Re: [PATCH 15/26] Btrfs: add a new source file with device replace code

[Liu Bo <bo.li.liu@oracle.com>]

On Tue, Nov 06, 2012 at 05:38:33PM +0100, Stefan Behrens wrote:
> This adds a new file to the sources together with the header file
> and the changes to ioctl.h that are required by the new C source
> file.
> 
> Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>
> ---
>  fs/btrfs/dev-replace.c | 843 +++++++++++++++++++++++++++++++++++++++++++++++++
>  fs/btrfs/dev-replace.h |  44 +++
>  fs/btrfs/ioctl.h       |  45 +++
>  3 files changed, 932 insertions(+)
> 
> diff --git a/fs/btrfs/dev-replace.c b/fs/btrfs/dev-replace.c
> new file mode 100644
> index 0000000..1d56163
> --- /dev/null
> +++ b/fs/btrfs/dev-replace.c
> @@ -0,0 +1,843 @@
> +/*
> + * Copyright (C) STRATO AG 2012.  All rights reserved.
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public
> + * License v2 as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> + * General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public
> + * License along with this program; if not, write to the
> + * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
> + * Boston, MA 021110-1307, USA.
> + */
> +#include <linux/sched.h>
> +#include <linux/bio.h>
> +#include <linux/slab.h>
> +#include <linux/buffer_head.h>
> +#include <linux/blkdev.h>
> +#include <linux/random.h>
> +#include <linux/iocontext.h>
> +#include <linux/capability.h>
> +#include <linux/kthread.h>
> +#include <linux/math64.h>
> +#include <asm/div64.h>
> +#include "compat.h"
> +#include "ctree.h"
> +#include "extent_map.h"
> +#include "disk-io.h"
> +#include "transaction.h"
> +#include "print-tree.h"
> +#include "volumes.h"
> +#include "async-thread.h"
> +#include "check-integrity.h"
> +#include "rcu-string.h"
> +#include "dev-replace.h"
> +
> +static u64 btrfs_get_seconds_since_1970(void);
> +static int btrfs_dev_replace_finishing(struct btrfs_fs_info *fs_info,
> +				       int scrub_ret);
> +static void btrfs_dev_replace_update_device_in_mapping_tree(
> +						struct btrfs_fs_info *fs_info,
> +						struct btrfs_device *srcdev,
> +						struct btrfs_device *tgtdev);
> +static int btrfs_dev_replace_find_srcdev(struct btrfs_root *root, u64 srcdevid,
> +					 char *srcdev_name,
> +					 struct btrfs_device **device);
> +static u64 __btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info);
> +static int btrfs_dev_replace_kthread(void *data);
> +static int btrfs_dev_replace_continue_on_mount(struct btrfs_fs_info *fs_info);
> +
> +
> +int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info)
> +{
> +	struct btrfs_key key;
> +	struct btrfs_root *dev_root = fs_info->dev_root;
> +	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
> +	struct extent_buffer *eb;
> +	int slot;
> +	int ret = 0;
> +	struct btrfs_path *path = NULL;
> +	int item_size;
> +	struct btrfs_dev_replace_item *ptr;
> +	u64 src_devid;
> +
> +	path = btrfs_alloc_path();
> +	if (!path) {
> +		ret = -ENOMEM;
> +		goto out;
> +	}
> +
> +	key.objectid = 0;
> +	key.type = BTRFS_DEV_REPLACE_KEY;
> +	key.offset = 0;
> +	ret = btrfs_search_slot(NULL, dev_root, &key, path, 0, 0);
> +	if (ret) {
> +no_valid_dev_replace_entry_found:
> +		ret = 0;
> +		dev_replace->replace_state =
> +			BTRFS_DEV_REPLACE_ITEM_STATE_NEVER_STARTED;
> +		dev_replace->cont_reading_from_srcdev_mode =
> +		    BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_ALWAYS;
> +		dev_replace->replace_state = 0;
> +		dev_replace->time_started = 0;
> +		dev_replace->time_stopped = 0;
