/*
 * Copyright (C) 2007 Oracle.  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/ratelimit.h>
#include <linux/kthread.h>
#include <linux/raid/pq.h>
#include <linux/semaphore.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 "raid56.h"
#include "async-thread.h"
#include "check-integrity.h"
#include "rcu-string.h"
#include "math.h"
#include "dev-replace.h"

static int init_first_rw_device(struct btrfs_trans_handle *trans,
                                struct btrfs_root *root,
                                struct btrfs_device *device);
static int btrfs_relocate_sys_chunks(struct btrfs_root *root);
static void __btrfs_reset_dev_stats(struct btrfs_device *dev);
static void btrfs_dev_stat_print_on_error(struct btrfs_device *dev);
static void btrfs_dev_stat_print_on_load(struct btrfs_device *device);

static DEFINE_MUTEX(uuid_mutex);
static LIST_HEAD(fs_uuids);

static void lock_chunks(struct btrfs_root *root)
{
        mutex_lock(&root->fs_info->chunk_mutex);
}

static void unlock_chunks(struct btrfs_root *root)
{
        mutex_unlock(&root->fs_info->chunk_mutex);
}

static struct btrfs_fs_devices *__alloc_fs_devices(void)
{
        struct btrfs_fs_devices *fs_devs;

        fs_devs = kzalloc(sizeof(*fs_devs), GFP_NOFS);
        if (!fs_devs)
                return ERR_PTR(-ENOMEM);

        mutex_init(&fs_devs->device_list_mutex);

        INIT_LIST_HEAD(&fs_devs->devices);
        INIT_LIST_HEAD(&fs_devs->alloc_list);
        INIT_LIST_HEAD(&fs_devs->list);

        return fs_devs;
}

/**
 * alloc_fs_devices - allocate struct btrfs_fs_devices
 * @fsid:       a pointer to UUID for this FS.  If NULL a new UUID is
 *              generated.
 *
 * Return: a pointer to a new &struct btrfs_fs_devices on success;
 * ERR_PTR() on error.  Returned struct is not linked onto any lists and
 * can be destroyed with kfree() right away.
 */
static struct btrfs_fs_devices *alloc_fs_devices(const u8 *fsid)
{
        struct btrfs_fs_devices *fs_devs;

        fs_devs = __alloc_fs_devices();
        if (IS_ERR(fs_devs))
                return fs_devs;

        if (fsid)
                memcpy(fs_devs->fsid, fsid, BTRFS_FSID_SIZE);
        else
                generate_random_uuid(fs_devs->fsid);

        return fs_devs;
}

static void free_fs_devices(struct btrfs_fs_devices *fs_devices)
{
        struct btrfs_device *device;
        WARN_ON(fs_devices->opened);
        while (!list_empty(&fs_devices->devices)) {
                device = list_entry(fs_devices->devices.next,
                                    struct btrfs_device, dev_list);
                list_del(&device->dev_list);
                rcu_string_free(device->name);
                kfree(device);
        }
        kfree(fs_devices);
}

static void btrfs_kobject_uevent(struct block_device *bdev,
                                 enum kobject_action action)
{
        int ret;

        ret = kobject_uevent(&disk_to_dev(bdev->bd_disk)->kobj, action);
        if (ret)
                pr_warn("Sending event '%d' to kobject: '%s' (%p): failed\n",
                        action,
                        kobject_name(&disk_to_dev(bdev->bd_disk)->kobj),
                        &disk_to_dev(bdev->bd_disk)->kobj);
}

void btrfs_cleanup_fs_uuids(void)
{
        struct btrfs_fs_devices *fs_devices;

        while (!list_empty(&fs_uuids)) {
                fs_devices = list_entry(fs_uuids.next,
                                        struct btrfs_fs_devices, list);
                list_del(&fs_devices->list);
                free_fs_devices(fs_devices);
        }
}

static struct btrfs_device *__alloc_device(void)
{
        struct btrfs_device *dev;

        dev = kzalloc(sizeof(*dev), GFP_NOFS);
        if (!dev)
                return ERR_PTR(-ENOMEM);

        INIT_LIST_HEAD(&dev->dev_list);
        INIT_LIST_HEAD(&dev->dev_alloc_list);

        spin_lock_init(&dev->io_lock);

        spin_lock_init(&dev->reada_lock);
        atomic_set(&dev->reada_in_flight, 0);
        INIT_RADIX_TREE(&dev->reada_zones, GFP_NOFS & ~__GFP_WAIT);
        INIT_RADIX_TREE(&dev->reada_extents, GFP_NOFS & ~__GFP_WAIT);

        return dev;
}

static noinline struct btrfs_device *__find_device(struct list_head *head,
                                                   u64 devid, u8 *uuid)
{
        struct btrfs_device *dev;

        list_for_each_entry(dev, head, dev_list) {
                if (dev->devid == devid &&
                    (!uuid || !memcmp(dev->uuid, uuid, BTRFS_UUID_SIZE))) {
                        return dev;
                }
        }
        return NULL;
}

static noinline struct btrfs_fs_devices *find_fsid(u8 *fsid)
{
        struct btrfs_fs_devices *fs_devices;

        list_for_each_entry(fs_devices, &fs_uuids, list) {
                if (memcmp(fsid, fs_devices->fsid, BTRFS_FSID_SIZE) == 0)
                        return fs_devices;
        }
        return NULL;
}

static int
btrfs_get_bdev_and_sb(const char *device_path, fmode_t flags, void *holder,
                      int flush, struct block_device **bdev,
                      struct buffer_head **bh)
{
        int ret;

        *bdev = blkdev_get_by_path(device_path, flags, holder);

        if (IS_ERR(*bdev)) {
                ret = PTR_ERR(*bdev);
                printk(KERN_INFO "btrfs: open %s failed\n", device_path);
                goto error;
        }

        if (flush)
                filemap_write_and_wait((*bdev)->bd_inode->i_mapping);
        ret = set_blocksize(*bdev, 4096);
        if (ret) {
                blkdev_put(*bdev, flags);
                goto error;
        }
        invalidate_bdev(*bdev);
        *bh = btrfs_read_dev_super(*bdev);
        if (!*bh) {
                ret = -EINVAL;
                blkdev_put(*bdev, flags);
                goto error;
        }

        return 0;

error:
        *bdev = NULL;
        *bh = NULL;
        return ret;
}

static void requeue_list(struct btrfs_pending_bios *pending_bios,
                        struct bio *head, struct bio *tail)
{

        struct bio *old_head;

        old_head = pending_bios->head;
        pending_bios->head = head;
        if (pending_bios->tail)
                tail->bi_next = old_head;
        else
                pending_bios->tail = tail;
}

/*
 * we try to collect pending bios for a device so we don't get a large
 * number of procs sending bios down to the same device.  This greatly
 * improves the schedulers ability to collect and merge the bios.
 *
 * But, it also turns into a long list of bios to process and that is sure
 * to eventually make the worker thread block.  The solution here is to
 * make some progress and then put this work struct back at the end of
 * the list if the block device is congested.  This way, multiple devices
 * can make progress from a single worker thread.
 */
static noinline void run_scheduled_bios(struct btrfs_device *device)
{
        struct bio *pending;
        struct backing_dev_info *bdi;
        struct btrfs_fs_info *fs_info;
        struct btrfs_pending_bios *pending_bios;
        struct bio *tail;
        struct bio *cur;
        int again = 0;
        unsigned long num_run;
        unsigned long batch_run = 0;
        unsigned long limit;
        unsigned long last_waited = 0;
        int force_reg = 0;
        int sync_pending = 0;
        struct blk_plug plug;

        /*
         * this function runs all the bios we've collected for
         * a particular device.  We don't want to wander off to
         * another device without first sending all of these down.
         * So, setup a plug here and finish it off before we return
         */
        blk_start_plug(&plug);

        bdi = blk_get_backing_dev_info(device->bdev);
        fs_info = device->dev_root->fs_info;
        limit = btrfs_async_submit_limit(fs_info);
        limit = limit * 2 / 3;

loop:
        spin_lock(&device->io_lock);

loop_lock:
        num_run = 0;

        /* take all the bios off the list at once and process them
         * later on (without the lock held).  But, remember the
         * tail and other pointers so the bios can be properly reinserted
         * into the list if we hit congestion
         */
        if (!force_reg && device->pending_sync_bios.head) {
                pending_bios = &device->pending_sync_bios;
                force_reg = 1;
        } else {
                pending_bios = &device->pending_bios;
                force_reg = 0;
        }

        pending = pending_bios->head;
        tail = pending_bios->tail;
        WARN_ON(pending && !tail);

        /*
         * if pending was null this time around, no bios need processing
         * at all and we can stop.  Otherwise it'll loop back up again
         * and do an additional check so no bios are missed.
         *
         * device->running_pending is used to synchronize with the
         * schedule_bio code.
         */
        if (device->pending_sync_bios.head == NULL &&
            device->pending_bios.head == NULL) {
                again = 0;
                device->running_pending = 0;
        } else {
                again = 1;
                device->running_pending = 1;
        }

        pending_bios->head = NULL;
        pending_bios->tail = NULL;

        spin_unlock(&device->io_lock);

        while (pending) {

                rmb();
                /* we want to work on both lists, but do more bios on the
                 * sync list than the regular list
                 */
                if ((num_run > 32 &&
                    pending_bios != &device->pending_sync_bios &&
                    device->pending_sync_bios.head) ||
                   (num_run > 64 && pending_bios == &device->pending_sync_bios &&
                    device->pending_bios.head)) {
                        spin_lock(&device->io_lock);
                        requeue_list(pending_bios, pending, tail);
                        goto loop_lock;
                }

                cur = pending;
                pending = pending->bi_next;
                cur->bi_next = NULL;

                if (atomic_dec_return(&fs_info->nr_async_bios) < limit &&
                    waitqueue_active(&fs_info->async_submit_wait))
                        wake_up(&fs_info->async_submit_wait);

                BUG_ON(atomic_read(&cur->bi_cnt) == 0);

                /*
                 * if we're doing the sync list, record that our
                 * plug has some sync requests on it
                 *
                 * If we're doing the regular list and there are
                 * sync requests sitting around, unplug before
                 * we add more
                 */
                if (pending_bios == &device->pending_sync_bios) {
                        sync_pending = 1;
                } else if (sync_pending) {
                        blk_finish_plug(&plug);
                        blk_start_plug(&plug);
                        sync_pending = 0;
                }

                btrfsic_submit_bio(cur->bi_rw, cur);
                num_run++;
                batch_run++;
                if (need_resched())
                        cond_resched();

                /*
                 * we made progress, there is more work to do and the bdi
                 * is now congested.  Back off and let other work structs
                 * run instead
                 */
                if (pending && bdi_write_congested(bdi) && batch_run > 8 &&
                    fs_info->fs_devices->open_devices > 1) {
                        struct io_context *ioc;

                        ioc = current->io_context;

                        /*
                         * the main goal here is that we don't want to
                         * block if we're going to be able to submit
                         * more requests without blocking.
                         *
                         * This code does two great things, it pokes into
                         * the elevator code from a filesystem _and_
                         * it makes assumptions about how batching works.
                         */
                        if (ioc && ioc->nr_batch_requests > 0 &&
                            time_before(jiffies, ioc->last_waited + HZ/50UL) &&
                            (last_waited == 0 ||
                             ioc->last_waited == last_waited)) {
                                /*
                                 * we want to go through our batch of
                                 * requests and stop.  So, we copy out
                                 * the ioc->last_waited time and test
                                 * against it before looping
                                 */
                                last_waited = ioc->last_waited;
                                if (need_resched())
                                        cond_resched();
                                continue;
                        }
                        spin_lock(&device->io_lock);
                        requeue_list(pending_bios, pending, tail);
                        device->running_pending = 1;

