1 /* SPDX-License-Identifier: GPL-2.0-or-later */
4           Copyright (C) 1996-98 Ingo Molnar, Gadi Oxman
12 #include <linux/backing-dev.h>
22 #include "md-cluster.h"
37 /* Status of sync thread. */
41 	 * 1) after assemble, sync data from first rdev to other copies, this
42 	 * must be done first before other sync actions and will only execute
44 	 * 2) resize the array(notice that this is not reshape), sync data for
50 	 * 1) for new replacement, sync data based on the replace rdev or
52 	 * 2) for new member disk while the array is degraded, sync data from
54 	 * 3) reassemble after power failure or re-add a hot removed rdev, sync
81 	 * or internal usage like setting the array read-only, will forbid above
112 	 * If meta_bdev is non-NULL, it means that a separate device is
135 	 * in sync with array
148 					 * the array after a level-change completes.
201 				 * bitmap-based recovery to get fully in sync.
203 				 * has been passed to pers->hot_add_disk.
206 	AutoDetected,		/* added by auto-detect */
215 				 * Blocked.  The Fault is/will-be
221 				 * found an unacknowledged bad-block.
223 				 * time, and the writer will re-check.
226 				 * re-check.  So setting it as
231 				 * hot-replaced, either because it has
244 				 * raid-5/6.
266 	Nonrot,			/* non-rotational device (SSD) */
272 	if (unlikely(rdev->badblocks.count)) {  in is_badblock()
273 		int rv = badblocks_check(&rdev->badblocks, rdev->data_offset + s,  in is_badblock()
277 			*first_bad -= rdev->data_offset;  in is_badblock()
299  * enum mddev_flags - md device flags.
311  * @MD_NOT_READY: do_md_run() is active, so 'array_state', ust not report that
335 	MD_SB_CHANGE_PENDING,	/* switch from 'clean' to 'active' in progress */
349  * mddev->curr_resync stores the current sector of the resync but
357 	/* Delayed to check that there is no conflict with another sync */
359 	/* Any value greater than or equal to this is in an active resync */
407 	/* If the last 'event' was simply a clean->dirty transition, and
409 	 * to just decrement the event count on a dirty->clean transition.
430 	 * Set when a sync operation is started. It holds this value even
431 	 * when the sync thread is "frozen" (interrupted) or "idle" (stopped
432 	 * or finished). It is overwritten when a new sync operation is begun.
453 	/* allow user-space to request suspension of IO to regions of the array */
456 	/* if zero, use the system-wide default */
460 	/* resync even though the same disks are shared among md-devices */
478 	 * with disk->open_mutex.
480 	 *  reconfig_mutex -> disk->open_mutex
481 	 *  disk->open_mutex -> open_mutex:  e.g. __blkdev_get -> md_open
485 	atomic_t			active;		/* general refcount */  member
486 	atomic_t			openers;	/* number of active opens */
497 	sector_t			resync_min;	/* user requested sync
512 	/* used for register new sync thread */
519 	 *   in_sync - and related safemode and MD_CHANGE changes
521 	 *   clearing ->bitmap
522 	 *   clearing ->bitmap_info.file
523 	 *   changing ->resync_{min,max}
528 	atomic_t			pending_writes;	/* number of active superblock writes */
550 							 * hot-adding a bitmap.  It should
558 		unsigned long		max_write_behind; /* write-behind mode */
570 	struct bio_set			sync_set; /* for sync operations like
588 	/* The sequence number for sync thread */
597 	/* flags for sync thread running status */
600 	 * set when one of sync action is set and new sync thread need to be
604 	/* sync thread is running, or about to be started */
606 	/* sync thread needs to be aborted for some reason */
608 	/* sync thread is done and is waiting to be unregistered */
