_MAJOR);
-	int shift = partitioned ? MdpMinorShift : 0;
-	int unit = MINOR(dev) >> shift;
+	int partitioned;
+	int shift;
+	int unit;
 	int error;
 
 	if (!mddev)
-		return NULL;
+		return -ENODEV;
+
+	partitioned = (MAJOR(mddev->unit) != MD_MAJOR);
+	shift = partitioned ? MdpMinorShift : 0;
+	unit = MINOR(mddev->unit) >> shift;
+
+	/* wait for any previous instance if this device
+	 * to be completed removed (mddev_delayed_delete).
+	 */
+	flush_scheduled_work();
 
 	mutex_lock(&disks_mutex);
 	if (mddev->gendisk) {
 		mutex_unlock(&disks_mutex);
 		mddev_put(mddev);
-		return NULL;
+		return -EEXIST;
 	}
+
+	if (name) {
+		/* Need to ensure that 'name' is not a duplicate.
+		 */
+		mddev_t *mddev2;
+		spin_lock(&all_mddevs_lock);
+
+		list_for_each_entry(mddev2, &all_mddevs, all_mddevs)
+			if (mddev2->gendisk &&
+			    strcmp(mddev2->gendisk->disk_name, name) == 0) {
+				spin_unlock(&all_mddevs_lock);
+				return -EEXIST;
+			}
+		spin_unlock(&all_mddevs_lock);
+	}
+
+	mddev->queue = blk_alloc_queue(GFP_KERNEL);
+	if (!mddev->queue) {
+		mutex_unlock(&disks_mutex);
+		mddev_put(mddev);
+		return -ENOMEM;
+	}
+	/* Can be unlocked because the queue is new: no concurrency */
+	queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, mddev->queue);
+
+	blk_queue_make_request(mddev->queue, md_fail_request);
+
 	disk = alloc_disk(1 << shift);
 	if (!disk) {
 		mutex_unlock(&disks_mutex);
+		blk_cleanup_queue(mddev->queue);
+		mddev->queue = NULL;
 		mddev_put(mddev);
-		return NULL;
+		return -ENOMEM;
 	}
-	disk->major = MAJOR(dev);
+	disk->major = MAJOR(mddev->unit);
 	disk->first_minor = unit << shift;
-	if (partitioned)
+	if (name)
+		strcpy(disk->disk_name, name);
+	else if (partitioned)
 		sprintf(disk->disk_name, "md_d%d", unit);
 	else
 		sprintf(disk->disk_name, "md%d", unit);
@@ -3464,7 +3603,7 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data)
 	disk->private_data = mddev;
 	disk->queue = mddev->queue;
 	/* Allow extended partitions.  This makes the
-	 * 'mdp' device redundant, but we can really
+	 * 'mdp' device redundant, but we can't really
 	 * remove it now.
 	 */
 	disk->flags |= GENHD_FL_EXT_DEVT;
@@ -3480,9 +3619,35 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data)
 		kobject_uevent(&mddev->kobj, KOBJ_ADD);
 		mddev->sysfs_state = sysfs_get_dirent(mddev->kobj.sd, "array_state");
 	}
+	mddev_put(mddev);
+	return 0;
+}
+
+static struct kobject *md_probe(dev_t dev, int *part, void *data)
+{
+	md_alloc(dev, NULL);
 	return NULL;
 }
 
+static int add_named_array(const char *val, struct kernel_param *kp)
+{
+	/* val must be "md_*" where * is not all digits.
+	 * We allocate an array with a large free minor number, and
+	 * set the name to val.  val must not already be an active name.
+	 */
+	int len = strlen(val);
+	char buf[DISK_NAME_LEN];
+
+	while (len && val[len-1] == '\n')
+		len--;
+	if (len >= DISK_NAME_LEN)
+		return -E2BIG;
+	strlcpy(buf, val, len+1);
+	if (strncmp(buf, "md_", 3) != 0)
+		return -EINVAL;
+	return md_alloc(0, buf);
+}
+
 static void md_safemode_timeout(unsigned long data)
 {
 	mddev_t *mddev = (mddev_t *) data;
@@ -3501,7 +3666,6 @@ static int do_md_run(mddev_t * mddev)
 {
 	int err;
 	int chunk_size;
-	struct list_head *tmp;
 	mdk_rdev_t *rdev;
 	struct gendisk *disk;
 	struct mdk_personality *pers;
@@ -3540,7 +3704,7 @@ static int do_md_run(mddev_t * mddev)
 		}
 
 		/* devices must have minimum size of one chunk */
-		rdev_for_each(rdev, tmp, mddev) {
+		list_for_each_entry(rdev, &mddev->disks, same_set) {
 			if (test_bit(Faulty, &rdev->flags))
 				continue;
 			if (rdev->size < chunk_size / 1024) {
@@ -3565,7 +3729,7 @@ static int do_md_run(mddev_t * mddev)
 	 * the only valid external interface is through the md
 	 * device.
 	 */
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		if (test_bit(Faulty, &rdev->flags))
 			continue;
 		sync_blockdev(rdev->bdev);
@@ -3630,10 +3794,10 @@ static int do_md_run(mddev_t * mddev)
 		 */
 		char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
 		mdk_rdev_t *rdev2;
-		struct list_head *tmp2;
 		int warned = 0;
-		rdev_for_each(rdev, tmp, mddev) {
-			rdev_for_each(rdev2, tmp2, mddev) {
+
+		list_for_each_entry(rdev, &mddev->disks, same_set)
+			list_for_each_entry(rdev2, &mddev->disks, same_set) {
 				if (rdev < rdev2 &&
 				    rdev->bdev->bd_contains ==
 				    rdev2->bdev->bd_contains) {
@@ -3647,7 +3811,7 @@ static int do_md_run(mddev_t * mddev)
 					warned = 1;
 				}
 			}
-		}
+
 		if (warned)
 			printk(KERN_WARNING
 			       "True protection against single-disk"
@@ -3684,6 +3848,7 @@ static int do_md_run(mddev_t * mddev)
 			printk(KERN_WARNING
 			       "md: cannot register extra attributes for %s\n",
 			       mdname(mddev));
+		mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, "sync_action");
 	} else if (mddev->ro == 2) /* auto-readonly not meaningful */
 		mddev->ro = 0;
 