                        spin_unlock(&device->io_lock);
                        btrfs_requeue_work(&device->work);
                        goto done;
                }
                /* unplug every 64 requests just for good measure */
                if (batch_run % 64 == 0) {
                        blk_finish_plug(&plug);
                        blk_start_plug(&plug);
                        sync_pending = 0;
                }
        }

        cond_resched();
        if (again)
                goto loop;

        spin_lock(&device->io_lock);
        if (device->pending_bios.head || device->pending_sync_bios.head)
                goto loop_lock;
        spin_unlock(&device->io_lock);

done:
        blk_finish_plug(&plug);
}

static void pending_bios_fn(struct btrfs_work *work)
{
        struct btrfs_device *device;

        device = container_of(work, struct btrfs_device, work);
        run_scheduled_bios(device);
}

static noinline int device_list_add(const char *path,
                           struct btrfs_super_block *disk_super,
                           u64 devid, struct btrfs_fs_devices **fs_devices_ret)
{
        struct btrfs_device *device;
        struct btrfs_fs_devices *fs_devices;
        struct rcu_string *name;
        u64 found_transid = btrfs_super_generation(disk_super);

        fs_devices = find_fsid(disk_super->fsid);
        if (!fs_devices) {
                fs_devices = alloc_fs_devices(disk_super->fsid);
                if (IS_ERR(fs_devices))
                        return PTR_ERR(fs_devices);

                list_add(&fs_devices->list, &fs_uuids);
                fs_devices->latest_devid = devid;
                fs_devices->latest_trans = found_transid;

                device = NULL;
        } else {
                device = __find_device(&fs_devices->devices, devid,
                                       disk_super->dev_item.uuid);
        }
        if (!device) {
                if (fs_devices->opened)
                        return -EBUSY;

                device = btrfs_alloc_device(NULL, &devid,
                                            disk_super->dev_item.uuid);
                if (IS_ERR(device)) {
                        /* we can safely leave the fs_devices entry around */
                        return PTR_ERR(device);
                }

                name = rcu_string_strdup(path, GFP_NOFS);
                if (!name) {
                        kfree(device);
                        return -ENOMEM;
                }
                rcu_assign_pointer(device->name, name);

                mutex_lock(&fs_devices->device_list_mutex);
                list_add_rcu(&device->dev_list, &fs_devices->devices);
                fs_devices->num_devices++;
                mutex_unlock(&fs_devices->device_list_mutex);

                device->fs_devices = fs_devices;
        } else if (!device->name || strcmp(device->name->str, path)) {
                name = rcu_string_strdup(path, GFP_NOFS);
                if (!name)
                        return -ENOMEM;
                rcu_string_free(device->name);
                rcu_assign_pointer(device->name, name);
                if (device->missing) {
                        fs_devices->missing_devices--;
                        device->missing = 0;
                }
        }

        if (found_transid > fs_devices->latest_trans) {
                fs_devices->latest_devid = devid;
                fs_devices->latest_trans = found_transid;
        }
        *fs_devices_ret = fs_devices;
        return 0;
}

static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig)
{
        struct btrfs_fs_devices *fs_devices;
        struct btrfs_device *device;
        struct btrfs_device *orig_dev;

        fs_devices = alloc_fs_devices(orig->fsid);
        if (IS_ERR(fs_devices))
                return fs_devices;

        fs_devices->latest_devid = orig->latest_devid;
        fs_devices->latest_trans = orig->latest_trans;
        fs_devices->total_devices = orig->total_devices;

        /* We have held the volume lock, it is safe to get the devices. */
        list_for_each_entry(orig_dev, &orig->devices, dev_list) {
                struct rcu_string *name;

                device = btrfs_alloc_device(NULL, &orig_dev->devid,
                                            orig_dev->uuid);
                if (IS_ERR(device))
                        goto error;

                /*
                 * This is ok to do without rcu read locked because we hold the
                 * uuid mutex so nothing we touch in here is going to disappear.
                 */
                name = rcu_string_strdup(orig_dev->name->str, GFP_NOFS);
                if (!name) {
                        kfree(device);
                        goto error;
                }
                rcu_assign_pointer(device->name, name);

                list_add(&device->dev_list, &fs_devices->devices);
                device->fs_devices = fs_devices;
                fs_devices->num_devices++;
        }
        return fs_devices;
error:
        free_fs_devices(fs_devices);
        return ERR_PTR(-ENOMEM);
}

void btrfs_close_extra_devices(struct btrfs_fs_info *fs_info,
                               struct btrfs_fs_devices *fs_devices, int step)
{
        struct btrfs_device *device, *next;

        struct block_device *latest_bdev = NULL;
        u64 latest_devid = 0;
        u64 latest_transid = 0;

        mutex_lock(&uuid_mutex);
again:
        /* This is the initialized path, it is safe to release the devices. */
        list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
                if (device->in_fs_metadata) {
                        if (!device->is_tgtdev_for_dev_replace &&
                            (!latest_transid ||
                             device->generation > latest_transid)) {
                                latest_devid = device->devid;
                                latest_transid = device->generation;
                                latest_bdev = device->bdev;
                        }
                        continue;
                }

                if (device->devid == BTRFS_DEV_REPLACE_DEVID) {
                        /*
                         * In the first step, keep the device which has
                         * the correct fsid and the devid that is used
                         * for the dev_replace procedure.
                         * In the second step, the dev_replace state is
                         * read from the device tree and it is known
                         * whether the procedure is really active or
                         * not, which means whether this device is
                         * used or whether it should be removed.
                         */
                        if (step == 0 || device->is_tgtdev_for_dev_replace) {
                                continue;
                        }
                }
                if (device->bdev) {
                        blkdev_put(device->bdev, device->mode);
                        device->bdev = NULL;
                        fs_devices->open_devices--;
                }
                if (device->writeable) {
                        list_del_init(&device->dev_alloc_list);
                        device->writeable = 0;
                        if (!device->is_tgtdev_for_dev_replace)
                                fs_devices->rw_devices--;
                }
                list_del_init(&device->dev_list);
                fs_devices->num_devices--;
                rcu_string_free(device->name);
                kfree(device);
        }

        if (fs_devices->seed) {
                fs_devices = fs_devices->seed;
                goto again;
        }

        fs_devices->latest_bdev = latest_bdev;
        fs_devices->latest_devid = latest_devid;
        fs_devices->latest_trans = latest_transid;

        mutex_unlock(&uuid_mutex);
}

static void __free_device(struct work_struct *work)
{
        struct btrfs_device *device;

        device = container_of(work, struct btrfs_device, rcu_work);

        if (device->bdev)
                blkdev_put(device->bdev, device->mode);

        rcu_string_free(device->name);
        kfree(device);
}

static void free_device(struct rcu_head *head)
{
        struct btrfs_device *device;

        device = container_of(head, struct btrfs_device, rcu);

        INIT_WORK(&device->rcu_work, __free_device);
        schedule_work(&device->rcu_work);
}

static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
{
        struct btrfs_device *device;

        if (--fs_devices->opened > 0)
                return 0;

        mutex_lock(&fs_devices->device_list_mutex);
        list_for_each_entry(device, &fs_devices->devices, dev_list) {
                struct btrfs_device *new_device;
                struct rcu_string *name;

                if (device->bdev)
                        fs_devices->open_devices--;

                if (device->writeable && !device->is_tgtdev_for_dev_replace) {
                        list_del_init(&device->dev_alloc_list);
                        fs_devices->rw_devices--;
                }

                if (device->can_discard)
                        fs_devices->num_can_discard--;
                if (device->missing)
                        fs_devices->missing_devices--;

                new_device = btrfs_alloc_device(NULL, &device->devid,
                                                device->uuid);
                BUG_ON(IS_ERR(new_device)); /* -ENOMEM */

                /* Safe because we are under uuid_mutex */
                if (device->name) {
                        name = rcu_string_strdup(device->name->str, GFP_NOFS);
                        BUG_ON(!name); /* -ENOMEM */
                        rcu_assign_pointer(new_device->name, name);
                }

                list_replace_rcu(&device->dev_list, &new_device->dev_list);
                new_device->fs_devices = device->fs_devices;

                call_rcu(&device->rcu, free_device);
        }
        mutex_unlock(&fs_devices->device_list_mutex);

        WARN_ON(fs_devices->open_devices);
        WARN_ON(fs_devices->rw_devices);
        fs_devices->opened = 0;
        fs_devices->seeding = 0;

        return 0;
}

int btrfs_close_devices(struct btrfs_fs_devices *fs_devices)
{
        struct btrfs_fs_devices *seed_devices = NULL;
        int ret;

        mutex_lock(&uuid_mutex);
        ret = __btrfs_close_devices(fs_devices);
        if (!fs_devices->opened) {
                seed_devices = fs_devices->seed;
                fs_devices->seed = NULL;
        }
        mutex_unlock(&uuid_mutex);

        while (seed_devices) {
                fs_devices = seed_devices;
                seed_devices = fs_devices->seed;
                __btrfs_close_devices(fs_devices);
                free_fs_devices(fs_devices);
        }
        /*
         * Wait for rcu kworkers under __btrfs_close_devices
         * to finish all blkdev_puts so device is really
         * free when umount is done.
         */
        rcu_barrier();
        return ret;
}

static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
                                fmode_t flags, void *holder)
{
        struct request_queue *q;
        struct block_device *bdev;
        struct list_head *head = &fs_devices->devices;
        struct btrfs_device *device;
        struct block_device *latest_bdev = NULL;
        struct buffer_head *bh;
        struct btrfs_super_block *disk_super;
        u64 latest_devid = 0;
        u64 latest_transid = 0;
        u64 devid;
        int seeding = 1;
        int ret = 0;

        flags |= FMODE_EXCL;

        list_for_each_entry(device, head, dev_list) {
                if (device->bdev)
                        continue;
                if (!device->name)
                        continue;

