static int btrfs_dev_replace_kthread(void *data)
{
struct btrfs_fs_info *fs_info = data;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
u64 progress;
int ret;
progress = btrfs_dev_replace_progress(fs_info);
progress = div_u64(progress, 10);
btrfs_info_in_rcu(fs_info,
"continuing dev_replace from %s (devid %llu) to target %s @%u%%",
btrfs_dev_name(dev_replace->srcdev),
dev_replace->srcdev->devid,
btrfs_dev_name(dev_replace->tgtdev),
(unsigned int)progress);
ret = btrfs_scrub_dev(fs_info, dev_replace->srcdev->devid,
dev_replace->committed_cursor_left,
btrfs_device_get_total_bytes(dev_replace->srcdev),
&dev_replace->scrub_progress, 0, 1);
ret = btrfs_dev_replace_finishing(fs_info, ret);
WARN_ON(ret && ret != -ECANCELED);
clear_bit(BTRFS_FS_EXCL_OP, &fs_info->flags);
return 0;
}
static int btrfs_dev_replace_kthread(void *data)
{
struct btrfs_fs_info *fs_info = data;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
struct btrfs_ioctl_dev_replace_args *status_args;
u64 progress;
status_args = kzalloc(sizeof(*status_args), GFP_KERNEL);
if (status_args) {
btrfs_dev_replace_status(fs_info, status_args);
progress = status_args->status.progress_1000;
kfree(status_args);
progress = div_u64(progress, 10);
btrfs_info_in_rcu(fs_info,
"continuing dev_replace from %s (devid %llu) to %s @%u%%",
dev_replace->srcdev->missing ? "<missing disk>" :
rcu_str_deref(dev_replace->srcdev->name),
dev_replace->srcdev->devid,
dev_replace->tgtdev ?
rcu_str_deref(dev_replace->tgtdev->name) :
"<missing target disk>",
(unsigned int)progress);
}
btrfs_dev_replace_continue_on_mount(fs_info);
atomic_set(&fs_info->mutually_exclusive_operation_running, 0);
return 0;
}
int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
struct btrfs_device *tgt_device = NULL;
struct btrfs_device *src_device = NULL;
struct btrfs_trans_handle *trans;
struct btrfs_root *root = fs_info->tree_root;
int result;
int ret;
if (sb_rdonly(fs_info->sb))
return -EROFS;
mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
down_write(&dev_replace->rwsem);
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NOT_STARTED;
up_write(&dev_replace->rwsem);
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
tgt_device = dev_replace->tgtdev;
src_device = dev_replace->srcdev;
up_write(&dev_replace->rwsem);
ret = btrfs_scrub_cancel(fs_info);
if (ret < 0) {
result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NOT_STARTED;
} else {
result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
/*
* btrfs_dev_replace_finishing() will handle the
* cleanup part
*/
btrfs_info_in_rcu(fs_info,
"dev_replace from %s (devid %llu) to %s canceled",
btrfs_dev_name(src_device), src_device->devid,
btrfs_dev_name(tgt_device));
}
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
/*
* Scrub doing the replace isn't running so we need to do the
* cleanup step of btrfs_dev_replace_finishing() here
*/
result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
tgt_device = dev_replace->tgtdev;
src_device = dev_replace->srcdev;
dev_replace->tgtdev = NULL;
dev_replace->srcdev = NULL;
dev_replace->replace_state =
BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED;
dev_replace->time_stopped = ktime_get_real_seconds();
dev_replace->item_needs_writeback = 1;
up_write(&dev_replace->rwsem);
/* Scrub for replace must not be running in suspended state */
ret = btrfs_scrub_cancel(fs_info);
ASSERT(ret != -ENOTCONN);
trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return PTR_ERR(trans);
}
ret = btrfs_commit_transaction(trans);
WARN_ON(ret);
btrfs_info_in_rcu(fs_info,
"suspended dev_replace from %s (devid %llu) to %s canceled",
btrfs_dev_name(src_device), src_device->devid,
btrfs_dev_name(tgt_device));
if (tgt_device)
btrfs_destroy_dev_replace_tgtdev(tgt_device);
break;
default:
up_write(&dev_replace->rwsem);
result = -EINVAL;
}
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return result;
}
static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
int scrub_ret)
{