@@ -3694,7 +3859,7 @@ static int do_md_run(mddev_t * mddev)
 	mddev->safemode_delay = (200 * HZ)/1000 +1; /* 200 msec delay */
 	mddev->in_sync = 1;
 
-	rdev_for_each(rdev, tmp, mddev)
+	list_for_each_entry(rdev, &mddev->disks, same_set)
 		if (rdev->raid_disk >= 0) {
 			char nm[20];
 			sprintf(nm, "rd%d", rdev->raid_disk);
@@ -3725,9 +3890,8 @@ static int do_md_run(mddev_t * mddev)
 	 * it will remove the drives and not do the right thing
 	 */
 	if (mddev->degraded && !mddev->sync_thread) {
-		struct list_head *rtmp;
 		int spares = 0;
-		rdev_for_each(rdev, rtmp, mddev)
+		list_for_each_entry(rdev, &mddev->disks, same_set)
 			if (rdev->raid_disk >= 0 &&
 			    !test_bit(In_sync, &rdev->flags) &&
 			    !test_bit(Faulty, &rdev->flags))
@@ -3754,7 +3918,8 @@ static int do_md_run(mddev_t * mddev)
 	mddev->changed = 1;
 	md_new_event(mddev);
 	sysfs_notify_dirent(mddev->sysfs_state);
-	sysfs_notify(&mddev->kobj, NULL, "sync_action");
+	if (mddev->sysfs_action)
+		sysfs_notify_dirent(mddev->sysfs_action);
 	sysfs_notify(&mddev->kobj, NULL, "degraded");
 	kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE);
 	return 0;
@@ -3854,9 +4019,12 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open)
 			mddev->queue->merge_bvec_fn = NULL;
 			mddev->queue->unplug_fn = NULL;
 			mddev->queue->backing_dev_info.congested_fn = NULL;
-			if (mddev->pers->sync_request)
+			if (mddev->pers->sync_request) {
 				sysfs_remove_group(&mddev->kobj, &md_redundancy_group);
-
+				if (mddev->sysfs_action)
+					sysfs_put(mddev->sysfs_action);
+				mddev->sysfs_action = NULL;
+			}
 			module_put(mddev->pers->owner);
 			mddev->pers = NULL;
 			/* tell userspace to handle 'inactive' */
@@ -3883,7 +4051,6 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open)
 	 */
 	if (mode == 0) {
 		mdk_rdev_t *rdev;
-		struct list_head *tmp;
 
 		printk(KERN_INFO "md: %s stopped.\n", mdname(mddev));
 
@@ -3895,7 +4062,7 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open)
 		}
 		mddev->bitmap_offset = 0;
 
-		rdev_for_each(rdev, tmp, mddev)
+		list_for_each_entry(rdev, &mddev->disks, same_set)
 			if (rdev->raid_disk >= 0) {
 				char nm[20];
 				sprintf(nm, "rd%d", rdev->raid_disk);
@@ -3941,6 +4108,8 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open)
 		mddev->barriers_work = 0;
 		mddev->safemode = 0;
 		kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE);
+		if (mddev->hold_active == UNTIL_STOP)
+			mddev->hold_active = 0;
 
 	} else if (mddev->pers)
 		printk(KERN_INFO "md: %s switched to read-only mode.\n",
@@ -3956,7 +4125,6 @@ out:
 static void autorun_array(mddev_t *mddev)
 {
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 	int err;
 
 	if (list_empty(&mddev->disks))
@@ -3964,7 +4132,7 @@ static void autorun_array(mddev_t *mddev)
 
 	printk(KERN_INFO "md: running: ");
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		char b[BDEVNAME_SIZE];
 		printk("<%s>", bdevname(rdev->bdev,b));
 	}
@@ -3991,8 +4159,7 @@ static void autorun_array(mddev_t *mddev)
  */
 static void autorun_devices(int part)
 {
-	struct list_head *tmp;
-	mdk_rdev_t *rdev0, *rdev;
+	mdk_rdev_t *rdev0, *rdev, *tmp;
 	mddev_t *mddev;
 	char b[BDEVNAME_SIZE];
 
@@ -4007,7 +4174,7 @@ static void autorun_devices(int part)
 		printk(KERN_INFO "md: considering %s ...\n",
 			bdevname(rdev0->bdev,b));
 		INIT_LIST_HEAD(&candidates);
-		rdev_for_each_list(rdev, tmp, pending_raid_disks)
+		rdev_for_each_list(rdev, tmp, &pending_raid_disks)
 			if (super_90_load(rdev, rdev0, 0) >= 0) {
 				printk(KERN_INFO "md:  adding %s ...\n",
 					bdevname(rdev->bdev,b));
@@ -4053,7 +4220,7 @@ static void autorun_devices(int part)
 		} else {
 			printk(KERN_INFO "md: created %s\n", mdname(mddev));
 			mddev->persistent = 1;
-			rdev_for_each_list(rdev, tmp, candidates) {
+			rdev_for_each_list(rdev, tmp, &candidates) {
 				list_del_init(&rdev->same_set);
 				if (bind_rdev_to_array(rdev, mddev))
 					export_rdev(rdev);
@@ -4064,7 +4231,7 @@ static void autorun_devices(int part)
 		/* on success, candidates will be empty, on error
 		 * it won't...
 		 */
-		rdev_for_each_list(rdev, tmp, candidates) {
+		rdev_for_each_list(rdev, tmp, &candidates) {
 			list_del_init(&rdev->same_set);
 			export_rdev(rdev);
 		}
@@ -4093,10 +4260,9 @@ static int get_array_info(mddev_t * mddev, void __user * arg)
 	mdu_array_info_t info;
 	int nr,working,active,failed,spare;
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 
 	nr=working=active=failed=spare=0;
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		nr++;
 		if (test_bit(Faulty, &rdev->flags))
 			failed++;
@@ -4614,9 +4780,8 @@ static int set_array_info(mddev_t * mddev, mdu_array_info_t *info)
 