                /* Just open everything we can; ignore failures here */
                if (btrfs_get_bdev_and_sb(device->name->str, flags, holder, 1,
                                            &bdev, &bh))
                        continue;

                disk_super = (struct btrfs_super_block *)bh->b_data;
                devid = btrfs_stack_device_id(&disk_super->dev_item);
                if (devid != device->devid)
                        goto error_brelse;

                if (memcmp(device->uuid, disk_super->dev_item.uuid,
                           BTRFS_UUID_SIZE))
                        goto error_brelse;

                device->generation = btrfs_super_generation(disk_super);
                if (!latest_transid || device->generation > latest_transid) {
                        latest_devid = devid;
                        latest_transid = device->generation;
                        latest_bdev = bdev;
                }

                if (btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING) {
                        device->writeable = 0;
                } else {
                        device->writeable = !bdev_read_only(bdev);
                        seeding = 0;
                }

                q = bdev_get_queue(bdev);
                if (blk_queue_discard(q)) {
                        device->can_discard = 1;
                        fs_devices->num_can_discard++;
                }

                device->bdev = bdev;
                device->in_fs_metadata = 0;
                device->mode = flags;

                if (!blk_queue_nonrot(bdev_get_queue(bdev)))
                        fs_devices->rotating = 1;

                fs_devices->open_devices++;
                if (device->writeable &&
                    device->devid != BTRFS_DEV_REPLACE_DEVID) {
                        fs_devices->rw_devices++;
                        list_add(&device->dev_alloc_list,
                                 &fs_devices->alloc_list);
                }
                brelse(bh);
                continue;

error_brelse:
                brelse(bh);
                blkdev_put(bdev, flags);
                continue;
        }
        if (fs_devices->open_devices == 0) {
                ret = -EINVAL;
                goto out;
        }
        fs_devices->seeding = seeding;
        fs_devices->opened = 1;
        fs_devices->latest_bdev = latest_bdev;
        fs_devices->latest_devid = latest_devid;
        fs_devices->latest_trans = latest_transid;
        fs_devices->total_rw_bytes = 0;
out:
        return ret;
}

int btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
                       fmode_t flags, void *holder)
{
        int ret;

        mutex_lock(&uuid_mutex);
        if (fs_devices->opened) {
                fs_devices->opened++;
                ret = 0;
        } else {
                ret = __btrfs_open_devices(fs_devices, flags, holder);
        }
        mutex_unlock(&uuid_mutex);
        return ret;
}

/*
 * Look for a btrfs signature on a device. This may be called out of the mount path
 * and we are not allowed to call set_blocksize during the scan. The superblock
 * is read via pagecache
 */
int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder,
                          struct btrfs_fs_devices **fs_devices_ret)
{
        struct btrfs_super_block *disk_super;
        struct block_device *bdev;
        struct page *page;
        void *p;
        int ret = -EINVAL;
        u64 devid;
        u64 transid;
        u64 total_devices;
        u64 bytenr;
        pgoff_t index;

        /*
         * we would like to check all the supers, but that would make
         * a btrfs mount succeed after a mkfs from a different FS.
         * So, we need to add a special mount option to scan for
         * later supers, using BTRFS_SUPER_MIRROR_MAX instead
         */
        bytenr = btrfs_sb_offset(0);
        flags |= FMODE_EXCL;
        mutex_lock(&uuid_mutex);

        bdev = blkdev_get_by_path(path, flags, holder);

        if (IS_ERR(bdev)) {
                ret = PTR_ERR(bdev);
                goto error;
        }

        /* make sure our super fits in the device */
        if (bytenr + PAGE_CACHE_SIZE >= i_size_read(bdev->bd_inode))
                goto error_bdev_put;

        /* make sure our super fits in the page */
        if (sizeof(*disk_super) > PAGE_CACHE_SIZE)
                goto error_bdev_put;

        /* make sure our super doesn't straddle pages on disk */
        index = bytenr >> PAGE_CACHE_SHIFT;
        if ((bytenr + sizeof(*disk_super) - 1) >> PAGE_CACHE_SHIFT != index)
                goto error_bdev_put;

        /* pull in the page with our super */
        page = read_cache_page_gfp(bdev->bd_inode->i_mapping,
                                   index, GFP_NOFS);

        if (IS_ERR_OR_NULL(page))
                goto error_bdev_put;

        p = kmap(page);

        /* align our pointer to the offset of the super block */
        disk_super = p + (bytenr & ~PAGE_CACHE_MASK);

        if (btrfs_super_bytenr(disk_super) != bytenr ||
            btrfs_super_magic(disk_super) != BTRFS_MAGIC)
                goto error_unmap;

        devid = btrfs_stack_device_id(&disk_super->dev_item);
        transid = btrfs_super_generation(disk_super);
        total_devices = btrfs_super_num_devices(disk_super);

        if (disk_super->label[0]) {
                if (disk_super->label[BTRFS_LABEL_SIZE - 1])
                        disk_super->label[BTRFS_LABEL_SIZE - 1] = '\0';
                printk(KERN_INFO "btrfs: device label %s ", disk_super->label);
        } else {
                printk(KERN_INFO "btrfs: device fsid %pU ", disk_super->fsid);
        }

        printk(KERN_CONT "devid %llu transid %llu %s\n", devid, transid, path);

        ret = device_list_add(path, disk_super, devid, fs_devices_ret);
        if (!ret && fs_devices_ret)
                (*fs_devices_ret)->total_devices = total_devices;

error_unmap:
        kunmap(page);
        page_cache_release(page);

error_bdev_put:
        blkdev_put(bdev, flags);
error:
        mutex_unlock(&uuid_mutex);
        return ret;
}

/* helper to account the used device space in the range */
int btrfs_account_dev_extents_size(struct btrfs_device *device, u64 start,
                                   u64 end, u64 *length)
{
        struct btrfs_key key;
        struct btrfs_root *root = device->dev_root;
        struct btrfs_dev_extent *dev_extent;
        struct btrfs_path *path;
        u64 extent_end;
        int ret;
        int slot;
        struct extent_buffer *l;

        *length = 0;

        if (start >= device->total_bytes || device->is_tgtdev_for_dev_replace)
                return 0;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
        path->reada = 2;

        key.objectid = device->devid;
        key.offset = start;
        key.type = BTRFS_DEV_EXTENT_KEY;

        ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
        if (ret < 0)
                goto out;
        if (ret > 0) {
                ret = btrfs_previous_item(root, path, key.objectid, key.type);
                if (ret < 0)
                        goto out;
        }

        while (1) {
                l = path->nodes[0];
                slot = path->slots[0];
                if (slot >= btrfs_header_nritems(l)) {
                        ret = btrfs_next_leaf(root, path);
                        if (ret == 0)
                                continue;
                        if (ret < 0)
                                goto out;

                        break;
                }
                btrfs_item_key_to_cpu(l, &key, slot);

                if (key.objectid < device->devid)
                        goto next;

                if (key.objectid > device->devid)
                        break;

                if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
                        goto next;

                dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
                extent_end = key.offset + btrfs_dev_extent_length(l,
                                                                  dev_extent);
                if (key.offset <= start && extent_end > end) {
                        *length = end - start + 1;

                        break;
                } else if (key.offset <= start && extent_end > start)
                        *length += extent_end - start;
                else if (key.offset > start && extent_end <= end)
                        *length += extent_end - key.offset;
                else if (key.offset > start && key.offset <= end) {
                        *length += end - key.offset + 1;
                        break;
                } else if (key.offset > end)
                        break;

next:
                path->slots[0]++;
        }
        ret = 0;
out:
        btrfs_free_path(path);
        return ret;
}

static int contains_pending_extent(struct btrfs_trans_handle *trans,
                                   struct btrfs_device *device,
                                   u64 *start, u64 len)
{
        struct extent_map *em;
        int ret = 0;

        list_for_each_entry(em, &trans->transaction->pending_chunks, list) {
                struct map_lookup *map;
                int i;

                map = (struct map_lookup *)em->bdev;
                for (i = 0; i < map->num_stripes; i++) {
                        if (map->stripes[i].dev != device)
                                continue;
                        if (map->stripes[i].physical >= *start + len ||
                            map->stripes[i].physical + em->orig_block_len <=
                            *start)
                                continue;
                        *start = map->stripes[i].physical +
                                em->orig_block_len;
                        ret = 1;
                }
        }

        return ret;
}


/*
 * find_free_dev_extent - find free space in the specified device
 * @device:     the device which we search the free space in
 * @num_bytes:  the size of the free space that we need
 * @start:      store the start of the free space.
 * @len:        the size of the free space. that we find, or the size of the max
 *              free space if we don't find suitable free space
 *
 * this uses a pretty simple search, the expectation is that it is
 * called very infrequently and that a given device has a small number
 * of extents
 *
 * @start is used to store the start of the free space if we find. But if we
 * don't find suitable free space, it will be used to store the start position
 * of the max free space.
 *
 * @len is used to store the size of the free space that we find.
 * But if we don't find suitable free space, it is used to store the size of
 * the max free space.
 */
int find_free_dev_extent(struct btrfs_trans_handle *trans,
                         struct btrfs_device *device, u64 num_bytes,
                         u64 *start, u64 *len)
{
        struct btrfs_key key;
        struct btrfs_root *root = device->dev_root;
        struct btrfs_dev_extent *dev_extent;
        struct btrfs_path *path;
        u64 hole_size;
        u64 max_hole_start;
        u64 max_hole_size;
        u64 extent_end;
        u64 search_start;
        u64 search_end = device->total_bytes;
        int ret;
        int slot;
        struct extent_buffer *l;

        /* FIXME use last free of some kind */

        /* we don't want to overwrite the superblock on the drive,
         * so we make sure to start at an offset of at least 1MB
         */
        search_start = max(root->fs_info->alloc_start, 1024ull * 1024);

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;
again:
        max_hole_start = search_start;
        max_hole_size = 0;
        hole_size = 0;

        if (search_start >= search_end || device->is_tgtdev_for_dev_replace) {
                ret = -ENOSPC;
                goto out;
        }

        path->reada = 2;
        path->search_commit_root = 1;
        path->skip_locking = 1;

        key.objectid = device->devid;
        key.offset = search_start;
        key.type = BTRFS_DEV_EXTENT_KEY;

        ret = btrfs_search_slot(NULL, root, &key, path, 0, 0);
        if (ret < 0)
                goto out;
        if (ret > 0) {
                ret = btrfs_previous_item(root, path, key.objectid, key.type);
                if (ret < 0)
                        goto out;
        }

        while (1) {
                l = path->nodes[0];
                slot = path->slots[0];
                if (slot >= btrfs_header_nritems(l)) {
                        ret = btrfs_next_leaf(root, path);
                        if (ret == 0)
                                continue;
                        if (ret < 0)
                                goto out;

                        break;
                }
                btrfs_item_key_to_cpu(l, &key, slot);

                if (key.objectid < device->devid)
                        goto next;

                if (key.objectid > device->devid)
                        break;

                if (btrfs_key_type(&key) != BTRFS_DEV_EXTENT_KEY)
                        goto next;

                if (key.offset > search_start) {
                        hole_size = key.offset - search_start;

                        /*
                         * Have to check before we set max_hole_start, otherwise
                         * we could end up sending back this offset anyway.
                         */
                        if (contains_pending_extent(trans, device,
                                                    &search_start,
                                                    hole_size))
                                hole_size = 0;

                        if (hole_size > max_hole_size) {
                                max_hole_start = search_start;
                                max_hole_size = hole_size;
                        }

                        /*
                         * If this free space is greater than which we need,
                         * it must be the max free space that we have found
                         * until now, so max_hole_start must point to the start
                         * of this free space and the length of this free space
                         * is stored in max_hole_size. Thus, we return
                         * max_hole_start and max_hole_size and go back to the
                         * caller.
                         */
                        if (hole_size >= num_bytes) {
                                ret = 0;
                                goto out;
                        }
                }

                dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
                extent_end = key.offset + btrfs_dev_extent_length(l,
                                                                  dev_extent);
                if (extent_end > search_start)
                        search_start = extent_end;
next:
                path->slots[0]++;
                cond_resched();
        }

        /*
         * At this point, search_start should be the end of
         * allocated dev extents, and when shrinking the device,
         * search_end may be smaller than search_start.
         */
        if (search_end > search_start)
                hole_size = search_end - search_start;

        if (hole_size > max_hole_size) {
                max_hole_start = search_start;
                max_hole_size = hole_size;
        }

        if (contains_pending_extent(trans, device, &search_start, hole_size)) {
                btrfs_release_path(path);
                goto again;
        }

