Merge branch 'block-6.17' into for-next

* block-6.17:
  block: tone down bio_check_eod
  loop: use vfs_getattr_nosec for accurate file size
  loop: Consolidate size calculation logic into lo_calculate_size()
  block: remove newlines from the warnings in blk_validate_integrity_limits
  block: handle pi_tuple_size in queue_limits_stack_integrity
  selftests: ublk: Use ARRAY_SIZE() macro to improve code
  md: fix sync_action incorrect display during resync
  md: add helper rdev_needs_recovery()
  md: keep recovery_cp in mdp_superblock_s
  md: add legacy_async_del_gendisk mode

Author: axboe

block/blk-core.c

@@ -557,7 +557,7 @@ static inline int bio_check_eod(struct bio *bio)
 	sector_t maxsector = bdev_nr_sectors(bio->bi_bdev);
 	unsigned int nr_sectors = bio_sectors(bio);
 
-	if (nr_sectors &&
+	if (nr_sectors && maxsector &&
 	    (nr_sectors > maxsector ||
 	     bio->bi_iter.bi_sector > maxsector - nr_sectors)) {
 		pr_info_ratelimited("%s: attempt to access beyond end of device\n"

block/blk-settings.c

@@ -157,25 +157,22 @@ static int blk_validate_integrity_limits(struct queue_limits *lim)
 	switch (bi->csum_type) {
 	case BLK_INTEGRITY_CSUM_NONE:
 		if (bi->pi_tuple_size) {
-			pr_warn("pi_tuple_size must be 0 when checksum type \
-				 is none\n");
+			pr_warn("pi_tuple_size must be 0 when checksum type is none\n");
 			return -EINVAL;
 		}
 		break;
 	case BLK_INTEGRITY_CSUM_CRC:
 	case BLK_INTEGRITY_CSUM_IP:
 		if (bi->pi_tuple_size != sizeof(struct t10_pi_tuple)) {
-			pr_warn("pi_tuple_size mismatch for T10 PI: expected \
-				 %zu, got %u\n",
+			pr_warn("pi_tuple_size mismatch for T10 PI: expected %zu, got %u\n",
 				 sizeof(struct t10_pi_tuple),
 				 bi->pi_tuple_size);
 			return -EINVAL;
 		}
 		break;
 	case BLK_INTEGRITY_CSUM_CRC64:
 		if (bi->pi_tuple_size != sizeof(struct crc64_pi_tuple)) {
-			pr_warn("pi_tuple_size mismatch for CRC64 PI: \
-				 expected %zu, got %u\n",
+			pr_warn("pi_tuple_size mismatch for CRC64 PI: expected %zu, got %u\n",
 				 sizeof(struct crc64_pi_tuple),
 				 bi->pi_tuple_size);
 			return -EINVAL;
@@ -972,6 +969,8 @@ bool queue_limits_stack_integrity(struct queue_limits *t,
 			goto incompatible;
 		if (ti->csum_type != bi->csum_type)
 			goto incompatible;
+		if (ti->pi_tuple_size != bi->pi_tuple_size)
+			goto incompatible;
 		if ((ti->flags & BLK_INTEGRITY_REF_TAG) !=
 		    (bi->flags & BLK_INTEGRITY_REF_TAG))
 			goto incompatible;
@@ -980,6 +979,7 @@ bool queue_limits_stack_integrity(struct queue_limits *t,
 		ti->flags |= (bi->flags & BLK_INTEGRITY_DEVICE_CAPABLE) |
 			     (bi->flags & BLK_INTEGRITY_REF_TAG);
 		ti->csum_type = bi->csum_type;
+		ti->pi_tuple_size = bi->pi_tuple_size;
 		ti->metadata_size = bi->metadata_size;
 		ti->pi_offset = bi->pi_offset;
 		ti->interval_exp = bi->interval_exp;

drivers/block/loop.c

@@ -137,32 +137,36 @@ static void loop_global_unlock(struct loop_device *lo, bool global)
 static int max_part;
 static int part_shift;
 
-static loff_t get_size(loff_t offset, loff_t sizelimit, struct file *file)
+static loff_t lo_calculate_size(struct loop_device *lo, struct file *file)
 {
+	struct kstat stat;
 	loff_t loopsize;
+	int ret;
 