 static int update_size(mddev_t *mddev, sector_t num_sectors)
 {
-	mdk_rdev_t * rdev;
+	mdk_rdev_t *rdev;
 	int rv;
-	struct list_head *tmp;
 	int fit = (num_sectors == 0);
 
 	if (mddev->pers->resize == NULL)
@@ -4638,7 +4803,7 @@ static int update_size(mddev_t *mddev, sector_t num_sectors)
 		 * grow, and re-add.
 		 */
 		return -EBUSY;
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		sector_t avail;
 		avail = rdev->size * 2;
 
@@ -5000,6 +5165,9 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode,
 
 done_unlock:
 abort_unlock:
+	if (mddev->hold_active == UNTIL_IOCTL &&
+	    err != -EINVAL)
+		mddev->hold_active = 0;
 	mddev_unlock(mddev);
 
 	return err;
@@ -5016,14 +5184,25 @@ static int md_open(struct block_device *bdev, fmode_t mode)
 	 * Succeed if we can lock the mddev, which confirms that
 	 * it isn't being stopped right now.
 	 */
-	mddev_t *mddev = bdev->bd_disk->private_data;
+	mddev_t *mddev = mddev_find(bdev->bd_dev);
 	int err;
 
+	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();
+		/* Then retry the open from the top */
+		return -ERESTARTSYS;
+	}
+	BUG_ON(mddev != bdev->bd_disk->private_data);
+
 	if ((err = mutex_lock_interruptible_nested(&mddev->reconfig_mutex, 1)))
 		goto out;
 
 	err = 0;
-	mddev_get(mddev);
 	atomic_inc(&mddev->openers);
 	mddev_unlock(mddev);
 
@@ -5187,11 +5366,10 @@ static void status_unused(struct seq_file *seq)
 {
 	int i = 0;
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 
 	seq_printf(seq, "unused devices: ");
 
-	rdev_for_each_list(rdev, tmp, pending_raid_disks) {
+	list_for_each_entry(rdev, &pending_raid_disks, same_set) {
 		char b[BDEVNAME_SIZE];
 		i++;
 		seq_printf(seq, "%s ",
@@ -5350,7 +5528,6 @@ static int md_seq_show(struct seq_file *seq, void *v)
 {
 	mddev_t *mddev = v;
 	sector_t size;
-	struct list_head *tmp2;
 	mdk_rdev_t *rdev;
 	struct mdstat_info *mi = seq->private;
 	struct bitmap *bitmap;
@@ -5387,7 +5564,7 @@ static int md_seq_show(struct seq_file *seq, void *v)
 		}
 
 		size = 0;
-		rdev_for_each(rdev, tmp2, mddev) {
+		list_for_each_entry(rdev, &mddev->disks, same_set) {
 			char b[BDEVNAME_SIZE];
 			seq_printf(seq, " %s[%d]",
 				bdevname(rdev->bdev,b), rdev->desc_nr);
@@ -5694,7 +5871,6 @@ void md_do_sync(mddev_t *mddev)
 	struct list_head *tmp;
 	sector_t last_check;
 	int skipped = 0;
-	struct list_head *rtmp;
 	mdk_rdev_t *rdev;
 	char *desc;
 
@@ -5799,7 +5975,7 @@ void md_do_sync(mddev_t *mddev)
 		/* recovery follows the physical size of devices */
 		max_sectors = mddev->size << 1;
 		j = MaxSector;
-		rdev_for_each(rdev, rtmp, mddev)
+		list_for_each_entry(rdev, &mddev->disks, same_set)
 			if (rdev->raid_disk >= 0 &&
 			    !test_bit(Faulty, &rdev->flags) &&
 			    !test_bit(In_sync, &rdev->flags) &&
@@ -5949,7 +6125,7 @@ void md_do_sync(mddev_t *mddev)
 		} else {
 			if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery))
 				mddev->curr_resync = MaxSector;
-			rdev_for_each(rdev, rtmp, mddev)
+			list_for_each_entry(rdev, &mddev->disks, same_set)
 				if (rdev->raid_disk >= 0 &&
 				    !test_bit(Faulty, &rdev->flags) &&
 				    !test_bit(In_sync, &rdev->flags) &&
@@ -5985,10 +6161,9 @@ EXPORT_SYMBOL_GPL(md_do_sync);
 static int remove_and_add_spares(mddev_t *mddev)
 {
 	mdk_rdev_t *rdev;
-	struct list_head *rtmp;
 	int spares = 0;
 
-	rdev_for_each(rdev, rtmp, mddev)
+	list_for_each_entry(rdev, &mddev->disks, same_set)
 		if (rdev->raid_disk >= 0 &&
 		    !test_bit(Blocked, &rdev->flags) &&
 		    (test_bit(Faulty, &rdev->flags) ||
@@ -6003,8 +6178,8 @@ static int remove_and_add_spares(mddev_t *mddev)
 			}
 		}
 