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
struct btrfs_device *tgt_device;
struct btrfs_device *src_device;
struct btrfs_root *root = fs_info->tree_root;
u8 uuid_tmp[BTRFS_UUID_SIZE];
struct btrfs_trans_handle *trans;
int ret = 0;
/* don't allow cancel or unmount to disturb the finishing procedure */
mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
down_read(&dev_replace->rwsem);
/* was the operation canceled, or is it finished? */
if (dev_replace->replace_state !=
BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED) {
up_read(&dev_replace->rwsem);
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return 0;
}
tgt_device = dev_replace->tgtdev;
src_device = dev_replace->srcdev;
up_read(&dev_replace->rwsem);
/*
* flush all outstanding I/O and inode extent mappings before the
* copy operation is declared as being finished
*/
ret = btrfs_start_delalloc_roots(fs_info, -1);
if (ret) {
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return ret;
}
btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1);
trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return PTR_ERR(trans);
}
ret = btrfs_commit_transaction(trans);
WARN_ON(ret);
/* keep away write_all_supers() during the finishing procedure */
mutex_lock(&fs_info->fs_devices->device_list_mutex);
mutex_lock(&fs_info->chunk_mutex);
down_write(&dev_replace->rwsem);
dev_replace->replace_state =
scrub_ret ? BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED
: BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED;
dev_replace->tgtdev = NULL;
dev_replace->srcdev = NULL;
dev_replace->time_stopped = ktime_get_real_seconds();
dev_replace->item_needs_writeback = 1;
/* replace old device with new one in mapping tree */
if (!scrub_ret) {
btrfs_dev_replace_update_device_in_mapping_tree(fs_info,
src_device,
tgt_device);
} else {
if (scrub_ret != -ECANCELED)
btrfs_err_in_rcu(fs_info,
"btrfs_scrub_dev(%s, %llu, %s) failed %d",
btrfs_dev_name(src_device),
src_device->devid,
rcu_str_deref(tgt_device->name), scrub_ret);
up_write(&dev_replace->rwsem);
mutex_unlock(&fs_info->chunk_mutex);
mutex_unlock(&fs_info->fs_devices->device_list_mutex);
btrfs_rm_dev_replace_blocked(fs_info);
if (tgt_device)
btrfs_destroy_dev_replace_tgtdev(tgt_device);
btrfs_rm_dev_replace_unblocked(fs_info);
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return scrub_ret;
}
btrfs_info_in_rcu(fs_info,
"dev_replace from %s (devid %llu) to %s finished",
btrfs_dev_name(src_device),
src_device->devid,
rcu_str_deref(tgt_device->name));
clear_bit(BTRFS_DEV_STATE_REPLACE_TGT, &tgt_device->dev_state);
tgt_device->devid = src_device->devid;
src_device->devid = BTRFS_DEV_REPLACE_DEVID;
memcpy(uuid_tmp, tgt_device->uuid, sizeof(uuid_tmp));
memcpy(tgt_device->uuid, src_device->uuid, sizeof(tgt_device->uuid));
memcpy(src_device->uuid, uuid_tmp, sizeof(src_device->uuid));
btrfs_device_set_total_bytes(tgt_device, src_device->total_bytes);
btrfs_device_set_disk_total_bytes(tgt_device,
src_device->disk_total_bytes);
btrfs_device_set_bytes_used(tgt_device, src_device->bytes_used);
ASSERT(list_empty(&src_device->resized_list));
tgt_device->commit_total_bytes = src_device->commit_total_bytes;
tgt_device->commit_bytes_used = src_device->bytes_used;
btrfs_assign_next_active_device(src_device, tgt_device);
list_add(&tgt_device->dev_alloc_list, &fs_info->fs_devices->alloc_list);
fs_info->fs_devices->rw_devices++;
up_write(&dev_replace->rwsem);
btrfs_rm_dev_replace_blocked(fs_info);
btrfs_rm_dev_replace_remove_srcdev(src_device);
btrfs_rm_dev_replace_unblocked(fs_info);
/*
* Increment dev_stats_ccnt so that btrfs_run_dev_stats() will
* update on-disk dev stats value during commit transaction
*/
atomic_inc(&tgt_device->dev_stats_ccnt);
/*
* this is again a consistent state where no dev_replace procedure
* is running, the target device is part of the filesystem, the
* source device is not part of the filesystem anymore and its 1st
* superblock is scratched out so that it is no longer marked to
* belong to this filesystem.
