INIT_LIST_HEAD(&new->all_mddevs);
init_timer(&new->safemode_timer);
atomic_set(&new->active, 1);
+ atomic_set(&new->openers, 0);
spin_lock_init(&new->write_lock);
init_waitqueue_head(&new->sb_wait);
init_waitqueue_head(&new->recovery_wait);
return NULL;
}
+/* return the offset of the super block in 512byte sectors */
static inline sector_t calc_dev_sboffset(struct block_device *bdev)
{
- sector_t size = bdev->bd_inode->i_size >> BLOCK_SIZE_BITS;
- return MD_NEW_SIZE_BLOCKS(size);
+ sector_t num_sectors = bdev->bd_inode->i_size / 512;
+ return MD_NEW_SIZE_SECTORS(num_sectors);
}
-static sector_t calc_dev_size(mdk_rdev_t *rdev, unsigned chunk_size)
+static sector_t calc_num_sectors(mdk_rdev_t *rdev, unsigned chunk_size)
{
- sector_t size;
-
- size = rdev->sb_offset;
+ sector_t num_sectors = rdev->sb_start;
if (chunk_size)
- size &= ~((sector_t)chunk_size/1024 - 1);
- return size;
+ num_sectors &= ~((sector_t)chunk_size/512 - 1);
+ return num_sectors;
}
static int alloc_disk_sb(mdk_rdev_t * rdev)
put_page(rdev->sb_page);
rdev->sb_loaded = 0;
rdev->sb_page = NULL;
- rdev->sb_offset = 0;
+ rdev->sb_start = 0;
rdev->size = 0;
}
}
return 0;
- if (!sync_page_io(rdev->bdev, rdev->sb_offset<<1, size, rdev->sb_page, READ))
+ if (!sync_page_io(rdev->bdev, rdev->sb_start, size, rdev->sb_page, READ))
goto fail;
rdev->sb_loaded = 1;
return 0;
int (*validate_super)(mddev_t *mddev, mdk_rdev_t *rdev);
void (*sync_super)(mddev_t *mddev, mdk_rdev_t *rdev);
unsigned long long (*rdev_size_change)(mdk_rdev_t *rdev,
- unsigned long long size);
+ sector_t num_sectors);
};
/*
char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
mdp_super_t *sb;
int ret;
- sector_t sb_offset;
/*
- * Calculate the position of the superblock,
+ * Calculate the position of the superblock (512byte sectors),
* it's at the end of the disk.
*
* It also happens to be a multiple of 4Kb.
*/
- sb_offset = calc_dev_sboffset(rdev->bdev);
- rdev->sb_offset = sb_offset;
+ rdev->sb_start = calc_dev_sboffset(rdev->bdev);
ret = read_disk_sb(rdev, MD_SB_BYTES);
if (ret) return ret;
else
ret = 0;
}
- rdev->size = calc_dev_size(rdev, sb->chunk_size);
+ rdev->size = calc_num_sectors(rdev, sb->chunk_size) / 2;
if (rdev->size < sb->size && sb->level > 1)
/* "this cannot possibly happen" ... */
* rdev_size_change for 0.90.0
*/
static unsigned long long
-super_90_rdev_size_change(mdk_rdev_t *rdev, unsigned long long size)
+super_90_rdev_size_change(mdk_rdev_t *rdev, sector_t num_sectors)
{
- if (size && size < rdev->mddev->size)
+ if (num_sectors && num_sectors < rdev->mddev->size * 2)
return 0; /* component must fit device */
- size *= 2; /* convert to sectors */
if (rdev->mddev->bitmap_offset)
return 0; /* can't move bitmap */
- rdev->sb_offset = calc_dev_sboffset(rdev->bdev);
- if (!size || size > rdev->sb_offset*2)
- size = rdev->sb_offset*2;
- md_super_write(rdev->mddev, rdev, rdev->sb_offset << 1, rdev->sb_size,
+ rdev->sb_start = calc_dev_sboffset(rdev->bdev);
+ if (!num_sectors || num_sectors > rdev->sb_start)
+ num_sectors = rdev->sb_start;
+ md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
rdev->sb_page);
md_super_wait(rdev->mddev);
- return size/2; /* kB for sysfs */
+ return num_sectors / 2; /* kB for sysfs */
}
{
struct mdp_superblock_1 *sb;
int ret;
- sector_t sb_offset;
+ sector_t sb_start;
char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE];
int bmask;
/*
- * Calculate the position of the superblock.