610 	/* running sync thread must abort immediately, and not restart */
612 	/* waiting for pers->start() to finish */
614 	/* interrupted because io-error */
617 	/* flags determines sync action, see details in enum sync_action */
648 	return (mddev->ro == MD_RDWR);  in md_is_rdwr()
654 	if (mddev->reshape_position == MaxSector)  in reshape_interrupted()
658 	if (!test_bit(MD_RECOVERY_RUNNING, &mddev->recovery))  in reshape_interrupted()
662 	if (test_bit(MD_RECOVERY_WAIT, &mddev->recovery) ||  in reshape_interrupted()
663 	    test_bit(MD_RECOVERY_INTR, &mddev->recovery) ||  in reshape_interrupted()
664 	    test_bit(MD_RECOVERY_FROZEN, &mddev->recovery))  in reshape_interrupted()
672 	return mutex_lock_interruptible(&mddev->reconfig_mutex);  in mddev_lock()
680 	mutex_lock(&mddev->reconfig_mutex);  in mddev_lock_nointr()
685 	return mutex_trylock(&mddev->reconfig_mutex);  in mddev_trylock()
691 	if (blk_queue_io_stat(bdev->bd_disk->queue))  in md_sync_acct()
692 		atomic_add(nr_sectors, &bdev->bd_disk->sync_io);  in md_sync_acct()
697 	md_sync_acct(bio->bi_bdev, nr_sectors);  in md_sync_acct_bio()
716 	/* error_handler must set ->faulty and clear ->in_sync
733 	 * 1 - stop new actions and wait for action io to complete
734 	 * 0 - return to normal behaviour
740 	 * e.g. 2drive raid1 -> 2drive raid5
741 	 *      ndrive raid5 -> degraded n+1drive raid6 with special layout
743 	 * This needs to be installed and then ->run used to activate the
747 	/* Changes the consistency policy of an active array. */
772 	return mddev->gendisk ? mddev->gendisk->disk_name : "mdX";  in mdname()
778 	if (!test_bit(Replacement, &rdev->flags) &&  in sysfs_link_rdev()
779 	    !test_bit(Journal, &rdev->flags) &&  in sysfs_link_rdev()
780 	    mddev->kobj.sd) {  in sysfs_link_rdev()
781 		sprintf(nm, "rd%d", rdev->raid_disk);  in sysfs_link_rdev()
782 		return sysfs_create_link(&mddev->kobj, &rdev->kobj, nm);  in sysfs_link_rdev()
790 	if (!test_bit(Replacement, &rdev->flags) &&  in sysfs_unlink_rdev()
791 	    !test_bit(Journal, &rdev->flags) &&  in sysfs_unlink_rdev()
792 	    mddev->kobj.sd) {  in sysfs_unlink_rdev()
793 		sprintf(nm, "rd%d", rdev->raid_disk);  in sysfs_unlink_rdev()
794 		sysfs_remove_link(&mddev->kobj, nm);  in sysfs_unlink_rdev()
809 	list_for_each_entry(rdev, &((mddev)->disks), same_set)
812 	list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set)
815 	list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set)
914 	return !disk_live(rdev->bdev->bd_disk);  in is_rdev_broken()
919 	int faulty = test_bit(Faulty, &rdev->flags);  in rdev_dec_pending()
920 	if (atomic_dec_and_test(&rdev->nr_pending) && faulty) {  in rdev_dec_pending()
921 		set_bit(MD_RECOVERY_NEEDED, &mddev->recovery);  in rdev_dec_pending()
922 		md_wakeup_thread(mddev->thread);  in rdev_dec_pending()
929 	return mddev->cluster_info && mddev->bitmap_info.nodes > 1;  in mddev_is_clustered()
936 	mddev->flags &= ~unsupported_flags;  in mddev_clear_unsupported_flags()
942 	    !bio->bi_bdev->bd_disk->queue->limits.max_write_zeroes_sectors)  in mddev_check_write_zeroes()
943 		mddev->gendisk->queue->limits.max_write_zeroes_sectors = 0;  in mddev_check_write_zeroes()
964 	mutex_lock(&mddev->reconfig_mutex);  in mddev_suspend_and_lock_nointr()
991  * MD devices can be used undeneath by DM, in which case ->gendisk is NULL.
995 	return !mddev->gendisk;  in mddev_is_dm()
1002 		trace_block_bio_remap(bio, disk_devt(mddev->gendisk), sector);  in mddev_trace_remap()
1008 		blk_add_trace_msg((mddev)->gendisk->queue, fmt, ##args); \