        /* See above. */
        if (hole_size < num_bytes)
                ret = -ENOSPC;
        else
                ret = 0;

out:
        btrfs_free_path(path);
        *start = max_hole_start;
        if (len)
                *len = max_hole_size;
        return ret;
}

static int btrfs_free_dev_extent(struct btrfs_trans_handle *trans,
                          struct btrfs_device *device,
                          u64 start)
{
        int ret;
        struct btrfs_path *path;
        struct btrfs_root *root = device->dev_root;
        struct btrfs_key key;
        struct btrfs_key found_key;
        struct extent_buffer *leaf = NULL;
        struct btrfs_dev_extent *extent = NULL;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        key.objectid = device->devid;
        key.offset = start;
        key.type = BTRFS_DEV_EXTENT_KEY;
again:
        ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
        if (ret > 0) {
                ret = btrfs_previous_item(root, path, key.objectid,
                                          BTRFS_DEV_EXTENT_KEY);
                if (ret)
                        goto out;
                leaf = path->nodes[0];
                btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);
                extent = btrfs_item_ptr(leaf, path->slots[0],
                                        struct btrfs_dev_extent);
                BUG_ON(found_key.offset > start || found_key.offset +
                       btrfs_dev_extent_length(leaf, extent) < start);
                key = found_key;
                btrfs_release_path(path);
                goto again;
        } else if (ret == 0) {
                leaf = path->nodes[0];
                extent = btrfs_item_ptr(leaf, path->slots[0],
                                        struct btrfs_dev_extent);
        } else {
                btrfs_error(root->fs_info, ret, "Slot search failed");
                goto out;
        }

        if (device->bytes_used > 0) {
                u64 len = btrfs_dev_extent_length(leaf, extent);
                device->bytes_used -= len;
                spin_lock(&root->fs_info->free_chunk_lock);
                root->fs_info->free_chunk_space += len;
                spin_unlock(&root->fs_info->free_chunk_lock);
        }
        ret = btrfs_del_item(trans, root, path);
        if (ret) {
                btrfs_error(root->fs_info, ret,
                            "Failed to remove dev extent item");
        }
out:
        btrfs_free_path(path);
        return ret;
}

static int btrfs_alloc_dev_extent(struct btrfs_trans_handle *trans,
                                  struct btrfs_device *device,
                                  u64 chunk_tree, u64 chunk_objectid,
                                  u64 chunk_offset, u64 start, u64 num_bytes)
{
        int ret;
        struct btrfs_path *path;
        struct btrfs_root *root = device->dev_root;
        struct btrfs_dev_extent *extent;
        struct extent_buffer *leaf;
        struct btrfs_key key;

        WARN_ON(!device->in_fs_metadata);
        WARN_ON(device->is_tgtdev_for_dev_replace);
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        key.objectid = device->devid;
        key.offset = start;
        key.type = BTRFS_DEV_EXTENT_KEY;
        ret = btrfs_insert_empty_item(trans, root, path, &key,
                                      sizeof(*extent));
        if (ret)
                goto out;

        leaf = path->nodes[0];
        extent = btrfs_item_ptr(leaf, path->slots[0],
                                struct btrfs_dev_extent);
        btrfs_set_dev_extent_chunk_tree(leaf, extent, chunk_tree);
        btrfs_set_dev_extent_chunk_objectid(leaf, extent, chunk_objectid);
        btrfs_set_dev_extent_chunk_offset(leaf, extent, chunk_offset);

        write_extent_buffer(leaf, root->fs_info->chunk_tree_uuid,
                    btrfs_dev_extent_chunk_tree_uuid(extent), BTRFS_UUID_SIZE);

        btrfs_set_dev_extent_length(leaf, extent, num_bytes);
        btrfs_mark_buffer_dirty(leaf);
out:
        btrfs_free_path(path);
        return ret;
}

static u64 find_next_chunk(struct btrfs_fs_info *fs_info)
{
        struct extent_map_tree *em_tree;
        struct extent_map *em;
        struct rb_node *n;
        u64 ret = 0;

        em_tree = &fs_info->mapping_tree.map_tree;
        read_lock(&em_tree->lock);
        n = rb_last(&em_tree->map);
        if (n) {
                em = rb_entry(n, struct extent_map, rb_node);
                ret = em->start + em->len;
        }
        read_unlock(&em_tree->lock);

        return ret;
}

static noinline int find_next_devid(struct btrfs_fs_info *fs_info,
                                    u64 *devid_ret)
{
        int ret;
        struct btrfs_key key;
        struct btrfs_key found_key;
        struct btrfs_path *path;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
        key.type = BTRFS_DEV_ITEM_KEY;
        key.offset = (u64)-1;

        ret = btrfs_search_slot(NULL, fs_info->chunk_root, &key, path, 0, 0);
        if (ret < 0)
                goto error;

        BUG_ON(ret == 0); /* Corruption */

        ret = btrfs_previous_item(fs_info->chunk_root, path,
                                  BTRFS_DEV_ITEMS_OBJECTID,
                                  BTRFS_DEV_ITEM_KEY);
        if (ret) {
                *devid_ret = 1;
        } else {
                btrfs_item_key_to_cpu(path->nodes[0], &found_key,
                                      path->slots[0]);
                *devid_ret = found_key.offset + 1;
        }
        ret = 0;
error:
        btrfs_free_path(path);
        return ret;
}

/*
 * the device information is stored in the chunk root
 * the btrfs_device struct should be fully filled in
 */
static int btrfs_add_device(struct btrfs_trans_handle *trans,
                            struct btrfs_root *root,
                            struct btrfs_device *device)
{
        int ret;
        struct btrfs_path *path;
        struct btrfs_dev_item *dev_item;
        struct extent_buffer *leaf;
        struct btrfs_key key;
        unsigned long ptr;

        root = root->fs_info->chunk_root;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
        key.type = BTRFS_DEV_ITEM_KEY;
        key.offset = device->devid;

        ret = btrfs_insert_empty_item(trans, root, path, &key,
                                      sizeof(*dev_item));
        if (ret)
                goto out;

        leaf = path->nodes[0];
        dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);

        btrfs_set_device_id(leaf, dev_item, device->devid);
        btrfs_set_device_generation(leaf, dev_item, 0);
        btrfs_set_device_type(leaf, dev_item, device->type);
        btrfs_set_device_io_align(leaf, dev_item, device->io_align);
        btrfs_set_device_io_width(leaf, dev_item, device->io_width);
        btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
        btrfs_set_device_total_bytes(leaf, dev_item, device->total_bytes);
        btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
        btrfs_set_device_group(leaf, dev_item, 0);
        btrfs_set_device_seek_speed(leaf, dev_item, 0);
        btrfs_set_device_bandwidth(leaf, dev_item, 0);
        btrfs_set_device_start_offset(leaf, dev_item, 0);

        ptr = btrfs_device_uuid(dev_item);
        write_extent_buffer(leaf, device->uuid, ptr, BTRFS_UUID_SIZE);
        ptr = btrfs_device_fsid(dev_item);
        write_extent_buffer(leaf, root->fs_info->fsid, ptr, BTRFS_UUID_SIZE);
        btrfs_mark_buffer_dirty(leaf);

        ret = 0;
out:
        btrfs_free_path(path);
        return ret;
}

static int btrfs_rm_dev_item(struct btrfs_root *root,
                             struct btrfs_device *device)
{
        int ret;
        struct btrfs_path *path;
        struct btrfs_key key;
        struct btrfs_trans_handle *trans;

        root = root->fs_info->chunk_root;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        trans = btrfs_start_transaction(root, 0);
        if (IS_ERR(trans)) {
                btrfs_free_path(path);
                return PTR_ERR(trans);
        }
        key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
        key.type = BTRFS_DEV_ITEM_KEY;
        key.offset = device->devid;
        lock_chunks(root);

        ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
        if (ret < 0)
                goto out;

        if (ret > 0) {
                ret = -ENOENT;
                goto out;
        }

        ret = btrfs_del_item(trans, root, path);
        if (ret)
                goto out;
out:
        btrfs_free_path(path);
        unlock_chunks(root);
        btrfs_commit_transaction(trans, root);
        return ret;
}

int btrfs_rm_device(struct btrfs_root *root, char *device_path)
{
        struct btrfs_device *device;
        struct btrfs_device *next_device;
        struct block_device *bdev;
        struct buffer_head *bh = NULL;
        struct btrfs_super_block *disk_super;
        struct btrfs_fs_devices *cur_devices;
        u64 all_avail;
        u64 devid;
        u64 num_devices;
        u8 *dev_uuid;
        unsigned seq;
        int ret = 0;
        bool clear_super = false;

        mutex_lock(&uuid_mutex);

        do {
                seq = read_seqbegin(&root->fs_info->profiles_lock);

                all_avail = root->fs_info->avail_data_alloc_bits |
                            root->fs_info->avail_system_alloc_bits |
                            root->fs_info->avail_metadata_alloc_bits;
        } while (read_seqretry(&root->fs_info->profiles_lock, seq));

        num_devices = root->fs_info->fs_devices->num_devices;
        btrfs_dev_replace_lock(&root->fs_info->dev_replace);
        if (btrfs_dev_replace_is_ongoing(&root->fs_info->dev_replace)) {
                WARN_ON(num_devices < 1);
                num_devices--;
        }
        btrfs_dev_replace_unlock(&root->fs_info->dev_replace);

        if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices <= 4) {
                ret = BTRFS_ERROR_DEV_RAID10_MIN_NOT_MET;
                goto out;
        }

        if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && num_devices <= 2) {
                ret = BTRFS_ERROR_DEV_RAID1_MIN_NOT_MET;
                goto out;
        }

        if ((all_avail & BTRFS_BLOCK_GROUP_RAID5) &&
            root->fs_info->fs_devices->rw_devices <= 2) {
                ret = BTRFS_ERROR_DEV_RAID5_MIN_NOT_MET;
                goto out;
        }
        if ((all_avail & BTRFS_BLOCK_GROUP_RAID6) &&
            root->fs_info->fs_devices->rw_devices <= 3) {
                ret = BTRFS_ERROR_DEV_RAID6_MIN_NOT_MET;
                goto out;
        }

        if (strcmp(device_path, "missing") == 0) {
                struct list_head *devices;
                struct btrfs_device *tmp;

                device = NULL;
                devices = &root->fs_info->fs_devices->devices;
                /*
                 * It is safe to read the devices since the volume_mutex
                 * is held.
                 */
                list_for_each_entry(tmp, devices, dev_list) {
                        if (tmp->in_fs_metadata &&
                            !tmp->is_tgtdev_for_dev_replace &&
                            !tmp->bdev) {
                                device = tmp;
                                break;
                        }
                }
                bdev = NULL;
                bh = NULL;
                disk_super = NULL;
                if (!device) {
                        ret = BTRFS_ERROR_DEV_MISSING_NOT_FOUND;
                        goto out;
                }
        } else {
                ret = btrfs_get_bdev_and_sb(device_path,
                                            FMODE_WRITE | FMODE_EXCL,
                                            root->fs_info->bdev_holder, 0,
                                            &bdev, &bh);
                if (ret)
                        goto out;
                disk_super = (struct btrfs_super_block *)bh->b_data;
                devid = btrfs_stack_device_id(&disk_super->dev_item);
                dev_uuid = disk_super->dev_item.uuid;
                device = btrfs_find_device(root->fs_info, devid, dev_uuid,
                                           disk_super->fsid);
                if (!device) {
                        ret = -ENOENT;
                        goto error_brelse;
                }
        }

        if (device->is_tgtdev_for_dev_replace) {
                ret = BTRFS_ERROR_DEV_TGT_REPLACE;
                goto error_brelse;
        }

        if (device->writeable && root->fs_info->fs_devices->rw_devices == 1) {
                ret = BTRFS_ERROR_DEV_ONLY_WRITABLE;
                goto error_brelse;
        }

        if (device->writeable) {
                lock_chunks(root);
                list_del_init(&device->dev_alloc_list);
                unlock_chunks(root);
                root->fs_info->fs_devices->rw_devices--;
                clear_super = true;
        }

        mutex_unlock(&uuid_mutex);
        ret = btrfs_shrink_device(device, 0);
        mutex_lock(&uuid_mutex);
        if (ret)
                goto error_undo;