> +		atomic64_set(&dev_replace->num_write_errors, 0);
> +		atomic64_set(&dev_replace->num_uncorrectable_read_errors, 0);
> +		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->srcdev = NULL;
> +		dev_replace->tgtdev = NULL;
> +		dev_replace->is_valid = 0;
> +		dev_replace->item_needs_writeback = 0;
> +		goto out;
> +	}
> +	slot = path->slots[0];
> +	eb = path->nodes[0];
> +	item_size = btrfs_item_size_nr(eb, slot);
> +	ptr = btrfs_item_ptr(eb, slot, struct btrfs_dev_replace_item);
> +
> +	if (item_size != sizeof(struct btrfs_dev_replace_item)) {
> +		pr_warn("btrfs: dev_replace entry found has unexpected size, ignore entry\n");
> +		goto no_valid_dev_replace_entry_found;
> +	}
> +
> +	src_devid = btrfs_dev_replace_src_devid(eb, ptr);
> +	dev_replace->cont_reading_from_srcdev_mode =
> +		btrfs_dev_replace_cont_reading_from_srcdev_mode(eb, ptr);
> +	dev_replace->replace_state = btrfs_dev_replace_replace_state(eb, ptr);
> +	dev_replace->time_started = btrfs_dev_replace_time_started(eb, ptr);
> +	dev_replace->time_stopped =
> +		btrfs_dev_replace_time_stopped(eb, ptr);
> +	atomic64_set(&dev_replace->num_write_errors,
> +		     btrfs_dev_replace_num_write_errors(eb, ptr));
> +	atomic64_set(&dev_replace->num_uncorrectable_read_errors,
> +		     btrfs_dev_replace_num_uncorrectable_read_errors(eb, ptr));
> +	dev_replace->cursor_left = btrfs_dev_replace_cursor_left(eb, ptr);
> +	dev_replace->committed_cursor_left = dev_replace->cursor_left;
> +	dev_replace->cursor_left_last_write_of_item = dev_replace->cursor_left;
> +	dev_replace->cursor_right = btrfs_dev_replace_cursor_right(eb, ptr);
> +	dev_replace->is_valid = 1;
> +
> +	dev_replace->item_needs_writeback = 0;
> +	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:
> +		dev_replace->srcdev = NULL;
> +		dev_replace->tgtdev = NULL;
> +		break;
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
> +		dev_replace->srcdev = btrfs_find_device(fs_info, src_devid,
> +							NULL, NULL);
> +		dev_replace->tgtdev = btrfs_find_device(fs_info,
> +							BTRFS_DEV_REPLACE_DEVID,
> +							NULL, NULL);
> +		/*
> +		 * allow 'btrfs dev replace_cancel' if src/tgt device is
> +		 * missing
> +		 */
> +		if (!dev_replace->srcdev &&
> +		    !btrfs_test_opt(dev_root, DEGRADED)) {
> +			ret = -EIO;
> +			pr_warn("btrfs: cannot mount because device replace operation is ongoing and\n" "srcdev (devid %llu) is missing, need to run 'btrfs dev scan'?\n",
> +				(unsigned long long)src_devid);
> +		}
> +		if (!dev_replace->tgtdev &&
> +		    !btrfs_test_opt(dev_root, DEGRADED)) {
> +			ret = -EIO;
> +			pr_warn("btrfs: cannot mount because device replace operation is ongoing and\n" "tgtdev (devid %llu) is missing, need to run btrfs dev scan?\n",
> +				(unsigned long long)BTRFS_DEV_REPLACE_DEVID);
> +		}
> +		if (dev_replace->tgtdev) {
> +			if (dev_replace->srcdev) {
> +				dev_replace->tgtdev->total_bytes =
> +					dev_replace->srcdev->total_bytes;
> +				dev_replace->tgtdev->disk_total_bytes =
> +					dev_replace->srcdev->disk_total_bytes;
> +				dev_replace->tgtdev->bytes_used =
> +					dev_replace->srcdev->bytes_used;
> +			}
> +			dev_replace->tgtdev->is_tgtdev_for_dev_replace = 1;
> +			btrfs_init_dev_replace_tgtdev_for_resume(fs_info,
> +				dev_replace->tgtdev);
> +		}
> +		break;
> +	}
> +
> +out:
> +	if (path) {
> +		btrfs_release_path(path);
> +		btrfs_free_path(path);

btrfs_free_path(path) will do release for you :)

> +	}
> +	return ret;
> +}
> +
> +/*
> + * called from commit_transaction. Writes changed device replace state to
> + * disk.