+ * Calculate the position of the superblock in 512byte sectors.
* It is always aligned to a 4K boundary and
* depeding on minor_version, it can be:
* 0: At least 8K, but less than 12K, from end of device
*/
switch(minor_version) {
case 0:
- sb_offset = rdev->bdev->bd_inode->i_size >> 9;
- sb_offset -= 8*2;
- sb_offset &= ~(sector_t)(4*2-1);
- /* convert from sectors to K */
- sb_offset /= 2;
+ sb_start = rdev->bdev->bd_inode->i_size >> 9;
+ sb_start -= 8*2;
+ sb_start &= ~(sector_t)(4*2-1);
break;
case 1:
- sb_offset = 0;
+ sb_start = 0;
break;
case 2:
- sb_offset = 4;
+ sb_start = 8;
break;
default:
return -EINVAL;
}
- rdev->sb_offset = sb_offset;
+ rdev->sb_start = sb_start;
/* superblock is rarely larger than 1K, but it can be larger,
* and it is safe to read 4k, so we do that
if (sb->magic != cpu_to_le32(MD_SB_MAGIC) ||
sb->major_version != cpu_to_le32(1) ||
le32_to_cpu(sb->max_dev) > (4096-256)/2 ||
- le64_to_cpu(sb->super_offset) != (rdev->sb_offset<<1) ||
+ le64_to_cpu(sb->super_offset) != rdev->sb_start ||
(le32_to_cpu(sb->feature_map) & ~MD_FEATURE_ALL) != 0)
return -EINVAL;
rdev->sb_size = (rdev->sb_size | bmask) + 1;
if (minor_version
- && rdev->data_offset < sb_offset + (rdev->sb_size/512))
+ && rdev->data_offset < sb_start + (rdev->sb_size/512))
return -EINVAL;
if (sb->level == cpu_to_le32(LEVEL_MULTIPATH))
if (minor_version)
rdev->size = ((rdev->bdev->bd_inode->i_size>>9) - le64_to_cpu(sb->data_offset)) / 2;
else
- rdev->size = rdev->sb_offset;
+ rdev->size = rdev->sb_start / 2;
if (rdev->size < le64_to_cpu(sb->data_size)/2)
return -EINVAL;
rdev->size = le64_to_cpu(sb->data_size)/2;
}
static unsigned long long
-super_1_rdev_size_change(mdk_rdev_t *rdev, unsigned long long size)
+super_1_rdev_size_change(mdk_rdev_t *rdev, sector_t num_sectors)
{
struct mdp_superblock_1 *sb;
- unsigned long long max_size;
- if (size && size < rdev->mddev->size)
+ sector_t max_sectors;
+ if (num_sectors && num_sectors < rdev->mddev->size * 2)
return 0; /* component must fit device */
- size *= 2; /* convert to sectors */
- if (rdev->sb_offset < rdev->data_offset/2) {
+ if (rdev->sb_start < rdev->data_offset) {
/* minor versions 1 and 2; superblock before data */
- max_size = (rdev->bdev->bd_inode->i_size >> 9);
- max_size -= rdev->data_offset;
- if (!size || size > max_size)
- size = max_size;
+ max_sectors = rdev->bdev->bd_inode->i_size >> 9;
+ max_sectors -= rdev->data_offset;