-	if (mddev->degraded && ! mddev->ro) {
-		rdev_for_each(rdev, rtmp, mddev) {
+	if (mddev->degraded && ! mddev->ro && !mddev->recovery_disabled) {
+		list_for_each_entry(rdev, &mddev->disks, same_set) {
 			if (rdev->raid_disk >= 0 &&
 			    !test_bit(In_sync, &rdev->flags) &&
 			    !test_bit(Blocked, &rdev->flags))
@@ -6056,7 +6231,6 @@ static int remove_and_add_spares(mddev_t *mddev)
 void md_check_recovery(mddev_t *mddev)
 {
 	mdk_rdev_t *rdev;
-	struct list_head *rtmp;
 
 
 	if (mddev->bitmap)
@@ -6120,7 +6294,7 @@ void md_check_recovery(mddev_t *mddev)
 		if (mddev->flags)
 			md_update_sb(mddev, 0);
 
-		rdev_for_each(rdev, rtmp, mddev)
+		list_for_each_entry(rdev, &mddev->disks, same_set)
 			if (test_and_clear_bit(StateChanged, &rdev->flags))
 				sysfs_notify_dirent(rdev->sysfs_state);
 
@@ -6149,13 +6323,13 @@ void md_check_recovery(mddev_t *mddev)
 			 * information must be scrapped
 			 */
 			if (!mddev->degraded)
-				rdev_for_each(rdev, rtmp, mddev)
+				list_for_each_entry(rdev, &mddev->disks, same_set)
 					rdev->saved_raid_disk = -1;
 
 			mddev->recovery = 0;
 			/* flag recovery needed just to double check */
 			set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);
-			sysfs_notify(&mddev->kobj, NULL, "sync_action");
+			sysfs_notify_dirent(mddev->sysfs_action);
 			md_new_event(mddev);
 			goto unlock;
 		}
@@ -6216,7 +6390,7 @@ void md_check_recovery(mddev_t *mddev)
 				mddev->recovery = 0;
 			} else
 				md_wakeup_thread(mddev->sync_thread);
-			sysfs_notify(&mddev->kobj, NULL, "sync_action");
+			sysfs_notify_dirent(mddev->sysfs_action);
 			md_new_event(mddev);
 		}
 	unlock:
@@ -6224,7 +6398,8 @@ void md_check_recovery(mddev_t *mddev)
 			clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery);
 			if (test_and_clear_bit(MD_RECOVERY_RECOVER,
 					       &mddev->recovery))
-				sysfs_notify(&mddev->kobj, NULL, "sync_action");
+				if (mddev->sysfs_action)
+					sysfs_notify_dirent(mddev->sysfs_action);
 		}
 		mddev_unlock(mddev);
 	}
@@ -6386,14 +6561,8 @@ static __exit void md_exit(void)
 	unregister_sysctl_table(raid_table_header);
 	remove_proc_entry("mdstat", NULL);
 	for_each_mddev(mddev, tmp) {
-		struct gendisk *disk = mddev->gendisk;
-		if (!disk)
-			continue;
 		export_array(mddev);
-		del_gendisk(disk);
-		put_disk(disk);
-		mddev->gendisk = NULL;
-		mddev_put(mddev);
+		mddev->hold_active = 0;
 	}
 }
 
@@ -6418,6 +6587,7 @@ static int set_ro(const char *val, struct kernel_param *kp)
 module_param_call(start_ro, set_ro, get_ro, NULL, S_IRUSR|S_IWUSR);
 module_param(start_dirty_degraded, int, S_IRUGO|S_IWUSR);
 
+module_param_call(new_array, add_named_array, NULL, NULL, S_IWUSR);
 
 EXPORT_SYMBOL(register_md_personality);
 EXPORT_SYMBOL(unregister_md_personality);

drivers/md/multipath.c

@@ -408,7 +408,6 @@ static int multipath_run (mddev_t *mddev)
 	int disk_idx;
 	struct multipath_info *disk;
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 
 	if (mddev->level != LEVEL_MULTIPATH) {
 		printk("multipath: %s: raid level not set to multipath IO (%d)\n",
@@ -441,7 +440,7 @@ static int multipath_run (mddev_t *mddev)
 	}
 
 	conf->working_disks = 0;
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		disk_idx = rdev->raid_disk;
 		if (disk_idx < 0 ||
 		    disk_idx >= mddev->raid_disks)

drivers/md/raid0.c

@@ -53,11 +53,10 @@ static int raid0_congested(void *data, int bits)
 static int create_strip_zones (mddev_t *mddev)
 {
 	int i, c, j;
-	sector_t current_offset, curr_zone_offset;
+	sector_t current_start, curr_zone_start;
 	sector_t min_spacing;
 	raid0_conf_t *conf = mddev_to_conf(mddev);
 	mdk_rdev_t *smallest, *rdev1, *rdev2, *rdev;
-	struct list_head *tmp1, *tmp2;
 	struct strip_zone *zone;
 	int cnt;
 	char b[BDEVNAME_SIZE];
@@ -67,19 +66,19 @@ static int create_strip_zones (mddev_t *mddev)
 	 */
 	conf->nr_strip_zones = 0;
  
-	rdev_for_each(rdev1, tmp1, mddev) {
-		printk("raid0: looking at %s\n",
+	list_for_each_entry(rdev1, &mddev->disks, same_set) {
+		printk(KERN_INFO "raid0: looking at %s\n",
 			bdevname(rdev1->bdev,b));
 		c = 0;
-		rdev_for_each(rdev2, tmp2, mddev) {
-			printk("raid0:   comparing %s(%llu)",
+		list_for_each_entry(rdev2, &mddev->disks, same_set) {
+			printk(KERN_INFO "raid0:   comparing %s(%llu)",
 			       bdevname(rdev1->bdev,b),
 			       (unsigned long long)rdev1->size);
-			printk(" with %s(%llu)\n",
+			printk(KERN_INFO " with %s(%llu)\n",
 			       bdevname(rdev2->bdev,b),
 			       (unsigned long long)rdev2->size);
 			if (rdev2 == rdev1) {
-				printk("raid0:   END\n");
+				printk(KERN_INFO "raid0:   END\n");
 				break;
 			}
 			if (rdev2->size == rdev1->size)
@@ -88,19 +87,20 @@ static int create_strip_zones (mddev_t *mddev)
 				 * Not unique, don't count it as a new
 				 * group
 				 */
-				printk("raid0:   EQUAL\n");
+				printk(KERN_INFO "raid0:   EQUAL\n");
 				c = 1;
 				break;
 			}
-			printk("raid0:   NOT EQUAL\n");
+			printk(KERN_INFO "raid0:   NOT EQUAL\n");
 		}
 		if (!c) {
-			printk("raid0:   ==> UNIQUE\n");
+			printk(KERN_INFO "raid0:   ==> UNIQUE\n");
 			conf->nr_strip_zones++;
-			printk("raid0: %d zones\n", conf->nr_strip_zones);
+			printk(KERN_INFO "raid0: %d zones\n",
+				conf->nr_strip_zones);
 		}
 	}
-	printk("raid0: FINAL %d zones\n", conf->nr_strip_zones);
+	printk(KERN_INFO "raid0: FINAL %d zones\n", conf->nr_strip_zones);
 