> + */
> +int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
> +			  struct btrfs_fs_info *fs_info)
> +{
> +	int ret;
> +	struct btrfs_root *dev_root = fs_info->dev_root;
> +	struct btrfs_path *path;
> +	struct btrfs_key key;
> +	struct extent_buffer *eb;
> +	struct btrfs_dev_replace_item *ptr;
> +	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
> +
> +	btrfs_dev_replace_lock(dev_replace);
> +	if (!dev_replace->is_valid ||
> +	    !dev_replace->item_needs_writeback) {
> +		btrfs_dev_replace_unlock(dev_replace);
> +		return 0;
> +	}
> +	btrfs_dev_replace_unlock(dev_replace);
> +
> +	key.objectid = 0;
> +	key.type = BTRFS_DEV_REPLACE_KEY;
> +	key.offset = 0;
> +
> +	path = btrfs_alloc_path();
> +	if (!path) {
> +		ret = -ENOMEM;
> +		goto out;
> +	}
> +	ret = btrfs_search_slot(trans, dev_root, &key, path, -1, 1);
> +	if (ret < 0) {
> +		pr_warn("btrfs: error %d while searching for dev_replace item!\n",
> +			ret);
> +		goto out;
> +	}
> +
> +	if (ret == 0 &&
> +	    btrfs_item_size_nr(path->nodes[0], path->slots[0]) < sizeof(*ptr)) {
> +		/*
> +		 * need to delete old one and insert a new one.
> +		 * Since no attempt is made to recover any old state, if the
> +		 * dev_replace state is 'running', the data on the target
> +		 * drive is lost.
> +		 * It would be possible to recover the state: just make sure
> +		 * that the beginning of the item is never changed and always
> +		 * contains all the essential information. Then read this
> +		 * minimal set of information and use it as a base for the
> +		 * new state.
> +		 */
> +		ret = btrfs_del_item(trans, dev_root, path);
> +		if (ret != 0) {
> +			pr_warn("btrfs: delete too small dev_replace item failed %d!\n",
> +				ret);
> +			goto out;
> +		}
> +		ret = 1;
> +	}
> +
> +	if (ret == 1) {
> +		/* need to insert a new item */
> +		btrfs_release_path(path);
> +		ret = btrfs_insert_empty_item(trans, dev_root, path,
> +					      &key, sizeof(*ptr));
> +		if (ret < 0) {
> +			pr_warn("btrfs: insert dev_replace item failed %d!\n",
> +				ret);
> +			goto out;
> +		}
> +	}
> +
> +	eb = path->nodes[0];
> +	ptr = btrfs_item_ptr(eb, path->slots[0],
> +			     struct btrfs_dev_replace_item);
> +
> +	btrfs_dev_replace_lock(dev_replace);
> +	if (dev_replace->srcdev)
> +		btrfs_set_dev_replace_src_devid(eb, ptr,
> +			dev_replace->srcdev->devid);
> +	else
> +		btrfs_set_dev_replace_src_devid(eb, ptr, (u64)-1);
> +	btrfs_set_dev_replace_cont_reading_from_srcdev_mode(eb, ptr,
> +		dev_replace->cont_reading_from_srcdev_mode);
> +	btrfs_set_dev_replace_replace_state(eb, ptr,
> +		dev_replace->replace_state);
> +	btrfs_set_dev_replace_time_started(eb, ptr, dev_replace->time_started);
> +	btrfs_set_dev_replace_time_stopped(eb, ptr, dev_replace->time_stopped);
> +	btrfs_set_dev_replace_num_write_errors(eb, ptr,
> +		atomic64_read(&dev_replace->num_write_errors));
> +	btrfs_set_dev_replace_num_uncorrectable_read_errors(eb, ptr,
> +		atomic64_read(&dev_replace->num_uncorrectable_read_errors));
> +	dev_replace->cursor_left_last_write_of_item =
> +		dev_replace->cursor_left;
> +	btrfs_set_dev_replace_cursor_left(eb, ptr,
> +		dev_replace->cursor_left_last_write_of_item);
> +	btrfs_set_dev_replace_cursor_right(eb, ptr,
> +		dev_replace->cursor_right);
> +	dev_replace->item_needs_writeback = 0;
> +	btrfs_dev_replace_unlock(dev_replace);
> +
> +	btrfs_mark_buffer_dirty(eb);
> +
> +out:
> +	btrfs_free_path(path);
> +
> +	return ret;
> +}
> +
> +void btrfs_after_dev_replace_commit(struct btrfs_fs_info *fs_info)
> +{
> +	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
> +
> +	dev_replace->committed_cursor_left =
> +		dev_replace->cursor_left_last_write_of_item;
> +}
> +
> +static u64 btrfs_get_seconds_since_1970(void)
> +{
> +	struct timespec t = CURRENT_TIME_SEC;
> +
> +	return t.tv_sec;
> +}
> +
> +int btrfs_dev_replace_start(struct btrfs_root *root,
> +			    struct btrfs_ioctl_dev_replace_args *args)
> +{
> +	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;
> +
> +	switch (args->start.cont_reading_from_srcdev_mode) {
> +	case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_ALWAYS:
> +	case BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_AVOID:
> +		break;
> +	default:
> +		return -EINVAL;
> +	}
> +
> +	if ((args->start.srcdevid == 0 && args->start.srcdev_name[0] == '\0') ||
> +	    args->start.tgtdev_name[0] == '\0')