+ if (!num_sectors || num_sectors > max_sectors)
+ num_sectors = max_sectors;
} else if (rdev->mddev->bitmap_offset) {
/* minor version 0 with bitmap we can't move */
return 0;
} else {
/* minor version 0; superblock after data */
- sector_t sb_offset;
- sb_offset = (rdev->bdev->bd_inode->i_size >> 9) - 8*2;
- sb_offset &= ~(sector_t)(4*2 - 1);
- max_size = rdev->size*2 + sb_offset - rdev->sb_offset*2;
- if (!size || size > max_size)
- size = max_size;
- rdev->sb_offset = sb_offset/2;
+ sector_t sb_start;
+ sb_start = (rdev->bdev->bd_inode->i_size >> 9) - 8*2;
+ sb_start &= ~(sector_t)(4*2 - 1);
+ max_sectors = rdev->size * 2 + sb_start - rdev->sb_start;
+ if (!num_sectors || num_sectors > max_sectors)
+ num_sectors = max_sectors;
+ rdev->sb_start = sb_start;
}
sb = (struct mdp_superblock_1 *) page_address(rdev->sb_page);
- sb->data_size = cpu_to_le64(size);
- sb->super_offset = rdev->sb_offset*2;
+ sb->data_size = cpu_to_le64(num_sectors);
+ sb->super_offset = rdev->sb_start;
sb->sb_csum = calc_sb_1_csum(sb);
- md_super_write(rdev->mddev, rdev, rdev->sb_offset << 1, rdev->sb_size,
+ md_super_write(rdev->mddev, rdev, rdev->sb_start, rdev->sb_size,
rdev->sb_page);
md_super_wait(rdev->mddev);
- return size/2; /* kB for sysfs */
+ return num_sectors / 2; /* kB for sysfs */
}
static struct super_type super_types[] = {
static int match_mddev_units(mddev_t *mddev1, mddev_t *mddev2)
{
- struct list_head *tmp, *tmp2;
mdk_rdev_t *rdev, *rdev2;
- rdev_for_each(rdev, tmp, mddev1)
- rdev_for_each(rdev2, tmp2, mddev2)
+ rcu_read_lock();
+ rdev_for_each_rcu(rdev, mddev1)
+ rdev_for_each_rcu(rdev2, mddev2)
if (rdev->bdev->bd_contains ==
- rdev2->bdev->bd_contains)
+ rdev2->bdev->bd_contains) {
+ rcu_read_unlock();
return 1;
-
+ }
+ rcu_read_unlock();
return 0;
}
kobject_del(&rdev->kobj);
goto fail;
}
- list_add(&rdev->same_set, &mddev->disks);
+ list_add_rcu(&rdev->same_set, &mddev->disks);
bd_claim_by_disk(rdev->bdev, rdev->bdev->bd_holder, mddev->gendisk);
return 0;
return;
}
bd_release_from_disk(rdev->bdev, rdev->mddev->gendisk);
- list_del_init(&rdev->same_set);
+ list_del_rcu(&rdev->same_set);
printk(KERN_INFO "md: unbind<%s>\n", bdevname(rdev->bdev,b));
rdev->mddev = NULL;
sysfs_remove_link(&rdev->kobj, "block");
/* We need to delay this, otherwise we can deadlock when
- * writing to 'remove' to "dev/state"
+ * writing to 'remove' to "dev/state". We also need
+ * to delay it due to rcu usage.