        /*
         * TODO: the superblock still includes this device in its num_devices
         * counter although write_all_supers() is not locked out. This
         * could give a filesystem state which requires a degraded mount.
         */
        ret = btrfs_rm_dev_item(root->fs_info->chunk_root, device);
        if (ret)
                goto error_undo;

        spin_lock(&root->fs_info->free_chunk_lock);
        root->fs_info->free_chunk_space = device->total_bytes -
                device->bytes_used;
        spin_unlock(&root->fs_info->free_chunk_lock);

        device->in_fs_metadata = 0;
        btrfs_scrub_cancel_dev(root->fs_info, device);

        /*
         * the device list mutex makes sure that we don't change
         * the device list while someone else is writing out all
         * the device supers. Whoever is writing all supers, should
         * lock the device list mutex before getting the number of
         * devices in the super block (super_copy). Conversely,
         * whoever updates the number of devices in the super block
         * (super_copy) should hold the device list mutex.
         */

        cur_devices = device->fs_devices;
        mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
        list_del_rcu(&device->dev_list);

        device->fs_devices->num_devices--;
        device->fs_devices->total_devices--;

        if (device->missing)
                root->fs_info->fs_devices->missing_devices--;

        next_device = list_entry(root->fs_info->fs_devices->devices.next,
                                 struct btrfs_device, dev_list);
        if (device->bdev == root->fs_info->sb->s_bdev)
                root->fs_info->sb->s_bdev = next_device->bdev;
        if (device->bdev == root->fs_info->fs_devices->latest_bdev)
                root->fs_info->fs_devices->latest_bdev = next_device->bdev;

        if (device->bdev)
                device->fs_devices->open_devices--;

        call_rcu(&device->rcu, free_device);

        num_devices = btrfs_super_num_devices(root->fs_info->super_copy) - 1;
        btrfs_set_super_num_devices(root->fs_info->super_copy, num_devices);
        mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);

        if (cur_devices->open_devices == 0) {
                struct btrfs_fs_devices *fs_devices;
                fs_devices = root->fs_info->fs_devices;
                while (fs_devices) {
                        if (fs_devices->seed == cur_devices)
                                break;
                        fs_devices = fs_devices->seed;
                }
                fs_devices->seed = cur_devices->seed;
                cur_devices->seed = NULL;
                lock_chunks(root);
                __btrfs_close_devices(cur_devices);
                unlock_chunks(root);
                free_fs_devices(cur_devices);
        }

        root->fs_info->num_tolerated_disk_barrier_failures =
                btrfs_calc_num_tolerated_disk_barrier_failures(root->fs_info);

        /*
         * at this point, the device is zero sized.  We want to
         * remove it from the devices list and zero out the old super
         */
        if (clear_super && disk_super) {
                /* make sure this device isn't detected as part of
                 * the FS anymore
                 */
                memset(&disk_super->magic, 0, sizeof(disk_super->magic));
                set_buffer_dirty(bh);
                sync_dirty_buffer(bh);
        }

        ret = 0;

        /* Notify udev that device has changed */
        if (bdev)
                btrfs_kobject_uevent(bdev, KOBJ_CHANGE);

error_brelse:
        brelse(bh);
        if (bdev)
                blkdev_put(bdev, FMODE_READ | FMODE_EXCL);
out:
        mutex_unlock(&uuid_mutex);
        return ret;
error_undo:
        if (device->writeable) {
                lock_chunks(root);
                list_add(&device->dev_alloc_list,
                         &root->fs_info->fs_devices->alloc_list);
                unlock_chunks(root);
                root->fs_info->fs_devices->rw_devices++;
        }
        goto error_brelse;
}

void btrfs_rm_dev_replace_srcdev(struct btrfs_fs_info *fs_info,
                                 struct btrfs_device *srcdev)
{
        WARN_ON(!mutex_is_locked(&fs_info->fs_devices->device_list_mutex));

        list_del_rcu(&srcdev->dev_list);
        list_del_rcu(&srcdev->dev_alloc_list);
        fs_info->fs_devices->num_devices--;
        if (srcdev->missing) {
                fs_info->fs_devices->missing_devices--;
                fs_info->fs_devices->rw_devices++;
        }
        if (srcdev->can_discard)
                fs_info->fs_devices->num_can_discard--;
        if (srcdev->bdev) {
                fs_info->fs_devices->open_devices--;

                /* zero out the old super */
                btrfs_scratch_superblock(srcdev);
        }

        call_rcu(&srcdev->rcu, free_device);
}

void btrfs_destroy_dev_replace_tgtdev(struct btrfs_fs_info *fs_info,
                                      struct btrfs_device *tgtdev)
{
        struct btrfs_device *next_device;

        WARN_ON(!tgtdev);
        mutex_lock(&fs_info->fs_devices->device_list_mutex);
        if (tgtdev->bdev) {
                btrfs_scratch_superblock(tgtdev);
                fs_info->fs_devices->open_devices--;
        }
        fs_info->fs_devices->num_devices--;
        if (tgtdev->can_discard)
                fs_info->fs_devices->num_can_discard++;

        next_device = list_entry(fs_info->fs_devices->devices.next,
                                 struct btrfs_device, dev_list);
        if (tgtdev->bdev == fs_info->sb->s_bdev)
                fs_info->sb->s_bdev = next_device->bdev;
        if (tgtdev->bdev == fs_info->fs_devices->latest_bdev)
                fs_info->fs_devices->latest_bdev = next_device->bdev;
        list_del_rcu(&tgtdev->dev_list);

        call_rcu(&tgtdev->rcu, free_device);

        mutex_unlock(&fs_info->fs_devices->device_list_mutex);
}

static int btrfs_find_device_by_path(struct btrfs_root *root, char *device_path,
                                     struct btrfs_device **device)
{
        int ret = 0;
        struct btrfs_super_block *disk_super;
        u64 devid;
        u8 *dev_uuid;
        struct block_device *bdev;
        struct buffer_head *bh;

        *device = NULL;
        ret = btrfs_get_bdev_and_sb(device_path, FMODE_READ,
                                    root->fs_info->bdev_holder, 0, &bdev, &bh);
        if (ret)
                return ret;
        disk_super = (struct btrfs_super_block *)bh->b_data;
        devid = btrfs_stack_device_id(&disk_super->dev_item);
        dev_uuid = disk_super->dev_item.uuid;
        *device = btrfs_find_device(root->fs_info, devid, dev_uuid,
                                    disk_super->fsid);
        brelse(bh);
        if (!*device)
                ret = -ENOENT;
        blkdev_put(bdev, FMODE_READ);
        return ret;
}

int btrfs_find_device_missing_or_by_path(struct btrfs_root *root,
                                         char *device_path,
                                         struct btrfs_device **device)
{
        *device = NULL;
        if (strcmp(device_path, "missing") == 0) {
                struct list_head *devices;
                struct btrfs_device *tmp;

                devices = &root->fs_info->fs_devices->devices;
                /*
                 * It is safe to read the devices since the volume_mutex
                 * is held by the caller.
                 */
                list_for_each_entry(tmp, devices, dev_list) {
                        if (tmp->in_fs_metadata && !tmp->bdev) {
                                *device = tmp;
                                break;
                        }
                }

                if (!*device) {
                        pr_err("btrfs: no missing device found\n");
                        return -ENOENT;
                }

                return 0;
        } else {
                return btrfs_find_device_by_path(root, device_path, device);
        }
}

/*
 * does all the dirty work required for changing file system's UUID.
 */
static int btrfs_prepare_sprout(struct btrfs_root *root)
{
        struct btrfs_fs_devices *fs_devices = root->fs_info->fs_devices;
        struct btrfs_fs_devices *old_devices;
        struct btrfs_fs_devices *seed_devices;
        struct btrfs_super_block *disk_super = root->fs_info->super_copy;
        struct btrfs_device *device;
        u64 super_flags;

        BUG_ON(!mutex_is_locked(&uuid_mutex));
        if (!fs_devices->seeding)
                return -EINVAL;

        seed_devices = __alloc_fs_devices();
        if (IS_ERR(seed_devices))
                return PTR_ERR(seed_devices);

        old_devices = clone_fs_devices(fs_devices);
        if (IS_ERR(old_devices)) {
                kfree(seed_devices);
                return PTR_ERR(old_devices);
        }

        list_add(&old_devices->list, &fs_uuids);

        memcpy(seed_devices, fs_devices, sizeof(*seed_devices));
        seed_devices->opened = 1;
        INIT_LIST_HEAD(&seed_devices->devices);
        INIT_LIST_HEAD(&seed_devices->alloc_list);
        mutex_init(&seed_devices->device_list_mutex);

        mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
        list_splice_init_rcu(&fs_devices->devices, &seed_devices->devices,
                              synchronize_rcu);

        list_splice_init(&fs_devices->alloc_list, &seed_devices->alloc_list);
        list_for_each_entry(device, &seed_devices->devices, dev_list) {
                device->fs_devices = seed_devices;
        }

        fs_devices->seeding = 0;
        fs_devices->num_devices = 0;
        fs_devices->open_devices = 0;
        fs_devices->total_devices = 0;
        fs_devices->seed = seed_devices;

        generate_random_uuid(fs_devices->fsid);
        memcpy(root->fs_info->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
        memcpy(disk_super->fsid, fs_devices->fsid, BTRFS_FSID_SIZE);
        mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);

        super_flags = btrfs_super_flags(disk_super) &
                      ~BTRFS_SUPER_FLAG_SEEDING;
        btrfs_set_super_flags(disk_super, super_flags);

        return 0;
}

/*
 * strore the expected generation for seed devices in device items.
 */
static int btrfs_finish_sprout(struct btrfs_trans_handle *trans,
                               struct btrfs_root *root)
{
        struct btrfs_path *path;
        struct extent_buffer *leaf;
        struct btrfs_dev_item *dev_item;
        struct btrfs_device *device;
        struct btrfs_key key;
        u8 fs_uuid[BTRFS_UUID_SIZE];
        u8 dev_uuid[BTRFS_UUID_SIZE];
        u64 devid;
        int ret;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        root = root->fs_info->chunk_root;
        key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
        key.offset = 0;
        key.type = BTRFS_DEV_ITEM_KEY;

        while (1) {
                ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
                if (ret < 0)
                        goto error;

                leaf = path->nodes[0];
next_slot:
                if (path->slots[0] >= btrfs_header_nritems(leaf)) {
                        ret = btrfs_next_leaf(root, path);
                        if (ret > 0)
                                break;
                        if (ret < 0)
                                goto error;
                        leaf = path->nodes[0];
                        btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
                        btrfs_release_path(path);
                        continue;
                }

                btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
                if (key.objectid != BTRFS_DEV_ITEMS_OBJECTID ||
                    key.type != BTRFS_DEV_ITEM_KEY)
                        break;

                dev_item = btrfs_item_ptr(leaf, path->slots[0],
                                          struct btrfs_dev_item);
                devid = btrfs_device_id(leaf, dev_item);
                read_extent_buffer(leaf, dev_uuid, btrfs_device_uuid(dev_item),
                                   BTRFS_UUID_SIZE);
                read_extent_buffer(leaf, fs_uuid, btrfs_device_fsid(dev_item),
                                   BTRFS_UUID_SIZE);
                device = btrfs_find_device(root->fs_info, devid, dev_uuid,
                                           fs_uuid);
                BUG_ON(!device); /* Logic error */

                if (device->fs_devices->seeding) {
                        btrfs_set_device_generation(leaf, dev_item,
                                                    device->generation);
                        btrfs_mark_buffer_dirty(leaf);
                }

                path->slots[0]++;
                goto next_slot;
        }
        ret = 0;
error:
        btrfs_free_path(path);
        return ret;
}

int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
{
        struct request_queue *q;
        struct btrfs_trans_handle *trans;
        struct btrfs_device *device;
        struct block_device *bdev;
        struct list_head *devices;
        struct super_block *sb = root->fs_info->sb;
        struct rcu_string *name;
        u64 total_bytes;
        int seeding_dev = 0;
        int ret = 0;

        if ((sb->s_flags & MS_RDONLY) && !root->fs_info->fs_devices->seeding)
                return -EROFS;

        bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL,
                                  root->fs_info->bdev_holder);
        if (IS_ERR(bdev))
                return PTR_ERR(bdev);

        if (root->fs_info->fs_devices->seeding) {
                seeding_dev = 1;
                down_write(&sb->s_umount);
                mutex_lock(&uuid_mutex);
        }

        filemap_write_and_wait(bdev->bd_inode->i_mapping);

        devices = &root->fs_info->fs_devices->devices;

        mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
        list_for_each_entry(device, devices, dev_list) {
                if (device->bdev == bdev) {
                        ret = -EEXIST;
                        mutex_unlock(
                                &root->fs_info->fs_devices->device_list_mutex);
                        goto error;
                }

        }
        mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);

        device = btrfs_alloc_device(root->fs_info, NULL, NULL);
        if (IS_ERR(device)) {
                /* we can safely leave the fs_devices entry around */
                ret = PTR_ERR(device);
                goto error;
        }

        name = rcu_string_strdup(device_path, GFP_NOFS);
        if (!name) {
                kfree(device);
                ret = -ENOMEM;
                goto error;
        }
        rcu_assign_pointer(device->name, name);

        trans = btrfs_start_transaction(root, 0);
        if (IS_ERR(trans)) {
                rcu_string_free(device->name);
                kfree(device);
                ret = PTR_ERR(trans);
                goto error;
        }

        lock_chunks(root);

        q = bdev_get_queue(bdev);
        if (blk_queue_discard(q))
                device->can_discard = 1;
        device->writeable = 1;
        device->generation = trans->transid;
        device->io_width = root->sectorsize;
        device->io_align = root->sectorsize;
        device->sector_size = root->sectorsize;
        device->total_bytes = i_size_read(bdev->bd_inode);
        device->disk_total_bytes = device->total_bytes;
        device->dev_root = root->fs_info->dev_root;
        device->bdev = bdev;
        device->in_fs_metadata = 1;
        device->is_tgtdev_for_dev_replace = 0;
        device->mode = FMODE_EXCL;
        set_blocksize(device->bdev, 4096);

        if (seeding_dev) {
                sb->s_flags &= ~MS_RDONLY;
                ret = btrfs_prepare_sprout(root);
                BUG_ON(ret); /* -ENOMEM */
        }

        device->fs_devices = root->fs_info->fs_devices;

        mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
        list_add_rcu(&device->dev_list, &root->fs_info->fs_devices->devices);
        list_add(&device->dev_alloc_list,
                 &root->fs_info->fs_devices->alloc_list);
        root->fs_info->fs_devices->num_devices++;
        root->fs_info->fs_devices->open_devices++;
        root->fs_info->fs_devices->rw_devices++;
        root->fs_info->fs_devices->total_devices++;
        if (device->can_discard)
                root->fs_info->fs_devices->num_can_discard++;
        root->fs_info->fs_devices->total_rw_bytes += device->total_bytes;

        spin_lock(&root->fs_info->free_chunk_lock);
        root->fs_info->free_chunk_space += device->total_bytes;
        spin_unlock(&root->fs_info->free_chunk_lock);

        if (!blk_queue_nonrot(bdev_get_queue(bdev)))
                root->fs_info->fs_devices->rotating = 1;

        total_bytes = btrfs_super_total_bytes(root->fs_info->super_copy);
        btrfs_set_super_total_bytes(root->fs_info->super_copy,
                                    total_bytes + device->total_bytes);

        total_bytes = btrfs_super_num_devices(root->fs_info->super_copy);
        btrfs_set_super_num_devices(root->fs_info->super_copy,
                                    total_bytes + 1);
        mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);

        if (seeding_dev) {
                ret = init_first_rw_device(trans, root, device);
                if (ret) {
                        btrfs_abort_transaction(trans, root, ret);
                        goto error_trans;
                }
                ret = btrfs_finish_sprout(trans, root);
                if (ret) {
                        btrfs_abort_transaction(trans, root, ret);
                        goto error_trans;
                }
        } else {
                ret = btrfs_add_device(trans, root, device);
                if (ret) {
                        btrfs_abort_transaction(trans, root, ret);
                        goto error_trans;
                }
        }

        /*
         * we've got more storage, clear any full flags on the space
         * infos
         */
        btrfs_clear_space_info_full(root->fs_info);

        unlock_chunks(root);
        root->fs_info->num_tolerated_disk_barrier_failures =
                btrfs_calc_num_tolerated_disk_barrier_failures(root->fs_info);
        ret = btrfs_commit_transaction(trans, root);

        if (seeding_dev) {
                mutex_unlock(&uuid_mutex);
                up_write(&sb->s_umount);

                if (ret) /* transaction commit */
                        return ret;

                ret = btrfs_relocate_sys_chunks(root);
                if (ret < 0)
                        btrfs_error(root->fs_info, ret,
                                    "Failed to relocate sys chunks after "
                                    "device initialization. This can be fixed "
                                    "using the \"btrfs balance\" command.");
                trans = btrfs_attach_transaction(root);
                if (IS_ERR(trans)) {
                        if (PTR_ERR(trans) == -ENOENT)
                                return 0;
                        return PTR_ERR(trans);
                }
                ret = btrfs_commit_transaction(trans, root);
        }

        return ret;

error_trans:
        unlock_chunks(root);
        btrfs_end_transaction(trans, root);
        rcu_string_free(device->name);
        kfree(device);
error:
        blkdev_put(bdev, FMODE_EXCL);
        if (seeding_dev) {
                mutex_unlock(&uuid_mutex);
                up_write(&sb->s_umount);
        }
        return ret;
}

int btrfs_init_dev_replace_tgtdev(struct btrfs_root *root, char *device_path,
                                  struct btrfs_device **device_out)
{
        struct request_queue *q;
        struct btrfs_device *device;
        struct block_device *bdev;
        struct btrfs_fs_info *fs_info = root->fs_info;
        struct list_head *devices;
        struct rcu_string *name;
        u64 devid = BTRFS_DEV_REPLACE_DEVID;
        int ret = 0;

        *device_out = NULL;
        if (fs_info->fs_devices->seeding)
                return -EINVAL;

        bdev = blkdev_get_by_path(device_path, FMODE_WRITE | FMODE_EXCL,
                                  fs_info->bdev_holder);
        if (IS_ERR(bdev))
                return PTR_ERR(bdev);

        filemap_write_and_wait(bdev->bd_inode->i_mapping);

        devices = &fs_info->fs_devices->devices;
        list_for_each_entry(device, devices, dev_list) {
                if (device->bdev == bdev) {
                        ret = -EEXIST;
                        goto error;
                }
        }

        device = btrfs_alloc_device(NULL, &devid, NULL);
        if (IS_ERR(device)) {
                ret = PTR_ERR(device);
                goto error;
        }

        name = rcu_string_strdup(device_path, GFP_NOFS);
        if (!name) {
                kfree(device);
                ret = -ENOMEM;
                goto error;
        }
        rcu_assign_pointer(device->name, name);

        q = bdev_get_queue(bdev);
        if (blk_queue_discard(q))
                device->can_discard = 1;
        mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
        device->writeable = 1;
        device->generation = 0;
        device->io_width = root->sectorsize;
        device->io_align = root->sectorsize;
        device->sector_size = root->sectorsize;
        device->total_bytes = i_size_read(bdev->bd_inode);
        device->disk_total_bytes = device->total_bytes;
        device->dev_root = fs_info->dev_root;
        device->bdev = bdev;
        device->in_fs_metadata = 1;
        device->is_tgtdev_for_dev_replace = 1;
        device->mode = FMODE_EXCL;
        set_blocksize(device->bdev, 4096);
        device->fs_devices = fs_info->fs_devices;
        list_add(&device->dev_list, &fs_info->fs_devices->devices);
        fs_info->fs_devices->num_devices++;
        fs_info->fs_devices->open_devices++;
        if (device->can_discard)
                fs_info->fs_devices->num_can_discard++;
        mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);

        *device_out = device;
        return ret;

error:
        blkdev_put(bdev, FMODE_EXCL);
        return ret;
}

void btrfs_init_dev_replace_tgtdev_for_resume(struct btrfs_fs_info *fs_info,
                                              struct btrfs_device *tgtdev)
{
        WARN_ON(fs_info->fs_devices->rw_devices == 0);
        tgtdev->io_width = fs_info->dev_root->sectorsize;
        tgtdev->io_align = fs_info->dev_root->sectorsize;
        tgtdev->sector_size = fs_info->dev_root->sectorsize;
        tgtdev->dev_root = fs_info->dev_root;
        tgtdev->in_fs_metadata = 1;
}

static noinline int btrfs_update_device(struct btrfs_trans_handle *trans,
                                        struct btrfs_device *device)
{
        int ret;
        struct btrfs_path *path;
        struct btrfs_root *root;
        struct btrfs_dev_item *dev_item;
        struct extent_buffer *leaf;
        struct btrfs_key key;

        root = device->dev_root->fs_info->chunk_root;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        key.objectid = BTRFS_DEV_ITEMS_OBJECTID;
        key.type = BTRFS_DEV_ITEM_KEY;
        key.offset = device->devid;

        ret = btrfs_search_slot(trans, root, &key, path, 0, 1);
        if (ret < 0)
                goto out;

        if (ret > 0) {
                ret = -ENOENT;
                goto out;
        }

        leaf = path->nodes[0];
        dev_item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_dev_item);

        btrfs_set_device_id(leaf, dev_item, device->devid);
        btrfs_set_device_type(leaf, dev_item, device->type);
        btrfs_set_device_io_align(leaf, dev_item, device->io_align);
        btrfs_set_device_io_width(leaf, dev_item, device->io_width);
        btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
        btrfs_set_device_total_bytes(leaf, dev_item, device->disk_total_bytes);
        btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
        btrfs_mark_buffer_dirty(leaf);

out:
        btrfs_free_path(path);
        return ret;
}

static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
                      struct btrfs_device *device, u64 new_size)
{
        struct btrfs_super_block *super_copy =
                device->dev_root->fs_info->super_copy;
        u64 old_total = btrfs_super_total_bytes(super_copy);
        u64 diff = new_size - device->total_bytes;

        if (!device->writeable)
                return -EACCES;
        if (new_size <= device->total_bytes ||
            device->is_tgtdev_for_dev_replace)
                return -EINVAL;

        btrfs_set_super_total_bytes(super_copy, old_total + diff);
        device->fs_devices->total_rw_bytes += diff;

        device->total_bytes = new_size;
        device->disk_total_bytes = new_size;
        btrfs_clear_space_info_full(device->dev_root->fs_info);

        return btrfs_update_device(trans, device);
}

int btrfs_grow_device(struct btrfs_trans_handle *trans,
                      struct btrfs_device *device, u64 new_size)
{
        int ret;
        lock_chunks(device->dev_root);
        ret = __btrfs_grow_device(trans, device, new_size);
        unlock_chunks(device->dev_root);
        return ret;
}

static int btrfs_free_chunk(struct btrfs_trans_handle *trans,
                            struct btrfs_root *root,
                            u64 chunk_tree, u64 chunk_objectid,
                            u64 chunk_offset)
{
        int ret;
        struct btrfs_path *path;
        struct btrfs_key key;

        root = root->fs_info->chunk_root;
        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        key.objectid = chunk_objectid;
        key.offset = chunk_offset;
        key.type = BTRFS_CHUNK_ITEM_KEY;

        ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
        if (ret < 0)
                goto out;
        else if (ret > 0) { /* Logic error or corruption */
                btrfs_error(root->fs_info, -ENOENT,
                            "Failed lookup while freeing chunk.");
                ret = -ENOENT;
                goto out;
        }

        ret = btrfs_del_item(trans, root, path);
        if (ret < 0)
                btrfs_error(root->fs_info, ret,
                            "Failed to delete chunk item.");
out:
        btrfs_free_path(path);
        return ret;
}

static int btrfs_del_sys_chunk(struct btrfs_root *root, u64 chunk_objectid, u64
                        chunk_offset)
{
        struct btrfs_super_block *super_copy = root->fs_info->super_copy;
        struct btrfs_disk_key *disk_key;
        struct btrfs_chunk *chunk;
        u8 *ptr;
        int ret = 0;
        u32 num_stripes;
        u32 array_size;
        u32 len = 0;
        u32 cur;
        struct btrfs_key key;

        array_size = btrfs_super_sys_array_size(super_copy);

        ptr = super_copy->sys_chunk_array;
        cur = 0;

        while (cur < array_size) {
                disk_key = (struct btrfs_disk_key *)ptr;
                btrfs_disk_key_to_cpu(&key, disk_key);

                len = sizeof(*disk_key);

                if (key.type == BTRFS_CHUNK_ITEM_KEY) {
                        chunk = (struct btrfs_chunk *)(ptr + len);
                        num_stripes = btrfs_stack_chunk_num_stripes(chunk);
                        len += btrfs_chunk_item_size(num_stripes);
                } else {
                        ret = -EIO;
                        break;
                }
                if (key.objectid == chunk_objectid &&
                    key.offset == chunk_offset) {
                        memmove(ptr, ptr + len, array_size - (cur + len));
                        array_size -= len;
                        btrfs_set_super_sys_array_size(super_copy, array_size);
                } else {
                        ptr += len;
                        cur += len;
                }
        }
        return ret;
}

static int btrfs_relocate_chunk(struct btrfs_root *root,
                         u64 chunk_tree, u64 chunk_objectid,
                         u64 chunk_offset)
{
        struct extent_map_tree *em_tree;
        struct btrfs_root *extent_root;
        struct btrfs_trans_handle *trans;
        struct extent_map *em;
        struct map_lookup *map;
        int ret;
        int i;

        root = root->fs_info->chunk_root;
        extent_root = root->fs_info->extent_root;
        em_tree = &root->fs_info->mapping_tree.map_tree;

        ret = btrfs_can_relocate(extent_root, chunk_offset);
        if (ret)
                return -ENOSPC;

        /* step one, relocate all the extents inside this chunk */
        ret = btrfs_relocate_block_group(extent_root, chunk_offset);
        if (ret)
                return ret;

        trans = btrfs_start_transaction(root, 0);
        if (IS_ERR(trans)) {
                ret = PTR_ERR(trans);
                btrfs_std_error(root->fs_info, ret);
                return ret;
        }

        lock_chunks(root);

        /*
         * step two, delete the device extents and the
         * chunk tree entries
         */
        read_lock(&em_tree->lock);
        em = lookup_extent_mapping(em_tree, chunk_offset, 1);
        read_unlock(&em_tree->lock);

        BUG_ON(!em || em->start > chunk_offset ||
               em->start + em->len < chunk_offset);
        map = (struct map_lookup *)em->bdev;

        for (i = 0; i < map->num_stripes; i++) {
                ret = btrfs_free_dev_extent(trans, map->stripes[i].dev,
                                            map->stripes[i].physical);
                BUG_ON(ret);

                if (map->stripes[i].dev) {
                        ret = btrfs_update_device(trans, map->stripes[i].dev);
                        BUG_ON(ret);
                }
        }
        ret = btrfs_free_chunk(trans, root, chunk_tree, chunk_objectid,
                               chunk_offset);

        BUG_ON(ret);

        trace_btrfs_chunk_free(root, map, chunk_offset, em->len);

        if (map->type & BTRFS_BLOCK_GROUP_SYSTEM) {
                ret = btrfs_del_sys_chunk(root, chunk_objectid, chunk_offset);
                BUG_ON(ret);
        }

        ret = btrfs_remove_block_group(trans, extent_root, chunk_offset);
        BUG_ON(ret);

        write_lock(&em_tree->lock);
        remove_extent_mapping(em_tree, em);
        write_unlock(&em_tree->lock);

        kfree(map);
        em->bdev = NULL;

        /* once for the tree */
        free_extent_map(em);
        /* once for us */
        free_extent_map(em);

        unlock_chunks(root);
        btrfs_end_transaction(trans, root);
        return 0;
}

static int btrfs_relocate_sys_chunks(struct btrfs_root *root)
{
        struct btrfs_root *chunk_root = root->fs_info->chunk_root;
        struct btrfs_path *path;
        struct extent_buffer *leaf;
        struct btrfs_chunk *chunk;
        struct btrfs_key key;
        struct btrfs_key found_key;
        u64 chunk_tree = chunk_root->root_key.objectid;
        u64 chunk_type;
        bool retried = false;
        int failed = 0;
        int ret;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

again:
        key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
        key.offset = (u64)-1;
        key.type = BTRFS_CHUNK_ITEM_KEY;

        while (1) {
                ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
                if (ret < 0)
                        goto error;
                BUG_ON(ret == 0); /* Corruption */

                ret = btrfs_previous_item(chunk_root, path, key.objectid,
                                          key.type);
                if (ret < 0)
                        goto error;
                if (ret > 0)
                        break;

                leaf = path->nodes[0];
                btrfs_item_key_to_cpu(leaf, &found_key, path->slots[0]);

                chunk = btrfs_item_ptr(leaf, path->slots[0],
                                       struct btrfs_chunk);
                chunk_type = btrfs_chunk_type(leaf, chunk);
                btrfs_release_path(path);

                if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM) {
                        ret = btrfs_relocate_chunk(chunk_root, chunk_tree,
                                                   found_key.objectid,
                                                   found_key.offset);
                        if (ret == -ENOSPC)
                                failed++;
                        else if (ret)
                                BUG();
                }

                if (found_key.offset == 0)
                        break;
                key.offset = found_key.offset - 1;
        }
        ret = 0;
        if (failed && !retried) {
                failed = 0;
                retried = true;
                goto again;
        } else if (failed && retried) {
                WARN_ON(1);
                ret = -ENOSPC;
        }
error:
        btrfs_free_path(path);
        return ret;
}

static int insert_balance_item(struct btrfs_root *root,
                               struct btrfs_balance_control *bctl)
{
        struct btrfs_trans_handle *trans;
        struct btrfs_balance_item *item;
        struct btrfs_disk_balance_args disk_bargs;
        struct btrfs_path *path;
        struct extent_buffer *leaf;
        struct btrfs_key key;
        int ret, err;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        trans = btrfs_start_transaction(root, 0);
        if (IS_ERR(trans)) {
                btrfs_free_path(path);
                return PTR_ERR(trans);
        }

        key.objectid = BTRFS_BALANCE_OBJECTID;
        key.type = BTRFS_BALANCE_ITEM_KEY;
        key.offset = 0;

        ret = btrfs_insert_empty_item(trans, root, path, &key,
                                      sizeof(*item));
        if (ret)
                goto out;

        leaf = path->nodes[0];
        item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_balance_item);

        memset_extent_buffer(leaf, 0, (unsigned long)item, sizeof(*item));

        btrfs_cpu_balance_args_to_disk(&disk_bargs, &bctl->data);
        btrfs_set_balance_data(leaf, item, &disk_bargs);
        btrfs_cpu_balance_args_to_disk(&disk_bargs, &bctl->meta);
        btrfs_set_balance_meta(leaf, item, &disk_bargs);
        btrfs_cpu_balance_args_to_disk(&disk_bargs, &bctl->sys);
        btrfs_set_balance_sys(leaf, item, &disk_bargs);

        btrfs_set_balance_flags(leaf, item, bctl->flags);

        btrfs_mark_buffer_dirty(leaf);
out:
        btrfs_free_path(path);
        err = btrfs_commit_transaction(trans, root);
        if (err && !ret)
                ret = err;
        return ret;
}

static int del_balance_item(struct btrfs_root *root)
{
        struct btrfs_trans_handle *trans;
        struct btrfs_path *path;
        struct btrfs_key key;
        int ret, err;

        path = btrfs_alloc_path();
        if (!path)
                return -ENOMEM;

        trans = btrfs_start_transaction(root, 0);
        if (IS_ERR(trans)) {
                btrfs_free_path(path);
                return PTR_ERR(trans);
        }

        key.objectid = BTRFS_BALANCE_OBJECTID;
        key.type = BTRFS_BALANCE_ITEM_KEY;
        key.offset = 0;

        ret = btrfs_search_slot(trans, root, &key, path, -1, 1);
        if (ret < 0)
                goto out;
        if (ret > 0) {
                ret = -ENOENT;
                goto out;
        }

        ret = btrfs_del_item(trans, root, path);
out:
        btrfs_free_path(path);
        err = btrfs_commit_transaction(trans, root);
        if (err && !ret)
                ret = err;
        return ret;
}

/*
 * This is a heuristic used to reduce the number of chunks balanced on
 * resume after balance was interrupted.
 */
static void update_balance_args(struct btrfs_balance_control *bctl)
{
        /*
         * Turn on soft mode for chunk types that were being converted.
         */
        if (bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT)
                bctl->data.flags |= BTRFS_BALANCE_ARGS_SOFT;
        if (bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT)
                bctl->sys.flags |= BTRFS_BALANCE_ARGS_SOFT;
        if (bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT)
                bctl->meta.flags |= BTRFS_BALANCE_ARGS_SOFT;