> +		return -EINVAL;
> +
> +	mutex_lock(&fs_info->volume_mutex);
> +	ret = btrfs_init_dev_replace_tgtdev(root, args->start.tgtdev_name,
> +					    &tgt_device);
> +	if (ret) {
> +		pr_err("btrfs: target device %s is invalid!\n",
> +		       args->start.tgtdev_name);
> +		mutex_unlock(&fs_info->volume_mutex);
> +		return -EINVAL;
> +	}
> +
> +	ret = btrfs_dev_replace_find_srcdev(root, args->start.srcdevid,
> +					    args->start.srcdev_name,
> +					    &src_device);
> +	mutex_unlock(&fs_info->volume_mutex);
> +	if (ret) {
> +		ret = -EINVAL;
> +		goto leave_no_lock;
> +	}
> +
> +	if (tgt_device->total_bytes < src_device->total_bytes) {
> +		pr_err("btrfs: target device is smaller than source device!\n");
> +		ret = -EINVAL;
> +		goto leave_no_lock;
> +	}
> +
> +	btrfs_dev_replace_lock(dev_replace);
> +	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:
> +		args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED;
> +		goto leave;
> +	}
> +
> +	dev_replace->cont_reading_from_srcdev_mode =
> +		args->start.cont_reading_from_srcdev_mode;
> +	WARN_ON(!src_device);
> +	dev_replace->srcdev = src_device;
> +	WARN_ON(!tgt_device);
> +	dev_replace->tgtdev = tgt_device;
> +
> +	tgt_device->total_bytes = src_device->total_bytes;
> +	tgt_device->disk_total_bytes = src_device->disk_total_bytes;
> +	tgt_device->bytes_used = src_device->bytes_used;
> +
> +	/*
> +	 * 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 = btrfs_get_seconds_since_1970();
> +	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;
> +	args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
> +	btrfs_dev_replace_unlock(dev_replace);
> +
> +	btrfs_wait_ordered_extents(root, 0);
> +
> +	/* force writing the updated state information to disk */
> +	trans = btrfs_start_transaction(root, 0);

why a start_transaction here?  Any reasons?
(same question also for some other places)

thanks,
liubo

> +	if (IS_ERR(trans)) {
> +		ret = PTR_ERR(trans);
> +		btrfs_dev_replace_lock(dev_replace);
> +		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,
> +			      src_device->total_bytes,
> +			      &dev_replace->scrub_progress, 0, 1);
> +
> +	ret = btrfs_dev_replace_finishing(root->fs_info, ret);
> +	WARN_ON(ret);
> +
> +	return 0;
> +
> +leave:
> +	dev_replace->srcdev = NULL;
> +	dev_replace->tgtdev = NULL;
> +	btrfs_dev_replace_unlock(dev_replace);
> +leave_no_lock:
> +	if (tgt_device)
> +		btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device);
> +	return ret;
> +}
> +
> +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);
> +
> +	btrfs_dev_replace_lock(dev_replace);
> +	/* was the operation canceled, or is it finished? */
> +	if (dev_replace->replace_state !=
> +	    BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED) {
> +		btrfs_dev_replace_unlock(dev_replace);
> +		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
> +		return 0;
> +	}
> +
> +	tgt_device = dev_replace->tgtdev;
> +	src_device = dev_replace->srcdev;
> +	btrfs_dev_replace_unlock(dev_replace);
> +
> +	/* 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);
> +
> +	/*
> +	 * flush all outstanding I/O and inode extent mappings before the
> +	 * copy operation is declared as being finished
> +	 */
> +	btrfs_start_delalloc_inodes(root, 0);
> +	btrfs_wait_ordered_extents(root, 0);
> +
> +	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, root);
> +	WARN_ON(ret);
> +
> +	/* keep away write_all_supers() during the finishing procedure */
> +	mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
> +	btrfs_dev_replace_lock(dev_replace);
> +	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 = btrfs_get_seconds_since_1970();
> +	dev_replace->item_needs_writeback = 1;
> +
> +	if (scrub_ret) {
> +		printk_in_rcu(KERN_ERR
> +			      "btrfs: btrfs_scrub_dev(%s, %llu, %s) failed %d\n",
> +			      rcu_str_deref(src_device->name),
> +			      src_device->devid,
> +			      rcu_str_deref(tgt_device->name), scrub_ret);
> +		btrfs_dev_replace_unlock(dev_replace);
> +		mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
> +		if (tgt_device)
> +			btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device);
> +		mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
> +
> +		return 0;
> +	}
> +
> +	tgt_device->is_tgtdev_for_dev_replace = 0;
> +	tgt_device->devid = src_device->devid;
> +	src_device->devid = BTRFS_DEV_REPLACE_DEVID;