*/
+ synchronize_rcu();
INIT_WORK(&rdev->del_work, md_delayed_delete);
kobject_get(&rdev->kobj);
schedule_work(&rdev->del_work);
if (rdev->mddev)
MD_BUG();
free_disk_sb(rdev);
- list_del_init(&rdev->same_set);
#ifndef MODULE
if (test_bit(AutoDetected, &rdev->flags))
md_autodetect_dev(rdev->bdev->bd_dev);
dprintk("%s ", bdevname(rdev->bdev,b));
if (!test_bit(Faulty, &rdev->flags)) {
md_super_write(mddev,rdev,
- rdev->sb_offset<<1, rdev->sb_size,
+ rdev->sb_start, rdev->sb_size,
rdev->sb_page);
dprintk(KERN_INFO "(write) %s's sb offset: %llu\n",
bdevname(rdev->bdev,b),
- (unsigned long long)rdev->sb_offset);
+ (unsigned long long)rdev->sb_start);
rdev->sb_events = mddev->events;
} else
static ssize_t
rdev_size_store(mdk_rdev_t *rdev, const char *buf, size_t len)
{
- char *e;
- unsigned long long size = simple_strtoull(buf, &e, 10);
+ unsigned long long size;
unsigned long long oldsize = rdev->size;
mddev_t *my_mddev = rdev->mddev;
- if (e==buf || (*e && *e != '\n'))
+ if (strict_strtoull(buf, 10, &size) < 0)
+ return -EINVAL;
+ if (size < my_mddev->size)
return -EINVAL;
if (my_mddev->pers && rdev->raid_disk >= 0) {
- if (rdev->mddev->persistent) {
- size = super_types[rdev->mddev->major_version].
- rdev_size_change(rdev, size);
+ if (my_mddev->persistent) {
+ size = super_types[my_mddev->major_version].
+ rdev_size_change(rdev, size * 2);
if (!size)
return -EBUSY;
} else if (!size) {
size = (rdev->bdev->bd_inode->i_size >> 10);
size -= rdev->data_offset/2;
}
- if (size < rdev->mddev->size)
+ if (size < my_mddev->size)
return -EINVAL; /* component must fit device */
}
rdev->size = size;
- if (size > oldsize && rdev->mddev->external) {
+ if (size > oldsize && my_mddev->external) {
/* need to check that all other rdevs with the same ->bdev
* do not overlap. We need to unlock the mddev to avoid
* a deadlock. We have already changed rdev->size, and if
if (test_bit(AllReserved, &rdev2->flags) ||
(rdev->bdev == rdev2->bdev &&
rdev != rdev2 &&
- overlaps(rdev->data_offset, rdev->size,
- rdev2->data_offset, rdev2->size))) {
+ overlaps(rdev->data_offset, rdev->size * 2,
+ rdev2->data_offset,
+ rdev2->size * 2))) {
overlap = 1;
break;
}
return -EBUSY;
}
}
- if (size < my_mddev->size || my_mddev->size == 0)
- my_mddev->size = size;
return len;
}
break;
case clear:
/* stopping an active array */
- if (atomic_read(&mddev->active) > 1)
+ if (atomic_read(&mddev->openers) > 0)
return -EBUSY;
err = do_md_stop(mddev, 0, 0);
break;
case inactive:
/* stopping an active array */
if (mddev->pers) {
- if (atomic_read(&mddev->active) > 1)
+ if (atomic_read(&mddev->openers) > 0)
return -EBUSY;
err = do_md_stop(mddev, 2, 0);
} else
* We don't want the data to overlap the metadata,
* Internal Bitmap issues has handled elsewhere.