-	/* Compute loopsize in bytes */
-	loopsize = i_size_read(file->f_mapping->host);
-	if (offset > 0)
-		loopsize -= offset;
+	/*
+	 * Get the accurate file size. This provides better results than
+	 * cached inode data, particularly for network filesystems where
+	 * metadata may be stale.
+	 */
+	ret = vfs_getattr_nosec(&file->f_path, &stat, STATX_SIZE, 0);
+	if (ret)
+		return 0;
+
+	loopsize = stat.size;
+	if (lo->lo_offset > 0)
+		loopsize -= lo->lo_offset;
 	/* offset is beyond i_size, weird but possible */
 	if (loopsize < 0)
 		return 0;
-
-	if (sizelimit > 0 && sizelimit < loopsize)
-		loopsize = sizelimit;
+	if (lo->lo_sizelimit > 0 && lo->lo_sizelimit < loopsize)
+		loopsize = lo->lo_sizelimit;
 	/*
 	 * Unfortunately, if we want to do I/O on the device,
 	 * the number of 512-byte sectors has to fit into a sector_t.
 	 */
 	return loopsize >> 9;
 }
 
-static loff_t get_loop_size(struct loop_device *lo, struct file *file)
-{
-	return get_size(lo->lo_offset, lo->lo_sizelimit, file);
-}
-
 /*
  * We support direct I/O only if lo_offset is aligned with the logical I/O size
  * of backing device, and the logical block size of loop is bigger than that of
@@ -569,7 +573,7 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev,
 	error = -EINVAL;
 
 	/* size of the new backing store needs to be the same */
-	if (get_loop_size(lo, file) != get_loop_size(lo, old_file))
+	if (lo_calculate_size(lo, file) != lo_calculate_size(lo, old_file))
 		goto out_err;
 
 	/*
@@ -1063,7 +1067,7 @@ static int loop_configure(struct loop_device *lo, blk_mode_t mode,
 	loop_update_dio(lo);
 	loop_sysfs_init(lo);
 
-	size = get_loop_size(lo, file);
+	size = lo_calculate_size(lo, file);
 	loop_set_size(lo, size);
 
 	/* Order wrt reading lo_state in loop_validate_file(). */
@@ -1255,8 +1259,7 @@ loop_set_status(struct loop_device *lo, const struct loop_info64 *info)
 	if (partscan)
 		clear_bit(GD_SUPPRESS_PART_SCAN, &lo->lo_disk->state);
 	if (!err && size_changed) {
-		loff_t new_size = get_size(lo->lo_offset, lo->lo_sizelimit,
-					   lo->lo_backing_file);
+		loff_t new_size = lo_calculate_size(lo, lo->lo_backing_file);
 		loop_set_size(lo, new_size);
 	}
 out_unlock:
@@ -1399,7 +1402,7 @@ static int loop_set_capacity(struct loop_device *lo)
 	if (unlikely(lo->lo_state != Lo_bound))
 		return -ENXIO;
 
-	size = get_loop_size(lo, lo->lo_backing_file);
+	size = lo_calculate_size(lo, lo->lo_backing_file);
 	loop_set_size(lo, size);
 
 	return 0;

drivers/md/md.c

@@ -339,6 +339,7 @@ static int start_readonly;
  * so all the races disappear.
  */
 static bool create_on_open = true;
+static bool legacy_async_del_gendisk = true;
 
 /*
  * We have a system wide 'event count' that is incremented
@@ -877,15 +878,18 @@ void mddev_unlock(struct mddev *mddev)
 		export_rdev(rdev, mddev);
 	}
 
-	/* Call del_gendisk after release reconfig_mutex to avoid
-	 * deadlock (e.g. call del_gendisk under the lock and an
-	 * access to sysfs files waits the lock)
-	 * And MD_DELETED is only used for md raid which is set in
-	 * do_md_stop. dm raid only uses md_stop to stop. So dm raid
-	 * doesn't need to check MD_DELETED when getting reconfig lock
-	 */
-	if (test_bit(MD_DELETED, &mddev->flags))
-		del_gendisk(mddev->gendisk);
+	if (!legacy_async_del_gendisk) {
+		/*
+		 * Call del_gendisk after release reconfig_mutex to avoid
+		 * deadlock (e.g. call del_gendisk under the lock and an
+		 * access to sysfs files waits the lock)
+		 * And MD_DELETED is only used for md raid which is set in
+		 * do_md_stop. dm raid only uses md_stop to stop. So dm raid
+		 * doesn't need to check MD_DELETED when getting reconfig lock
+		 */
+		if (test_bit(MD_DELETED, &mddev->flags))
+			del_gendisk(mddev->gendisk);
+	}
 }
 EXPORT_SYMBOL_GPL(mddev_unlock);
 