> +	tgt_device->bytes_used = src_device->bytes_used;
> +	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));
> +	tgt_device->total_bytes = src_device->total_bytes;
> +	tgt_device->disk_total_bytes = src_device->disk_total_bytes;
> +	tgt_device->bytes_used = src_device->bytes_used;
> +	if (fs_info->sb->s_bdev == src_device->bdev)
> +		fs_info->sb->s_bdev = tgt_device->bdev;
> +	if (fs_info->fs_devices->latest_bdev == src_device->bdev)
> +		fs_info->fs_devices->latest_bdev = tgt_device->bdev;
> +	list_add(&tgt_device->dev_alloc_list, &fs_info->fs_devices->alloc_list);
> +
> +	btrfs_rm_dev_replace_srcdev(fs_info, src_device);
> +	if (src_device->bdev) {
> +		/* zero out the old super */
> +		btrfs_scratch_superblock(src_device);
> +	}
> +	/*
> +	 * 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.
> +	 */
> +	btrfs_dev_replace_unlock(dev_replace);
> +	mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
> +
> +	/* write back the superblocks */
> +	trans = btrfs_start_transaction(root, 0);
> +	if (!IS_ERR(trans))
> +		btrfs_commit_transaction(trans, root);
> +
> +	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
> +
> +	return 0;
> +}
> +
> +static void btrfs_dev_replace_update_device_in_mapping_tree(
> +						struct btrfs_fs_info *fs_info,
> +						struct btrfs_device *srcdev,
> +						struct btrfs_device *tgtdev)
> +{
> +	struct extent_map_tree *em_tree = &fs_info->mapping_tree.map_tree;
> +	struct extent_map *em;
> +	struct map_lookup *map;
> +	u64 start = 0;
> +	int i;
> +
> +	write_lock(&em_tree->lock);
> +	do {
> +		em = lookup_extent_mapping(em_tree, start, (u64)-1);
> +		if (!em)
> +			break;
> +		map = (struct map_lookup *)em->bdev;
> +		for (i = 0; i < map->num_stripes; i++)
> +			if (srcdev == map->stripes[i].dev)
> +				map->stripes[i].dev = tgtdev;
> +		start = em->start + em->len;
> +		free_extent_map(em);
> +	} while (start);
> +	write_unlock(&em_tree->lock);
> +}
> +
> +static int btrfs_dev_replace_find_srcdev(struct btrfs_root *root, u64 srcdevid,
> +					 char *srcdev_name,
> +					 struct btrfs_device **device)
> +{
> +	int ret;
> +
> +	if (srcdevid) {
> +		*device = btrfs_find_device(root->fs_info, srcdevid, NULL,
> +					    NULL);
> +		if (!*device)
> +			ret = -ENOENT;
> +	} else {
> +		ret = btrfs_find_device_missing_or_by_path(root, srcdev_name,
> +							   device);
> +	}
> +	return ret;
> +}
> +
> +void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info,
> +			      struct btrfs_ioctl_dev_replace_args *args)
> +{
> +	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
> +
> +	btrfs_dev_replace_lock(dev_replace);
> +	/* even if !dev_replace_is_valid, the values are good enough for
> +	 * the replace_status ioctl */
> +	args->result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
> +	args->status.replace_state = dev_replace->replace_state;
> +	args->status.time_started = dev_replace->time_started;
> +	args->status.time_stopped = dev_replace->time_stopped;
> +	args->status.num_write_errors =
> +		atomic64_read(&dev_replace->num_write_errors);
> +	args->status.num_uncorrectable_read_errors =
> +		atomic64_read(&dev_replace->num_uncorrectable_read_errors);
> +	switch (dev_replace->replace_state) {
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED:
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED:
> +		args->status.progress_1000 = 0;
> +		break;
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED:
> +		args->status.progress_1000 = 1000;
> +		break;
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
> +		args->status.progress_1000 = div64_u64(dev_replace->cursor_left,
> +			div64_u64(dev_replace->srcdev->total_bytes, 1000));
> +		break;
> +	}
> +	btrfs_dev_replace_unlock(dev_replace);
> +}
> +
> +int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info,
> +			     struct btrfs_ioctl_dev_replace_args *args)
> +{
> +	args->result = __btrfs_dev_replace_cancel(fs_info);
> +	return 0;
> +}
> +
> +static u64 __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_trans_handle *trans;
> +	struct btrfs_root *root = fs_info->tree_root;
> +	u64 result;
> +	int ret;
> +
> +	mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
> +	btrfs_dev_replace_lock(dev_replace);
> +	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;
> +		btrfs_dev_replace_unlock(dev_replace);
> +		goto leave;
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