 	conf->strip_zone = kzalloc(sizeof(struct strip_zone)*
 				conf->nr_strip_zones, GFP_KERNEL);
@@ -119,16 +119,17 @@ static int create_strip_zones (mddev_t *mddev)
 	cnt = 0;
 	smallest = NULL;
 	zone->dev = conf->devlist;
-	rdev_for_each(rdev1, tmp1, mddev) {
+	list_for_each_entry(rdev1, &mddev->disks, same_set) {
 		int j = rdev1->raid_disk;
 
 		if (j < 0 || j >= mddev->raid_disks) {
-			printk("raid0: bad disk number %d - aborting!\n", j);
+			printk(KERN_ERR "raid0: bad disk number %d - "
+				"aborting!\n", j);
 			goto abort;
 		}
 		if (zone->dev[j]) {
-			printk("raid0: multiple devices for %d - aborting!\n",
-				j);
+			printk(KERN_ERR "raid0: multiple devices for %d - "
+				"aborting!\n", j);
 			goto abort;
 		}
 		zone->dev[j] = rdev1;
@@ -149,16 +150,16 @@ static int create_strip_zones (mddev_t *mddev)
 		cnt++;
 	}
 	if (cnt != mddev->raid_disks) {
-		printk("raid0: too few disks (%d of %d) - aborting!\n",
-			cnt, mddev->raid_disks);
+		printk(KERN_ERR "raid0: too few disks (%d of %d) - "
+			"aborting!\n", cnt, mddev->raid_disks);
 		goto abort;
 	}
 	zone->nb_dev = cnt;
-	zone->size = smallest->size * cnt;
-	zone->zone_offset = 0;
+	zone->sectors = smallest->size * cnt * 2;
+	zone->zone_start = 0;
 
-	current_offset = smallest->size;
-	curr_zone_offset = zone->size;
+	current_start = smallest->size * 2;
+	curr_zone_start = zone->sectors;
 
 	/* now do the other zones */
 	for (i = 1; i < conf->nr_strip_zones; i++)
@@ -166,40 +167,41 @@ static int create_strip_zones (mddev_t *mddev)
 		zone = conf->strip_zone + i;
 		zone->dev = conf->strip_zone[i-1].dev + mddev->raid_disks;
 
-		printk("raid0: zone %d\n", i);
-		zone->dev_offset = current_offset;
+		printk(KERN_INFO "raid0: zone %d\n", i);
+		zone->dev_start = current_start;
 		smallest = NULL;
 		c = 0;
 
 		for (j=0; j<cnt; j++) {
 			char b[BDEVNAME_SIZE];
 			rdev = conf->strip_zone[0].dev[j];
-			printk("raid0: checking %s ...", bdevname(rdev->bdev,b));
-			if (rdev->size > current_offset)
-			{
-				printk(" contained as device %d\n", c);
+			printk(KERN_INFO "raid0: checking %s ...",
+				bdevname(rdev->bdev, b));
+			if (rdev->size > current_start / 2) {
+				printk(KERN_INFO " contained as device %d\n",
+					c);
 				zone->dev[c] = rdev;
 				c++;
 				if (!smallest || (rdev->size <smallest->size)) {
 					smallest = rdev;
-					printk("  (%llu) is smallest!.\n", 
+					printk(KERN_INFO "  (%llu) is smallest!.\n",
 						(unsigned long long)rdev->size);
 				}
 			} else
-				printk(" nope.\n");
+				printk(KERN_INFO " nope.\n");
 		}
 
 		zone->nb_dev = c;
-		zone->size = (smallest->size - current_offset) * c;
-		printk("raid0: zone->nb_dev: %d, size: %llu\n",
-			zone->nb_dev, (unsigned long long)zone->size);
+		zone->sectors = (smallest->size * 2 - current_start) * c;
+		printk(KERN_INFO "raid0: zone->nb_dev: %d, sectors: %llu\n",
+			zone->nb_dev, (unsigned long long)zone->sectors);
 
-		zone->zone_offset = curr_zone_offset;
-		curr_zone_offset += zone->size;
+		zone->zone_start = curr_zone_start;
+		curr_zone_start += zone->sectors;
 
-		current_offset = smallest->size;
-		printk("raid0: current zone offset: %llu\n",
-			(unsigned long long)current_offset);
+		current_start = smallest->size * 2;
+		printk(KERN_INFO "raid0: current zone start: %llu\n",
+			(unsigned long long)current_start);
 	}
 
