diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 190a1b24c902014274084427705a7ed378e4b964..eb7a11ac5b73aa127805b1c8f280ffababc1d423 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2427,6 +2427,10 @@ struct btrfs_root *open_ctree(struct super_block *sb, if (!err) err = btrfs_orphan_cleanup(fs_info->tree_root); up_read(&fs_info->cleanup_work_sem); + + if (!err) + err = btrfs_recover_balance(fs_info->tree_root); + if (err) { close_ctree(tree_root); return ERR_PTR(err); diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 4c60ca0ae90b47e68323f1709b83b33c4b40da25..17e565388de0fa0e2fdfbe40942efa8e93d4357e 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -23,6 +23,7 @@ #include #include #include +#include #include #include "compat.h" #include "ctree.h" @@ -2164,6 +2165,46 @@ static int del_balance_item(struct btrfs_root *root) 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 @@ -2626,10 +2667,18 @@ int btrfs_balance(struct btrfs_balance_control *bctl, do_balance: ret = insert_balance_item(fs_info->tree_root, bctl); - if (ret) + if (ret && ret != -EEXIST) goto out; - set_balance_control(bctl); + if (!(bctl->flags & BTRFS_BALANCE_RESUME)) { + BUG_ON(ret == -EEXIST); + set_balance_control(bctl); + } else { + BUG_ON(ret != -EEXIST); + spin_lock(&fs_info->balance_lock); + update_balance_args(bctl); + spin_unlock(&fs_info->balance_lock); + } mutex_unlock(&fs_info->balance_mutex); @@ -2646,7 +2695,89 @@ int btrfs_balance(struct btrfs_balance_control *bctl, return ret; out: + if (bctl->flags & BTRFS_BALANCE_RESUME) + __cancel_balance(fs_info); + else + kfree(bctl); + return ret; +} + +static int balance_kthread(void *data) +{ + struct btrfs_balance_control *bctl = + (struct btrfs_balance_control *)data; + struct btrfs_fs_info *fs_info = bctl->fs_info; + int ret; + + mutex_lock(&fs_info->volume_mutex); + mutex_lock(&fs_info->balance_mutex); + + set_balance_control(bctl); + + printk(KERN_INFO "btrfs: continuing balance\n"); + ret = btrfs_balance(bctl, NULL); + + mutex_unlock(&fs_info->balance_mutex); + mutex_unlock(&fs_info->volume_mutex); + return ret; +} + +int btrfs_recover_balance(struct btrfs_root *tree_root) +{ + struct task_struct *tsk; + struct btrfs_balance_control *bctl; + 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; + + path = btrfs_alloc_path(); + if (!path) + return -ENOMEM; + + bctl = kzalloc(sizeof(*bctl), GFP_NOFS); + if (!bctl) { + ret = -ENOMEM; + goto out; + } + + key.objectid = BTRFS_BALANCE_OBJECTID; + key.type = BTRFS_BALANCE_ITEM_KEY; + key.offset = 0; + + ret = btrfs_search_slot(NULL, tree_root, &key, path, 0, 0); + if (ret < 0) + goto out_bctl; + if (ret > 0) { /* ret = -ENOENT; */ + ret = 0; + goto out_bctl; + } + + leaf = path->nodes[0]; + item = btrfs_item_ptr(leaf, path->slots[0], struct btrfs_balance_item); + + bctl->fs_info = tree_root->fs_info; + bctl->flags = btrfs_balance_flags(leaf, item) | BTRFS_BALANCE_RESUME; + + btrfs_balance_data(leaf, item, &disk_bargs); + btrfs_disk_balance_args_to_cpu(&bctl->data, &disk_bargs); + btrfs_balance_meta(leaf, item, &disk_bargs); + btrfs_disk_balance_args_to_cpu(&bctl->meta, &disk_bargs); + btrfs_balance_sys(leaf, item, &disk_bargs); + btrfs_disk_balance_args_to_cpu(&bctl->sys, &disk_bargs); + + tsk = kthread_run(balance_kthread, bctl, "btrfs-balance"); + if (IS_ERR(tsk)) + ret = PTR_ERR(tsk); + else + goto out; + +out_bctl: kfree(bctl); +out: + btrfs_free_path(path); return ret; } diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 6c143c98017a950848704595e0d92176e8f0c23c..cd25ea58ec35ef47c3fb0f88118c4b5a574ab349 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -198,6 +198,7 @@ struct map_lookup { BTRFS_BALANCE_METADATA) #define BTRFS_BALANCE_FORCE (1ULL << 3) +#define BTRFS_BALANCE_RESUME (1ULL << 4) /* * Balance filters @@ -271,6 +272,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size); int btrfs_init_new_device(struct btrfs_root *root, char *path); int btrfs_balance(struct btrfs_balance_control *bctl, struct btrfs_ioctl_balance_args *bargs); +int btrfs_recover_balance(struct btrfs_root *tree_root); int btrfs_chunk_readonly(struct btrfs_root *root, u64 chunk_offset); int find_free_dev_extent(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 num_bytes,