> +		result = BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR;
> +		tgt_device = dev_replace->tgtdev;
> +		dev_replace->tgtdev = NULL;
> +		dev_replace->srcdev = NULL;
> +		break;
> +	}
> +	dev_replace->replace_state = BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED;
> +	dev_replace->time_stopped = btrfs_get_seconds_since_1970();
> +	dev_replace->item_needs_writeback = 1;
> +	btrfs_dev_replace_unlock(dev_replace);
> +	btrfs_scrub_cancel(fs_info);
> +
> +	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, root);
> +	WARN_ON(ret);
> +	if (tgt_device)
> +		btrfs_destroy_dev_replace_tgtdev(fs_info, tgt_device);
> +
> +leave:
> +	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
> +	return result;
> +}
> +
> +void btrfs_dev_replace_suspend_for_unmount(struct btrfs_fs_info *fs_info)
> +{
> +	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
> +
> +	mutex_lock(&dev_replace->lock_finishing_cancel_unmount);
> +	btrfs_dev_replace_lock(dev_replace);
> +	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:
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
> +		break;
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
> +		dev_replace->replace_state =
> +			BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED;
> +		dev_replace->time_stopped = btrfs_get_seconds_since_1970();
> +		dev_replace->item_needs_writeback = 1;
> +		pr_info("btrfs: suspending dev_replace for unmount\n");
> +		break;
> +	}
> +
> +	btrfs_dev_replace_unlock(dev_replace);
> +	mutex_unlock(&dev_replace->lock_finishing_cancel_unmount);
> +}
> +
> +/* resume dev_replace procedure that was interrupted by unmount */
> +int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info)
> +{
> +	struct task_struct *task;
> +	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
> +
> +	btrfs_dev_replace_lock(dev_replace);
> +	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:
> +		btrfs_dev_replace_unlock(dev_replace);
> +		return 0;
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
> +		break;
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
> +		dev_replace->replace_state =
> +			BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED;
> +		break;
> +	}
> +	if (!dev_replace->tgtdev || !dev_replace->tgtdev->bdev) {
> +		pr_info("btrfs: cannot continue dev_replace, tgtdev is missing\n"
> +			"btrfs: you may cancel the operation after 'mount -o degraded'\n");
> +		btrfs_dev_replace_unlock(dev_replace);
> +		return 0;
> +	}
> +	btrfs_dev_replace_unlock(dev_replace);
> +
> +	WARN_ON(atomic_xchg(
> +		&fs_info->mutually_exclusive_operation_running, 1));
> +	task = kthread_run(btrfs_dev_replace_kthread, fs_info, "btrfs-devrepl");
> +	return PTR_RET(task);
> +}
> +
> +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_NOFS);
> +	if (status_args) {
> +		btrfs_dev_replace_status(fs_info, status_args);
> +		progress = status_args->status.progress_1000;
> +		kfree(status_args);
> +		do_div(progress, 10);
> +		printk_in_rcu(KERN_INFO
> +			      "btrfs: continuing dev_replace from %s (devid %llu) to %s @%u%%\n",
> +			      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;
> +}
> +
> +static int btrfs_dev_replace_continue_on_mount(struct btrfs_fs_info *fs_info)
> +{
> +	struct btrfs_dev_replace *dev_replace = &fs_info->dev_replace;
> +	int ret;
> +
> +	ret = btrfs_scrub_dev(fs_info, dev_replace->srcdev->devid,
> +			      dev_replace->committed_cursor_left,
> +			      dev_replace->srcdev->total_bytes,
> +			      &dev_replace->scrub_progress, 0, 1);
> +	ret = btrfs_dev_replace_finishing(fs_info, ret);
> +	WARN_ON(ret);
> +	return 0;
> +}
> +
> +int btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace)
> +{
> +	if (!dev_replace->is_valid)
> +		return 0;
> +
> +	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:
> +		return 0;
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED:
> +	case BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED:
> +		/*
> +		 * return true even if tgtdev is missing (this is
> +		 * something that can happen if the dev_replace
> +		 * procedure is suspended by an umount and then
> +		 * the tgtdev is missing (or "btrfs dev scan") was
> +		 * not called and the the filesystem is remounted
> +		 * in degraded state. This does not stop the
> +		 * dev_replace procedure. It needs to be canceled
> +		 * manually if the cancelation is wanted.