 	/* Now find appropriate hash spacing.
@@ -210,16 +212,16 @@ static int create_strip_zones (mddev_t *mddev)
 	 * strip though as it's size has no bearing on the efficacy of the hash
 	 * table.
 	 */
-	conf->hash_spacing = curr_zone_offset;
-	min_spacing = curr_zone_offset;
+	conf->spacing = curr_zone_start;
+	min_spacing = curr_zone_start;
 	sector_div(min_spacing, PAGE_SIZE/sizeof(struct strip_zone*));
 	for (i=0; i < conf->nr_strip_zones-1; i++) {
-		sector_t sz = 0;
-		for (j=i; j<conf->nr_strip_zones-1 &&
-			     sz < min_spacing ; j++)
-			sz += conf->strip_zone[j].size;
-		if (sz >= min_spacing && sz < conf->hash_spacing)
-			conf->hash_spacing = sz;
+		sector_t s = 0;
+		for (j = i; j < conf->nr_strip_zones - 1 &&
+				s < min_spacing; j++)
+			s += conf->strip_zone[j].sectors;
+		if (s >= min_spacing && s < conf->spacing)
+			conf->spacing = s;
 	}
 
 	mddev->queue->unplug_fn = raid0_unplug;
@@ -227,7 +229,7 @@ static int create_strip_zones (mddev_t *mddev)
 	mddev->queue->backing_dev_info.congested_fn = raid0_congested;
 	mddev->queue->backing_dev_info.congested_data = mddev;
 
-	printk("raid0: done.\n");
+	printk(KERN_INFO "raid0: done.\n");
 	return 0;
  abort:
 	return 1;
@@ -262,10 +264,9 @@ static int raid0_mergeable_bvec(struct request_queue *q,
 static int raid0_run (mddev_t *mddev)
 {
 	unsigned  cur=0, i=0, nb_zone;
-	s64 size;
+	s64 sectors;
 	raid0_conf_t *conf;
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 
 	if (mddev->chunk_size == 0) {
 		printk(KERN_ERR "md/raid0: non-zero chunk size required.\n");
@@ -291,54 +292,54 @@ static int raid0_run (mddev_t *mddev)
 
 	/* calculate array device size */
 	mddev->array_sectors = 0;
-	rdev_for_each(rdev, tmp, mddev)
+	list_for_each_entry(rdev, &mddev->disks, same_set)
 		mddev->array_sectors += rdev->size * 2;
 
-	printk("raid0 : md_size is %llu blocks.\n", 
-		(unsigned long long)mddev->array_sectors / 2);
-	printk("raid0 : conf->hash_spacing is %llu blocks.\n",
-		(unsigned long long)conf->hash_spacing);
+	printk(KERN_INFO "raid0 : md_size is %llu sectors.\n",
+		(unsigned long long)mddev->array_sectors);
+	printk(KERN_INFO "raid0 : conf->spacing is %llu sectors.\n",
+		(unsigned long long)conf->spacing);
 	{
-		sector_t s = mddev->array_sectors / 2;
-		sector_t space = conf->hash_spacing;
+		sector_t s = mddev->array_sectors;
+		sector_t space = conf->spacing;
 		int round;
-		conf->preshift = 0;
+		conf->sector_shift = 0;
 		if (sizeof(sector_t) > sizeof(u32)) {
 			/*shift down space and s so that sector_div will work */
 			while (space > (sector_t) (~(u32)0)) {
 				s >>= 1;
 				space >>= 1;
 				s += 1; /* force round-up */
-				conf->preshift++;
+				conf->sector_shift++;
 			}
 		}
 		round = sector_div(s, (u32)space) ? 1 : 0;
 		nb_zone = s + round;
 	}
-	printk("raid0 : nb_zone is %d.\n", nb_zone);
+	printk(KERN_INFO "raid0 : nb_zone is %d.\n", nb_zone);
 
-	printk("raid0 : Allocating %Zd bytes for hash.\n",
+	printk(KERN_INFO "raid0 : Allocating %zu bytes for hash.\n",
 				nb_zone*sizeof(struct strip_zone*));
 	conf->hash_table = kmalloc (sizeof (struct strip_zone *)*nb_zone, GFP_KERNEL);
 	if (!conf->hash_table)
 		goto out_free_conf;
-	size = conf->strip_zone[cur].size;
+	sectors = conf->strip_zone[cur].sectors;
 
 	conf->hash_table[0] = conf->strip_zone + cur;
 	for (i=1; i< nb_zone; i++) {
-		while (size <= conf->hash_spacing) {
+		while (sectors <= conf->spacing) {
 			cur++;
-			size += conf->strip_zone[cur].size;
+			sectors += conf->strip_zone[cur].sectors;
 		}
-		size -= conf->hash_spacing;
+		sectors -= conf->spacing;
 		conf->hash_table[i] = conf->strip_zone + cur;
 	}
-	if (conf->preshift) {
-		conf->hash_spacing >>= conf->preshift;
-		/* round hash_spacing up so when we divide by it, we
+	if (conf->sector_shift) {
+		conf->spacing >>= conf->sector_shift;
+		/* round spacing up so when we divide by it, we
 		 * err on the side of too-low, which is safest
 		 */
-		conf->hash_spacing++;
+		conf->spacing++;
 	}
 
 	/* calculate the max read-ahead size.
@@ -387,12 +388,12 @@ static int raid0_stop (mddev_t *mddev)
 static int raid0_make_request (struct request_queue *q, struct bio *bio)
 {
 	mddev_t *mddev = q->queuedata;
-	unsigned int sect_in_chunk, chunksize_bits,  chunk_size, chunk_sects;
+	unsigned int sect_in_chunk, chunksect_bits, chunk_sects;
 	raid0_conf_t *conf = mddev_to_conf(mddev);
 	struct strip_zone *zone;
 	mdk_rdev_t *tmp_dev;
 	sector_t chunk;
-	sector_t block, rsect;
+	sector_t sector, rsect;
 	const int rw = bio_data_dir(bio);
 	int cpu;
 
@@ -407,11 +408,9 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio)
 		      bio_sectors(bio));
 	part_stat_unlock();
 