*/
mutex_unlock(&fs_info->chunk_mutex);
mutex_unlock(&fs_info->fs_devices->device_list_mutex);
/* replace the sysfs entry */
btrfs_sysfs_rm_device_link(fs_info->fs_devices, src_device);
btrfs_rm_dev_replace_free_srcdev(fs_info, src_device);
/* write back the superblocks */
trans = btrfs_start_transaction(root, 0);
if (!IS_ERR(trans))
btrfs_commit_transaction(trans);
mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
return 0;
}
static int btrfs_dev_replace_start(struct btrfs_fs_info *fs_info,
const char *tgtdev_name, u64 srcdevid, const char *srcdev_name,
int read_src)
{
struct btrfs_root *root = fs_info->dev_root;
struct btrfs_trans_handle *trans;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
int ret;
struct btrfs_device *tgt_device = NULL;
struct btrfs_device *src_device = NULL;
bool need_unlock;
src_device = btrfs_find_device_by_devspec(fs_info, srcdevid,
srcdev_name);
if (IS_ERR(src_device))
return PTR_ERR(src_device);
if (btrfs_pinned_by_swapfile(fs_info, src_device)) {
btrfs_warn_in_rcu(fs_info,
"cannot replace device %s (devid %llu) due to active swapfile",
btrfs_dev_name(src_device), src_device->devid);
return -ETXTBSY;
}
ret = btrfs_init_dev_replace_tgtdev(fs_info, tgtdev_name,
src_device, &tgt_device);
if (ret)
return ret;
/*
* Here we commit the transaction to make sure commit_total_bytes
* of all the devices are updated.
*/
trans = btrfs_attach_transaction(root);
if (!IS_ERR(trans)) {
ret = btrfs_commit_transaction(trans);
if (ret)
return ret;
} else if (PTR_ERR(trans) != -ENOENT) {
return PTR_ERR(trans);
}
need_unlock = true;
down_write(&dev_replace->rwsem);
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
ASSERT(0);
ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED;
goto leave;
}
dev_replace->cont_reading_from_srcdev_mode = read_src;
WARN_ON(!src_device);
dev_replace->srcdev = src_device;
dev_replace->tgtdev = tgt_device;
btrfs_info_in_rcu(fs_info,
"dev_replace from %s (devid %llu) to %s started",
btrfs_dev_name(src_device),
src_device->devid,
rcu_str_deref(tgt_device->name));
/*
* from now on, the writes to the srcdev are all duplicated to
* go to the tgtdev as well (refer to btrfs_map_block()).
*/
dev_replace->replace_state = BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED;
dev_replace->time_started = ktime_get_real_seconds();
dev_replace->cursor_left = 0;
dev_replace->committed_cursor_left = 0;
dev_replace->cursor_left_last_write_of_item = 0;
dev_replace->cursor_right = 0;
dev_replace->is_valid = 1;
dev_replace->item_needs_writeback = 1;
atomic64_set(&dev_replace->num_write_errors, 0);
atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0);
up_write(&dev_replace->rwsem);
need_unlock = false;
ret = btrfs_sysfs_add_device_link(tgt_device->fs_devices, tgt_device);
if (ret)
btrfs_err(fs_info, "kobj add dev failed %d", ret);
btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1);
/* force writing the updated state information to disk */
trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
ret = PTR_ERR(trans);
need_unlock = true;
down_write(&dev_replace->rwsem);
dev_replace->replace_state =
BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED;
dev_replace->srcdev = NULL;
dev_replace->tgtdev = NULL;
goto leave;
}
ret = btrfs_commit_transaction(trans);
WARN_ON(ret);
/* the disk copy procedure reuses the scrub code */
ret = btrfs_scrub_dev(fs_info, src_device->devid, 0,
btrfs_device_get_total_bytes(src_device),
&dev_replace->scrub_progress, 0, 1);