*/
- if (rdev->data_offset < rdev->sb_offset) {
+ if (rdev->data_offset < rdev->sb_start) {
if (mddev->size &&
rdev->data_offset + mddev->size*2
- > rdev->sb_offset*2) {
+ > rdev->sb_start) {
printk("md: %s: data overlaps metadata\n",
mdname(mddev));
return -EINVAL;
}
} else {
- if (rdev->sb_offset*2 + rdev->sb_size/512
+ if (rdev->sb_start + rdev->sb_size/512
> rdev->data_offset) {
printk("md: %s: metadata overlaps data\n",
mdname(mddev));
if (mddev->flags)
md_update_sb(mddev, 0);
- set_capacity(disk, mddev->array_size<<1);
+ set_capacity(disk, mddev->array_sectors);
/* If we call blk_queue_make_request here, it will
* re-initialise max_sectors etc which may have been
int err = 0;
struct gendisk *disk = mddev->gendisk;
- if (atomic_read(&mddev->active) > 1 + is_open) {
+ if (atomic_read(&mddev->openers) > is_open) {
printk("md: %s still in use.\n",mdname(mddev));
return -EBUSY;
}
export_array(mddev);
- mddev->array_size = 0;
+ mddev->array_sectors = 0;
mddev->size = 0;
mddev->raid_disks = 0;
mddev->recovery_cp = 0;
/* 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);
+ }
mddev_put(mddev);
}
printk(KERN_INFO "md: ... autorun DONE.\n");
if (!mddev->persistent) {
printk(KERN_INFO "md: nonpersistent superblock ...\n");
- rdev->sb_offset = rdev->bdev->bd_inode->i_size >> BLOCK_SIZE_BITS;
+ rdev->sb_start = rdev->bdev->bd_inode->i_size / 512;
} else
- rdev->sb_offset = calc_dev_sboffset(rdev->bdev);
- rdev->size = calc_dev_size(rdev, mddev->chunk_size);
+ rdev->sb_start = calc_dev_sboffset(rdev->bdev);
+ rdev->size = calc_num_sectors(rdev, mddev->chunk_size) / 2;
err = bind_rdev_to_array(rdev, mddev);
if (err) {
{
char b[BDEVNAME_SIZE];
int err;
- unsigned int size;
mdk_rdev_t *rdev;
if (!mddev->pers)
}
if (mddev->persistent)
- rdev->sb_offset = calc_dev_sboffset(rdev->bdev);
+ rdev->sb_start = calc_dev_sboffset(rdev->bdev);
else
- rdev->sb_offset =
- rdev->bdev->bd_inode->i_size >> BLOCK_SIZE_BITS;
+ rdev->sb_start = rdev->bdev->bd_inode->i_size / 512;
- size = calc_dev_size(rdev, mddev->chunk_size);
- rdev->size = size;
+ rdev->size = calc_num_sectors(rdev, mddev->chunk_size) / 2;
if (test_bit(Faulty, &rdev->flags)) {
printk(KERN_WARNING
* linear and raid0 always use whatever space is available. We can only
* consider changing this number if no resync or reconstruction is
* happening, and if the new size is acceptable. It must fit before the
- * sb_offset or, if that is <data_offset, it must fit before the size
+ * sb_start or, if that is <data_offset, it must fit before the size
* of each device. If num_sectors is zero, we find the largest size
* that fits.
bdev = bdget_disk(mddev->gendisk, 0);
if (bdev) {
mutex_lock(&bdev->bd_inode->i_mutex);
- i_size_write(bdev->bd_inode, (loff_t)mddev->array_size << 10);
+ i_size_write(bdev->bd_inode,
+ (loff_t)mddev->array_sectors << 9);
mutex_unlock(&bdev->bd_inode->i_mutex);
bdput(bdev);
}
err = 0;
mddev_get(mddev);
+ atomic_inc(&mddev->openers);
mddev_unlock(mddev);
check_disk_change(inode->i_bdev);
mddev_t *mddev = inode->i_bdev->bd_disk->private_data;
BUG_ON(!mddev);
+ atomic_dec(&mddev->openers);
mddev_put(mddev);
return 0;
if (!list_empty(&mddev->disks)) {
if (mddev->pers)
seq_printf(seq, "\n %llu blocks",
- (unsigned long long)mddev->array_size);
+ (unsigned long long)
+ mddev->array_sectors / 2);
else
seq_printf(seq, "\n %llu blocks",
- (unsigned long long)size);
+ (unsigned long long)size);
}
if (mddev->persistent) {
if (mddev->major_version != 0 ||
static int is_mddev_idle(mddev_t *mddev)
{
mdk_rdev_t * rdev;
- struct list_head *tmp;
int idle;
long curr_events;
idle = 1;
- rdev_for_each(rdev, tmp, mddev) {
+ rcu_read_lock();
+ rdev_for_each_rcu(rdev, mddev) {
struct gendisk *disk = rdev->bdev->bd_contains->bd_disk;
curr_events = disk_stat_read(disk, sectors[0]) +
disk_stat_read(disk, sectors[1]) -
idle = 0;
}
}
+ rcu_read_unlock();
return idle;
}