-	chunk_size = mddev->chunk_size >> 10;
 	chunk_sects = mddev->chunk_size >> 9;
-	chunksize_bits = ffz(~chunk_size);
-	block = bio->bi_sector >> 1;
-	
+	chunksect_bits = ffz(~chunk_sects);
+	sector = bio->bi_sector;
 
 	if (unlikely(chunk_sects < (bio->bi_sector & (chunk_sects - 1)) + (bio->bi_size >> 9))) {
 		struct bio_pair *bp;
@@ -434,28 +433,27 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio)
  
 
 	{
-		sector_t x = block >> conf->preshift;
-		sector_div(x, (u32)conf->hash_spacing);
+		sector_t x = sector >> conf->sector_shift;
+		sector_div(x, (u32)conf->spacing);
 		zone = conf->hash_table[x];
 	}
 
-	while (block >= (zone->zone_offset + zone->size)) 
+	while (sector >= zone->zone_start + zone->sectors)
 		zone++;
 
-	sect_in_chunk = bio->bi_sector & ((chunk_size<<1) -1);
+	sect_in_chunk = bio->bi_sector & (chunk_sects - 1);
 
 
 	{
-		sector_t x =  (block - zone->zone_offset) >> chunksize_bits;
+		sector_t x = (sector - zone->zone_start) >> chunksect_bits;
 
 		sector_div(x, zone->nb_dev);
 		chunk = x;
 
-		x = block >> chunksize_bits;
+		x = sector >> chunksect_bits;
 		tmp_dev = zone->dev[sector_div(x, zone->nb_dev)];
 	}
-	rsect = (((chunk << chunksize_bits) + zone->dev_offset)<<1)
-		+ sect_in_chunk;
+	rsect = (chunk << chunksect_bits) + zone->dev_start + sect_in_chunk;
  
 	bio->bi_bdev = tmp_dev->bdev;
 	bio->bi_sector = rsect + tmp_dev->data_offset;
@@ -467,7 +465,7 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio)
 
 bad_map:
 	printk("raid0_make_request bug: can't convert block across chunks"
-		" or bigger than %dk %llu %d\n", chunk_size, 
+		" or bigger than %dk %llu %d\n", chunk_sects / 2,
 		(unsigned long long)bio->bi_sector, bio->bi_size >> 10);
 
 	bio_io_error(bio);
@@ -492,10 +490,10 @@ static void raid0_status (struct seq_file *seq, mddev_t *mddev)
 			seq_printf(seq, "%s/", bdevname(
 				conf->strip_zone[j].dev[k]->bdev,b));
 
-		seq_printf(seq, "] zo=%d do=%d s=%d\n",
-				conf->strip_zone[j].zone_offset,
-				conf->strip_zone[j].dev_offset,
-				conf->strip_zone[j].size);
+		seq_printf(seq, "] zs=%d ds=%d s=%d\n",
+				conf->strip_zone[j].zone_start,
+				conf->strip_zone[j].dev_start,
+				conf->strip_zone[j].sectors);
 	}
 #endif
 	seq_printf(seq, " %dk chunks", mddev->chunk_size/1024);

drivers/md/raid1.c

@@ -1016,12 +1016,16 @@ static void error(mddev_t *mddev, mdk_rdev_t *rdev)
 	 * else mark the drive as failed
 	 */
 	if (test_bit(In_sync, &rdev->flags)
-	    && (conf->raid_disks - mddev->degraded) == 1)
+	    && (conf->raid_disks - mddev->degraded) == 1) {
 		/*
 		 * Don't fail the drive, act as though we were just a
-		 * normal single drive
+		 * normal single drive.
+		 * However don't try a recovery from this drive as
+		 * it is very likely to fail.
 		 */
+		mddev->recovery_disabled = 1;
 		return;
+	}
 	if (test_and_clear_bit(In_sync, &rdev->flags)) {
 		unsigned long flags;
 		spin_lock_irqsave(&conf->device_lock, flags);
@@ -1919,7 +1923,6 @@ static int run(mddev_t *mddev)
 	int i, j, disk_idx;
 	mirror_info_t *disk;
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 
 	if (mddev->level != 1) {
 		printk("raid1: %s: raid level not set to mirroring (%d)\n",
@@ -1964,7 +1967,7 @@ static int run(mddev_t *mddev)
 	spin_lock_init(&conf->device_lock);
 	mddev->queue->queue_lock = &conf->device_lock;
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		disk_idx = rdev->raid_disk;
 		if (disk_idx >= mddev->raid_disks
 		    || disk_idx < 0)

drivers/md/raid10.c

@@ -2025,7 +2025,6 @@ static int run(mddev_t *mddev)
 	int i, disk_idx;
 	mirror_info_t *disk;
 	mdk_rdev_t *rdev;
-	struct list_head *tmp;
 	int nc, fc, fo;
 	sector_t stride, size;
 
@@ -2108,7 +2107,7 @@ static int run(mddev_t *mddev)
 	spin_lock_init(&conf->device_lock);
 	mddev->queue->queue_lock = &conf->device_lock;
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		disk_idx = rdev->raid_disk;
 		if (disk_idx >= mddev->raid_disks
 		    || disk_idx < 0)

drivers/md/raid5.c

@@ -3998,7 +3998,6 @@ static int run(mddev_t *mddev)
 	int raid_disk, memory;
 	mdk_rdev_t *rdev;
 	struct disk_info *disk;
-	struct list_head *tmp;
 	int working_disks = 0;
 
 	if (mddev->level != 5 && mddev->level != 4 && mddev->level != 6) {
@@ -4108,7 +4107,7 @@ static int run(mddev_t *mddev)
 
 	pr_debug("raid5: run(%s) called.\n", mdname(mddev));
 