ret = btrfs_dev_replace_finishing(fs_info, ret);
if (ret == -EINPROGRESS) {
ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS;
} else if (ret != -ECANCELED) {
WARN_ON(ret);
}
return ret;
leave:
if (need_unlock)
up_write(&dev_replace->rwsem);
btrfs_destroy_dev_replace_tgtdev(tgt_device);
return ret;
}
int btrfs_dev_replace_start(struct btrfs_root *root, char *tgtdev_name,
u64 srcdevid, char *srcdev_name, int read_src)
{
struct btrfs_trans_handle *trans;
struct btrfs_fs_info *fs_info = root->fs_info;
struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
int ret;
struct btrfs_device *tgt_device = NULL;
struct btrfs_device *src_device = NULL;
/* the disk copy procedure reuses the scrub code */
mutex_lock(&fs_info->volume_mutex);
ret = btrfs_find_device_by_devspec(root, srcdevid,
srcdev_name, &src_device);
if (ret) {
mutex_unlock(&fs_info->volume_mutex);
return ret;
}
ret = btrfs_init_dev_replace_tgtdev(root, tgtdev_name,
src_device, &tgt_device);
mutex_unlock(&fs_info->volume_mutex);
if (ret)
return ret;
/*
* Here we commit the transaction to make sure commit_total_bytes
* of all the devices are updated.
*/
trans = btrfs_attach_transaction(root);
if (!IS_ERR(trans)) {
ret = btrfs_commit_transaction(trans, root);
if (ret)
return ret;
} else if (PTR_ERR(trans) != -ENOENT) {
return PTR_ERR(trans);
}
btrfs_dev_replace_lock(dev_replace, 1);
switch (dev_replace->replace_state) {
case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
break;
case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED;
goto leave;
}
dev_replace->cont_reading_from_srcdev_mode = read_src;
WARN_ON(!src_device);
dev_replace->srcdev = src_device;
WARN_ON(!tgt_device);
dev_replace->tgtdev = tgt_device;
btrfs_info_in_rcu(fs_info,
"dev_replace from %s (devid %llu) to %s started",
src_device->missing ? "<missing disk>" :
rcu_str_deref(src_device->name),
src_device->devid,
rcu_str_deref(tgt_device->name));
/*
* from now on, the writes to the srcdev are all duplicated to
* go to the tgtdev as well (refer to btrfs_map_block()).
*/
dev_replace->replace_state = BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED;
dev_replace->time_started = get_seconds();
dev_replace->cursor_left = 0;
dev_replace->committed_cursor_left = 0;
dev_replace->cursor_left_last_write_of_item = 0;
dev_replace->cursor_right = 0;
dev_replace->is_valid = 1;
dev_replace->item_needs_writeback = 1;
atomic64_set(&dev_replace->num_write_errors, 0);
atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0);
btrfs_dev_replace_unlock(dev_replace, 1);
ret = btrfs_sysfs_add_device_link(tgt_device->fs_devices, tgt_device);
if (ret)
btrfs_err(fs_info, "kobj add dev failed %d\n", ret);
btrfs_wait_ordered_roots(root->fs_info, -1, 0, (u64)-1);
/* force writing the updated state information to disk */
trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
ret = PTR_ERR(trans);
btrfs_dev_replace_lock(dev_replace, 1);
goto leave;
}
ret = btrfs_commit_transaction(trans, root);
WARN_ON(ret);
/* the disk copy procedure reuses the scrub code */
ret = btrfs_scrub_dev(fs_info, src_device->devid, 0,
btrfs_device_get_total_bytes(src_device),
&dev_replace->scrub_progress, 0, 1);
ret = btrfs_dev_replace_finishing(fs_info, ret);
if (ret == -EINPROGRESS) {
ret = BTRFS_IOCTL_DEV_REPLACE_RESULT_SCRUB_INPROGRESS;
} else {
WARN_ON(ret);
}
return ret;
leave:
dev_replace->srcdev = NULL;
dev_replace->tgtdev = NULL;
btrfs_dev_replace_unlock(dev_replace, 1);
btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device);
return ret;
}