@@ -1419,7 +1423,7 @@ static int super_90_validate(struct mddev *mddev, struct md_rdev *freshest, stru
 		else {
 			if (sb->events_hi == sb->cp_events_hi &&
 				sb->events_lo == sb->cp_events_lo) {
-				mddev->resync_offset = sb->resync_offset;
+				mddev->resync_offset = sb->recovery_cp;
 			} else
 				mddev->resync_offset = 0;
 		}
@@ -1547,13 +1551,13 @@ static void super_90_sync(struct mddev *mddev, struct md_rdev *rdev)
 	mddev->minor_version = sb->minor_version;
 	if (mddev->in_sync)
 	{
-		sb->resync_offset = mddev->resync_offset;
+		sb->recovery_cp = mddev->resync_offset;
 		sb->cp_events_hi = (mddev->events>>32);
 		sb->cp_events_lo = (u32)mddev->events;
 		if (mddev->resync_offset == MaxSector)
 			sb->state = (1<< MD_SB_CLEAN);
 	} else
-		sb->resync_offset = 0;
+		sb->recovery_cp = 0;
 
 	sb->layout = mddev->layout;
 	sb->chunk_size = mddev->chunk_sectors << 9;
@@ -4835,9 +4839,42 @@ metadata_store(struct mddev *mddev, const char *buf, size_t len)
 static struct md_sysfs_entry md_metadata =
 __ATTR_PREALLOC(metadata_version, S_IRUGO|S_IWUSR, metadata_show, metadata_store);
 
+static bool rdev_needs_recovery(struct md_rdev *rdev, sector_t sectors)
+{
+	return rdev->raid_disk >= 0 &&
+	       !test_bit(Journal, &rdev->flags) &&
+	       !test_bit(Faulty, &rdev->flags) &&
+	       !test_bit(In_sync, &rdev->flags) &&
+	       rdev->recovery_offset < sectors;
+}
+
+static enum sync_action md_get_active_sync_action(struct mddev *mddev)
+{
+	struct md_rdev *rdev;
+	bool is_recover = false;
+
+	if (mddev->resync_offset < MaxSector)
+		return ACTION_RESYNC;
+
+	if (mddev->reshape_position != MaxSector)
+		return ACTION_RESHAPE;
+
+	rcu_read_lock();
+	rdev_for_each_rcu(rdev, mddev) {
+		if (rdev_needs_recovery(rdev, MaxSector)) {
+			is_recover = true;
+			break;
+		}
+	}
+	rcu_read_unlock();
+
+	return is_recover ? ACTION_RECOVER : ACTION_IDLE;
+}
+
 enum sync_action md_sync_action(struct mddev *mddev)
 {
 	unsigned long recovery = mddev->recovery;
+	enum sync_action active_action;
 
 	/*
 	 * frozen has the highest priority, means running sync_thread will be
@@ -4861,8 +4898,17 @@ enum sync_action md_sync_action(struct mddev *mddev)
 	    !test_bit(MD_RECOVERY_NEEDED, &recovery))
 		return ACTION_IDLE;
 
-	if (test_bit(MD_RECOVERY_RESHAPE, &recovery) ||
-	    mddev->reshape_position != MaxSector)
+	/*
+	 * Check if any sync operation (resync/recover/reshape) is
+	 * currently active. This ensures that only one sync operation
+	 * can run at a time. Returns the type of active operation, or
+	 * ACTION_IDLE if none are active.
+	 */
+	active_action = md_get_active_sync_action(mddev);
+	if (active_action != ACTION_IDLE)
+		return active_action;
+
+	if (test_bit(MD_RECOVERY_RESHAPE, &recovery))
 		return ACTION_RESHAPE;
 
 	if (test_bit(MD_RECOVERY_RECOVER, &recovery))
@@ -5818,6 +5864,13 @@ static void md_kobj_release(struct kobject *ko)
 {
 	struct mddev *mddev = container_of(ko, struct mddev, kobj);
 
+	if (legacy_async_del_gendisk) {
+		if (mddev->sysfs_state)
+			sysfs_put(mddev->sysfs_state);
+		if (mddev->sysfs_level)
+			sysfs_put(mddev->sysfs_level);
+		del_gendisk(mddev->gendisk);
+	}
 	put_disk(mddev->gendisk);
 }
 
@@ -6021,6 +6074,9 @@ static int md_alloc_and_put(dev_t dev, char *name)
 {
 	struct mddev *mddev = md_alloc(dev, name);
 