-	rdev_for_each(rdev, tmp, mddev) {
+	list_for_each_entry(rdev, &mddev->disks, same_set) {
 		raid_disk = rdev->raid_disk;
 		if (raid_disk >= conf->raid_disks
 		    || raid_disk < 0)
@@ -4533,7 +4532,6 @@ static int raid5_start_reshape(mddev_t *mddev)
 {
 	raid5_conf_t *conf = mddev_to_conf(mddev);
 	mdk_rdev_t *rdev;
-	struct list_head *rtmp;
 	int spares = 0;
 	int added_devices = 0;
 	unsigned long flags;
@@ -4541,7 +4539,7 @@ static int raid5_start_reshape(mddev_t *mddev)
 	if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))
 		return -EBUSY;
 
-	rdev_for_each(rdev, rtmp, mddev)
+	list_for_each_entry(rdev, &mddev->disks, same_set)
 		if (rdev->raid_disk < 0 &&
 		    !test_bit(Faulty, &rdev->flags))
 			spares++;
@@ -4563,7 +4561,7 @@ static int raid5_start_reshape(mddev_t *mddev)
 	/* Add some new drives, as many as will fit.
 	 * We know there are enough to make the newly sized array work.
 	 */
-	rdev_for_each(rdev, rtmp, mddev)
+	list_for_each_entry(rdev, &mddev->disks, same_set)
 		if (rdev->raid_disk < 0 &&
 		    !test_bit(Faulty, &rdev->flags)) {
 			if (raid5_add_disk(mddev, rdev) == 0) {

fs/block_dev.c

@@ -1005,6 +1005,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
 	}
 
 	lock_kernel();
+ restart:
 
 	ret = -ENXIO;
 	disk = get_gendisk(bdev->bd_dev, &partno);
@@ -1025,6 +1026,19 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part)
 
 			if (disk->fops->open) {
 				ret = disk->fops->open(bdev, mode);
+				if (ret == -ERESTARTSYS) {
+					/* Lost a race with 'disk' being
+					 * deleted, try again.
+					 * See md.c
+					 */
+					disk_put_part(bdev->bd_part);
+					bdev->bd_part = NULL;
+					module_put(disk->fops->owner);
+					put_disk(disk);
+					bdev->bd_disk = NULL;
+					mutex_unlock(&bdev->bd_mutex);
+					goto restart;
+				}
 				if (ret)
 					goto out_clear;
 			}

include/linux/raid/md_k.h

@@ -137,6 +137,9 @@ struct mddev_s
 	struct gendisk			*gendisk;
 
 	struct kobject			kobj;
+	int				hold_active;
+#define	UNTIL_IOCTL	1
+#define	UNTIL_STOP	2
 
 	/* Superblock information */
 	int				major_version,
@@ -215,6 +218,9 @@ struct mddev_s
 #define	MD_RECOVERY_FROZEN	9
 
 	unsigned long			recovery;
+	int				recovery_disabled; /* if we detect that recovery
+							    * will always fail, set this
+							    * so we don't loop trying */
 
 	int				in_sync;	/* know to not need resync */
 	struct mutex			reconfig_mutex;
@@ -244,6 +250,9 @@ struct mddev_s
 	struct sysfs_dirent		*sysfs_state;	/* handle for 'array_state'
 							 * file in sysfs.
 							 */
+	struct sysfs_dirent		*sysfs_action;  /* handle for 'sync_action' */
+
+	struct work_struct del_work;	/* used for delayed sysfs removal */
 
 	spinlock_t			write_lock;
 	wait_queue_head_t		sb_wait;	/* for waiting on superblock updates */
@@ -334,17 +343,14 @@ static inline char * mdname (mddev_t * mddev)
  * iterates through some rdev ringlist. It's safe to remove the
  * current 'rdev'. Dont touch 'tmp' though.
  */
-#define rdev_for_each_list(rdev, tmp, list)				\
-									\
-	for ((tmp) = (list).next;					\
-		(rdev) = (list_entry((tmp), mdk_rdev_t, same_set)),	\
-			(tmp) = (tmp)->next, (tmp)->prev != &(list)	\
-		; )
+#define rdev_for_each_list(rdev, tmp, head)				\
+	list_for_each_entry_safe(rdev, tmp, head, same_set)
+
 /*
  * iterates through the 'same array disks' ringlist
  */
 #define rdev_for_each(rdev, tmp, mddev)				\
-	rdev_for_each_list(rdev, tmp, (mddev)->disks)
+	list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set)
 
 #define rdev_for_each_rcu(rdev, mddev)				\
 	list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set)

include/linux/raid/md_p.h

@@ -194,6 +194,8 @@ static inline __u64 md_event(mdp_super_t *sb) {
 	return (ev<<32)| sb->events_lo;
 }
 
+#define MD_SUPERBLOCK_1_TIME_SEC_MASK ((1ULL<<40) - 1)
+
 /*
  * The version-1 superblock :
  * All numeric fields are little-endian.

include/linux/raid/raid0.h

@@ -5,9 +5,9 @@
 
 struct strip_zone
 {
-	sector_t zone_offset;	/* Zone offset in md_dev */
-	sector_t dev_offset;	/* Zone offset in real dev */
-	sector_t size;		/* Zone size */
+	sector_t zone_start;	/* Zone offset in md_dev (in sectors) */
+	sector_t dev_start;	/* Zone offset in real dev (in sectors) */
+	sector_t sectors;	/* Zone size in sectors */
 	int nb_dev;		/* # of devices attached to the zone */
 	mdk_rdev_t **dev;	/* Devices attached to the zone */
 };
@@ -19,8 +19,8 @@ struct raid0_private_data
 	mdk_rdev_t **devlist; /* lists of rdevs, pointed to by strip_zone->dev */
 	int nr_strip_zones;
 
-	sector_t hash_spacing;
-	int preshift;			/* shift this before divide by hash_spacing */
+	sector_t spacing;
+	int sector_shift; /* shift this before divide by spacing */
 };
 
 typedef struct raid0_private_data raid0_conf_t;