> +		 */
> +		break;
> +	}
> +	return 1;
> +}
> +
> +void btrfs_dev_replace_lock(struct btrfs_dev_replace *dev_replace)
> +{
> +	/* the beginning is just an optimization for the typical case */
> +	if (atomic_read(&dev_replace->nesting_level) == 0) {
> +acquire_lock:
> +		/* this is not a nested case where the same thread
> +		 * is trying to acqurire the same lock twice */
> +		mutex_lock(&dev_replace->lock);
> +		mutex_lock(&dev_replace->lock_management_lock);
> +		dev_replace->lock_owner = current->pid;
> +		atomic_inc(&dev_replace->nesting_level);
> +		mutex_unlock(&dev_replace->lock_management_lock);
> +		return;
> +	}
> +
> +	mutex_lock(&dev_replace->lock_management_lock);
> +	if (atomic_read(&dev_replace->nesting_level) > 0 &&
> +	    dev_replace->lock_owner == current->pid) {
> +		WARN_ON(!mutex_is_locked(&dev_replace->lock));
> +		atomic_inc(&dev_replace->nesting_level);
> +		mutex_unlock(&dev_replace->lock_management_lock);
> +		return;
> +	}
> +
> +	mutex_unlock(&dev_replace->lock_management_lock);
> +	goto acquire_lock;
> +}
> +
> +void btrfs_dev_replace_unlock(struct btrfs_dev_replace *dev_replace)
> +{
> +	WARN_ON(!mutex_is_locked(&dev_replace->lock));
> +	mutex_lock(&dev_replace->lock_management_lock);
> +	WARN_ON(atomic_read(&dev_replace->nesting_level) < 1);
> +	WARN_ON(dev_replace->lock_owner != current->pid);
> +	atomic_dec(&dev_replace->nesting_level);
> +	if (atomic_read(&dev_replace->nesting_level) == 0) {
> +		dev_replace->lock_owner = 0;
> +		mutex_unlock(&dev_replace->lock_management_lock);
> +		mutex_unlock(&dev_replace->lock);
> +	} else {
> +		mutex_unlock(&dev_replace->lock_management_lock);
> +	}
> +}
> diff --git a/fs/btrfs/dev-replace.h b/fs/btrfs/dev-replace.h
> new file mode 100644
> index 0000000..20035cb
> --- /dev/null
> +++ b/fs/btrfs/dev-replace.h
> @@ -0,0 +1,44 @@
> +/*
> + * Copyright (C) STRATO AG 2012.  All rights reserved.
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public
> + * License v2 as published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> + * General Public License for more details.
> + *
> + * You should have received a copy of the GNU General Public
> + * License along with this program; if not, write to the
> + * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
> + * Boston, MA 021110-1307, USA.