+	if (legacy_async_del_gendisk)
+		pr_warn("md: async del_gendisk mode will be removed in future, please upgrade to mdadm-4.5+\n");
+
 	if (IS_ERR(mddev))
 		return PTR_ERR(mddev);
 	mddev_put(mddev);
@@ -6431,10 +6487,22 @@ static void md_clean(struct mddev *mddev)
 	mddev->persistent = 0;
 	mddev->level = LEVEL_NONE;
 	mddev->clevel[0] = 0;
-	/* if UNTIL_STOP is set, it's cleared here */
-	mddev->hold_active = 0;
-	/* Don't clear MD_CLOSING, or mddev can be opened again. */
-	mddev->flags &= BIT_ULL_MASK(MD_CLOSING);
+
+	/*
+	 * For legacy_async_del_gendisk mode, it can stop the array in the
+	 * middle of assembling it, then it still can access the array. So
+	 * it needs to clear MD_CLOSING. If not legacy_async_del_gendisk,
+	 * it can't open the array again after stopping it. So it doesn't
+	 * clear MD_CLOSING.
+	 */
+	if (legacy_async_del_gendisk && mddev->hold_active) {
+		clear_bit(MD_CLOSING, &mddev->flags);
+	} else {
+		/* if UNTIL_STOP is set, it's cleared here */
+		mddev->hold_active = 0;
+		/* Don't clear MD_CLOSING, or mddev can be opened again. */
+		mddev->flags &= BIT_ULL_MASK(MD_CLOSING);
+	}
 	mddev->sb_flags = 0;
 	mddev->ro = MD_RDWR;
 	mddev->metadata_type[0] = 0;
@@ -6658,7 +6726,8 @@ static int do_md_stop(struct mddev *mddev, int mode)
 
 		export_array(mddev);
 		md_clean(mddev);
-		set_bit(MD_DELETED, &mddev->flags);
+		if (!legacy_async_del_gendisk)
+			set_bit(MD_DELETED, &mddev->flags);
 	}
 	md_new_event();
 	sysfs_notify_dirent_safe(mddev->sysfs_state);
@@ -8968,11 +9037,7 @@ static sector_t md_sync_position(struct mddev *mddev, enum sync_action action)
 		start = MaxSector;
 		rcu_read_lock();
 		rdev_for_each_rcu(rdev, mddev)
-			if (rdev->raid_disk >= 0 &&
-			    !test_bit(Journal, &rdev->flags) &&
-			    !test_bit(Faulty, &rdev->flags) &&
-			    !test_bit(In_sync, &rdev->flags) &&
-			    rdev->recovery_offset < start)
+			if (rdev_needs_recovery(rdev, start))
 				start = rdev->recovery_offset;
 		rcu_read_unlock();
 
@@ -9331,12 +9396,8 @@ void md_do_sync(struct md_thread *thread)
 			    test_bit(MD_RECOVERY_RECOVER, &mddev->recovery)) {
 				rcu_read_lock();
 				rdev_for_each_rcu(rdev, mddev)
-					if (rdev->raid_disk >= 0 &&
-					    mddev->delta_disks >= 0 &&
-					    !test_bit(Journal, &rdev->flags) &&
-					    !test_bit(Faulty, &rdev->flags) &&
-					    !test_bit(In_sync, &rdev->flags) &&
-					    rdev->recovery_offset < mddev->curr_resync)
+					if (mddev->delta_disks >= 0 &&
+					    rdev_needs_recovery(rdev, mddev->curr_resync))
 						rdev->recovery_offset = mddev->curr_resync;
 				rcu_read_unlock();
 			}
@@ -10392,6 +10453,7 @@ 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);
 module_param(create_on_open, bool, S_IRUSR|S_IWUSR);
+module_param(legacy_async_del_gendisk, bool, 0600);
 
 MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("MD RAID framework");

include/uapi/linux/raid/md_p.h

@@ -173,7 +173,7 @@ typedef struct mdp_superblock_s {
 #else
 #error unspecified endianness
 #endif
-	__u32 resync_offset;	/* 11 resync checkpoint sector count	      */
+	__u32 recovery_cp;	/* 11 resync checkpoint sector count	      */
 	/* There are only valid for minor_version > 90 */
 	__u64 reshape_position;	/* 12,13 next address in array-space for reshape */
 	__u32 new_level;	/* 14 new level we are reshaping to	      */

tools/testing/selftests/ublk/kublk.c

@@ -1400,7 +1400,7 @@ static int cmd_dev_get_features(void)
 
 			if (!((1ULL << i)  & features))
 				continue;
-			if (i < sizeof(feat_map) / sizeof(feat_map[0]))
+			if (i < ARRAY_SIZE(feat_map))
 				feat = feat_map[i];
 			else
 				feat = "unknown";
@@ -1477,7 +1477,7 @@ static void __cmd_create_help(char *exe, bool recovery)
 	printf("\tdefault: nr_queues=2(max 32), depth=128(max 1024), dev_id=-1(auto allocation)\n");
 	printf("\tdefault: nthreads=nr_queues");
 
-	for (i = 0; i < sizeof(tgt_ops_list) / sizeof(tgt_ops_list[0]); i++) {
+	for (i = 0; i < ARRAY_SIZE(tgt_ops_list); i++) {
 		const struct ublk_tgt_ops *ops = tgt_ops_list[i];
 
 		if (ops->usage)