        /*
         * Turn on usage filter if is not already used.  The idea is
         * that chunks that we have already balanced should be
         * reasonably full.  Don't do it for chunks that are being
         * converted - that will keep us from relocating unconverted
         * (albeit full) chunks.
         */
        if (!(bctl->data.flags & BTRFS_BALANCE_ARGS_USAGE) &&
            !(bctl->data.flags & BTRFS_BALANCE_ARGS_CONVERT)) {
                bctl->data.flags |= BTRFS_BALANCE_ARGS_USAGE;
                bctl->data.usage = 90;
        }
        if (!(bctl->sys.flags & BTRFS_BALANCE_ARGS_USAGE) &&
            !(bctl->sys.flags & BTRFS_BALANCE_ARGS_CONVERT)) {
                bctl->sys.flags |= BTRFS_BALANCE_ARGS_USAGE;
                bctl->sys.usage = 90;
        }
        if (!(bctl->meta.flags & BTRFS_BALANCE_ARGS_USAGE) &&
            !(bctl->meta.flags & BTRFS_BALANCE_ARGS_CONVERT)) {
                bctl->meta.flags |= BTRFS_BALANCE_ARGS_USAGE;
                bctl->meta.usage = 90;
        }
}

/*
 * Should be called with both balance and volume mutexes held to
 * serialize other volume operations (add_dev/rm_dev/resize) with
 * restriper.  Same goes for unset_balance_control.
 */
static void set_balance_control(struct btrfs_balance_control *bctl)
{
        struct btrfs_fs_info *fs_info = bctl->fs_info;

        BUG_ON(fs_info->balance_ctl);

        spin_lock(&fs_info->balance_lock);
        fs_info->balance_ctl = bctl;
        spin_unlock(&fs_info->balance_lock);
}

static void unset_balance_control(struct btrfs_fs_info *fs_info)
{
        struct btrfs_balance_control *bctl = fs_info->balance_ctl;

        BUG_ON(!fs_info->balance_ctl);

        spin_lock(&fs_info->balance_lock);
        fs_info->balance_ctl = NULL;
        spin_unlock(&fs_info->balance_lock);

        kfree(bctl);
}

/*
 * Balance filters.  Return 1 if chunk should be filtered out
 * (should not be balanced).
 */
static int chunk_profiles_filter(u64 chunk_type,
                                 struct btrfs_balance_args *bargs)
{
        chunk_type = chunk_to_extended(chunk_type) &
                                BTRFS_EXTENDED_PROFILE_MASK;

        if (bargs->profiles & chunk_type)
                return 0;

        return 1;
}

static int chunk_usage_filter(struct btrfs_fs_info *fs_info, u64 chunk_offset,
                              struct btrfs_balance_args *bargs)
{
        struct btrfs_block_group_cache *cache;
        u64 chunk_used, user_thresh;
        int ret = 1;

        cache = btrfs_lookup_block_group(fs_info, chunk_offset);
        chunk_used = btrfs_block_group_used(&cache->item);

        if (bargs->usage == 0)
                user_thresh = 1;
        else if (bargs->usage > 100)
                user_thresh = cache->key.offset;
        else
                user_thresh = div_factor_fine(cache->key.offset,
                                              bargs->usage);

        if (chunk_used < user_thresh)
                ret = 0;

        btrfs_put_block_group(cache);
        return ret;
}

static int chunk_devid_filter(struct extent_buffer *leaf,
                              struct btrfs_chunk *chunk,
                              struct btrfs_balance_args *bargs)
{
        struct btrfs_stripe *stripe;
        int num_stripes = btrfs_chunk_num_stripes(leaf, chunk);
        int i;

        for (i = 0; i < num_stripes; i++) {
                stripe = btrfs_stripe_nr(chunk, i);
                if (btrfs_stripe_devid(leaf, stripe) == bargs->devid)
                        return 0;
        }

        return 1;
}

/* [pstart, pend) */
static int chunk_drange_filter(struct extent_buffer *leaf,
                               struct btrfs_chunk *chunk,
                               u64 chunk_offset,
                               struct btrfs_balance_args *bargs)
{
        struct btrfs_stripe *stripe;
        int num_stripes = btrfs_chunk_num_stripes(leaf, chunk);
        u64 stripe_offset;
        u64 stripe_length;
        int factor;
        int i;

        if (!(bargs->flags & BTRFS_BALANCE_ARGS_DEVID))
                return 0;

        if (btrfs_chunk_type(leaf, chunk) & (BTRFS_BLOCK_GROUP_DUP |
             BTRFS_BLOCK_GROUP_RAID1 | BTRFS_BLOCK_GROUP_RAID10)) {
                factor = num_stripes / 2;
        } else if (btrfs_chunk_type(leaf, chunk) & BTRFS_BLOCK_GROUP_RAID5) {
                factor = num_stripes - 1;
        } else if (btrfs_chunk_type(leaf, chunk) & BTRFS_BLOCK_GROUP_RAID6) {
                factor = num_stripes - 2;
        } else {
                factor = num_stripes;
        }

        for (i = 0; i < num_stripes; i++) {
                stripe = btrfs_stripe_nr(chunk, i);
                if (btrfs_stripe_devid(leaf, stripe) != bargs->devid)
                        continue;

                stripe_offset = btrfs_stripe_offset(leaf, stripe);
                stripe_length = btrfs_chunk_length(leaf, chunk);
                do_div(stripe_length, factor);

                if (stripe_offset < bargs->pend &&
                    stripe_offset + stripe_length > bargs->pstart)
                        return 0;
        }

        return 1;
}

/* [vstart, vend) */
static int chunk_vrange_filter(struct extent_buffer *leaf,
                               struct btrfs_chunk *chunk,
                               u64 chunk_offset,
                               struct btrfs_balance_args *bargs)
{
        if (chunk_offset < bargs->vend &&
            chunk_offset + btrfs_chunk_length(leaf, chunk) > bargs->vstart)
                /* at least part of the chunk is inside this vrange */
                return 0;

        return 1;
}

static int chunk_soft_convert_filter(u64 chunk_type,
                                     struct btrfs_balance_args *bargs)
{
        if (!(bargs->flags & BTRFS_BALANCE_ARGS_CONVERT))
                return 0;

        chunk_type = chunk_to_extended(chunk_type) &
                                BTRFS_EXTENDED_PROFILE_MASK;

        if (bargs->target == chunk_type)
                return 1;

        return 0;
}

static int should_balance_chunk(struct btrfs_root *root,
                                struct extent_buffer *leaf,
                                struct btrfs_chunk *chunk, u64 chunk_offset)
{
        struct btrfs_balance_control *bctl = root->fs_info->balance_ctl;
        struct btrfs_balance_args *bargs = NULL;
        u64 chunk_type = btrfs_chunk_type(leaf, chunk);

        /* type filter */
        if (!((chunk_type & BTRFS_BLOCK_GROUP_TYPE_MASK) &
              (bctl->flags & BTRFS_BALANCE_TYPE_MASK))) {
                return 0;
        }

        if (chunk_type & BTRFS_BLOCK_GROUP_DATA)
                bargs = &bctl->data;
        else if (chunk_type & BTRFS_BLOCK_GROUP_SYSTEM)
                bargs = &bctl->sys;
        else if (chunk_type & BTRFS_BLOCK_GROUP_METADATA)
                bargs = &bctl->meta;

        /* profiles filter */
        if ((bargs->flags & BTRFS_BALANCE_ARGS_PROFILES) &&
            chunk_profiles_filter(chunk_type, bargs)) {
                return 0;
        }

        /* usage filter */
        if ((bargs->flags & BTRFS_BALANCE_ARGS_USAGE) &&
            chunk_usage_filter(bctl->fs_info, chunk_offset, bargs)) {
                return 0;
        }

        /* devid filter */
        if ((bargs->flags & BTRFS_BALANCE_ARGS_DEVID) &&
            chunk_devid_filter(leaf, chunk, bargs)) {
                return 0;
        }

        /* drange filter, makes sense only with devid filter */
        if ((bargs->flags & BTRFS_BALANCE_ARGS_DRANGE) &&
            chunk_drange_filter(leaf, chunk, chunk_offset, bargs)) {
                return 0;
        }

        /* vrange filter */
        if ((bargs->flags & BTRFS_BALANCE_ARGS_VRANGE) &&
            chunk_vrange_filter(leaf, chunk, chunk_offset, bargs)) {
                return 0;
        }

        /* soft profile changing mode */
        if ((bargs->flags & BTRFS_BALANCE_ARGS_SOFT) &&
            chunk_soft_convert_filter(chunk_type, bargs)) {
                return 0;
        }

        return 1;
}

static int __btrfs_balance(struct btrfs_fs_info *fs_info)
{
        struct btrfs_balance_control *bctl = fs_info->balance_ctl;
        struct btrfs_root *chunk_root = fs_info->chunk_root;
        struct btrfs_root *dev_root = fs_info->dev_root;
        struct list_head *devices;
        struct btrfs_device *device;
        u64 old_size;
        u64 size_to_free;
        struct btrfs_chunk *chunk;
        struct btrfs_path *path;
        struct btrfs_key key;
        struct btrfs_key found_key;
        struct btrfs_trans_handle *trans;
        struct extent_buffer *leaf;
        int slot;
        int ret;
        int enospc_errors = 0;
        bool counting = true;

        /* step one make some room on all the devices */
        devices = &fs_info->fs_devices->devices;
        list_for_each_entry(device, devices, dev_list) {
                old_size = device->total_bytes;
                size_to_free = div_factor(old_size, 1);
                size_to_free = min(size_to_free, (u64)1 * 1024 * 1024);
                if (!device->writeable ||
                    device->total_bytes - device->bytes_used > size_to_free ||
                    device->is_tgtdev_for_dev_replace)
                        continue;

                ret = btrfs_shrink_device(device, old_size - size_to_free);
                if (ret == -ENOSPC)
                        break;
                BUG_ON(ret);

                trans = btrfs_start_transaction(dev_root, 0);
                BUG_ON(IS_ERR(trans));

                ret = btrfs_grow_device(trans, device, old_size);
                BUG_ON(ret);

                btrfs_end_transaction(trans, dev_root);
        }

        /* step two, relocate all the chunks */
        path = btrfs_alloc_path();
        if (!path) {
                ret = -ENOMEM;
                goto error;
        }

        /* zero out stat counters */
        spin_lock(&fs_info->balance_lock);
        memset(&bctl->stat, 0, sizeof(bctl->stat));
        spin_unlock(&fs_info->balance_lock);
again:
        key.objectid = BTRFS_FIRST_CHUNK_TREE_OBJECTID;
        key.offset = (u64)-1;
        key.type = BTRFS_CHUNK_ITEM_KEY;

        while (1) {
                if ((!counting && atomic_read(&fs_info->balance_pause_req)) ||
                    atomic_read(&fs_info->balance_cancel_req)) {
                        ret = -ECANCELED;
                        goto error;
                }

                ret = btrfs_search_slot(NULL, chunk_root, &key, path, 0, 0);
                if (ret < 0)
                        goto error;

                /*
                 * this shouldn't happen, it means the last relocate
                 * failed
                 */
                if (ret == 0)
                        BUG(); /* FIXME break ? */

                ret = btrfs_previous_item(chunk_root, path, 0,
                                          BTRFS_CHUNK_ITEM_KEY);
                if (ret) {
                        ret = 0;
                        break;
                }

                leaf = path->nodes[0];
                slot = path->slots[0];