> + */
> +
> +#if !defined(__BTRFS_DEV_REPLACE__)
> +#define __BTRFS_DEV_REPLACE__
> +
> +struct btrfs_ioctl_dev_replace_args;
> +
> +int btrfs_init_dev_replace(struct btrfs_fs_info *fs_info);
> +int btrfs_run_dev_replace(struct btrfs_trans_handle *trans,
> +			  struct btrfs_fs_info *fs_info);
> +void btrfs_after_dev_replace_commit(struct btrfs_fs_info *fs_info);
> +int btrfs_dev_replace_start(struct btrfs_root *root,
> +			    struct btrfs_ioctl_dev_replace_args *args);
> +void btrfs_dev_replace_status(struct btrfs_fs_info *fs_info,
> +			      struct btrfs_ioctl_dev_replace_args *args);
> +int btrfs_dev_replace_cancel(struct btrfs_fs_info *fs_info,
> +			     struct btrfs_ioctl_dev_replace_args *args);
> +void btrfs_dev_replace_suspend_for_unmount(struct btrfs_fs_info *fs_info);
> +int btrfs_resume_dev_replace_async(struct btrfs_fs_info *fs_info);
> +int btrfs_dev_replace_is_ongoing(struct btrfs_dev_replace *dev_replace);
> +void btrfs_dev_replace_lock(struct btrfs_dev_replace *dev_replace);
> +void btrfs_dev_replace_unlock(struct btrfs_dev_replace *dev_replace);
> +
> +static inline void btrfs_dev_replace_stats_inc(atomic64_t *stat_value)
> +{
> +	atomic64_inc(stat_value);
> +}
> +#endif
> diff --git a/fs/btrfs/ioctl.h b/fs/btrfs/ioctl.h
> index 731e287..62006ba 100644
> --- a/fs/btrfs/ioctl.h
> +++ b/fs/btrfs/ioctl.h
> @@ -123,6 +123,48 @@ struct btrfs_ioctl_scrub_args {
>  	__u64 unused[(1024-32-sizeof(struct btrfs_scrub_progress))/8];
>  };
>  
> +#define BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_ALWAYS	0
> +#define BTRFS_IOCTL_DEV_REPLACE_CONT_READING_FROM_SRCDEV_MODE_AVOID	1
> +struct btrfs_ioctl_dev_replace_start_params {
> +	__u64 srcdevid;	/* in, if 0, use srcdev_name instead */
> +	__u8 srcdev_name[BTRFS_PATH_NAME_MAX + 1];	/* in */
> +	__u8 tgtdev_name[BTRFS_PATH_NAME_MAX + 1];	/* in */
> +	__u64 cont_reading_from_srcdev_mode;	/* in, see #define
> +						 * above */
> +};
> +
> +#define BTRFS_IOCTL_DEV_REPLACE_STATE_NEVER_STARTED	0
> +#define BTRFS_IOCTL_DEV_REPLACE_STATE_STARTED		1
> +#define BTRFS_IOCTL_DEV_REPLACE_STATE_FINISHED		2
> +#define BTRFS_IOCTL_DEV_REPLACE_STATE_CANCELED		3
> +#define BTRFS_IOCTL_DEV_REPLACE_STATE_SUSPENDED		4
> +struct btrfs_ioctl_dev_replace_status_params {
> +	__u64 replace_state;	/* out, see #define above */
> +	__u64 progress_1000;	/* out, 0 <= x <= 1000 */
> +	__u64 time_started;	/* out, seconds since 1-Jan-1970 */
> +	__u64 time_stopped;	/* out, seconds since 1-Jan-1970 */
> +	__u64 num_write_errors;	/* out */
> +	__u64 num_uncorrectable_read_errors;	/* out */
> +};
> +
> +#define BTRFS_IOCTL_DEV_REPLACE_CMD_START			0
> +#define BTRFS_IOCTL_DEV_REPLACE_CMD_STATUS			1
> +#define BTRFS_IOCTL_DEV_REPLACE_CMD_CANCEL			2
> +#define BTRFS_IOCTL_DEV_REPLACE_RESULT_NO_ERROR			0
> +#define BTRFS_IOCTL_DEV_REPLACE_RESULT_NOT_STARTED		1
> +#define BTRFS_IOCTL_DEV_REPLACE_RESULT_ALREADY_STARTED		2
> +struct btrfs_ioctl_dev_replace_args {
> +	__u64 cmd;	/* in */
> +	__u64 result;	/* out */
> +
> +	union {
> +		struct btrfs_ioctl_dev_replace_start_params start;
> +		struct btrfs_ioctl_dev_replace_status_params status;
> +	};	/* in/out */
> +
> +	__u64 spare[64];
> +};
> +
>  #define BTRFS_DEVICE_PATH_NAME_MAX 1024
>  struct btrfs_ioctl_dev_info_args {
>  	__u64 devid;				/* in/out */
> @@ -453,4 +495,7 @@ struct btrfs_ioctl_send_args {
>  			       struct btrfs_ioctl_qgroup_limit_args)
>  #define BTRFS_IOC_GET_DEV_STATS _IOWR(BTRFS_IOCTL_MAGIC, 52, \
>  				      struct btrfs_ioctl_get_dev_stats)
> +#define BTRFS_IOC_DEV_REPLACE _IOWR(BTRFS_IOCTL_MAGIC, 53, \
> +				    struct btrfs_ioctl_dev_replace_args)
> +
>  #endif
> -- 
> 1.8.0
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html