diff options
Diffstat (limited to 'fs/btrfs/transaction.c')
| -rw-r--r-- | fs/btrfs/transaction.c | 2011 |
1 files changed, 1171 insertions, 840 deletions
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index f615d59b0489..05ee4391c83a 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -1,81 +1,153 @@ +// SPDX-License-Identifier: GPL-2.0 /* * 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/fs.h> #include <linux/slab.h> #include <linux/sched.h> +#include <linux/sched/mm.h> #include <linux/writeback.h> #include <linux/pagemap.h> #include <linux/blkdev.h> #include <linux/uuid.h> +#include <linux/timekeeping.h> +#include "misc.h" #include "ctree.h" #include "disk-io.h" #include "transaction.h" #include "locking.h" #include "tree-log.h" -#include "inode-map.h" #include "volumes.h" #include "dev-replace.h" #include "qgroup.h" +#include "block-group.h" +#include "space-info.h" +#include "fs.h" +#include "accessors.h" +#include "extent-tree.h" +#include "root-tree.h" +#include "dir-item.h" +#include "uuid-tree.h" +#include "ioctl.h" +#include "relocation.h" +#include "scrub.h" +#include "ordered-data.h" +#include "delayed-inode.h" + +static struct kmem_cache *btrfs_trans_handle_cachep; -#define BTRFS_ROOT_TRANS_TAG 0 - +/* + * Transaction states and transitions + * + * No running transaction (fs tree blocks are not modified) + * | + * | To next stage: + * | Call start_transaction() variants. Except btrfs_join_transaction_nostart(). + * V + * Transaction N [[TRANS_STATE_RUNNING]] + * | + * | New trans handles can be attached to transaction N by calling all + * | start_transaction() variants. + * | + * | To next stage: + * | Call btrfs_commit_transaction() on any trans handle attached to + * | transaction N + * V + * Transaction N [[TRANS_STATE_COMMIT_PREP]] + * | + * | If there are simultaneous calls to btrfs_commit_transaction() one will win + * | the race and the rest will wait for the winner to commit the transaction. + * | + * | The winner will wait for previous running transaction to completely finish + * | if there is one. + * | + * Transaction N [[TRANS_STATE_COMMIT_START]] + * | + * | Then one of the following happens: + * | - Wait for all other trans handle holders to release. + * | The btrfs_commit_transaction() caller will do the commit work. + * | - Wait for current transaction to be committed by others. + * | Other btrfs_commit_transaction() caller will do the commit work. + * | + * | At this stage, only btrfs_join_transaction*() variants can attach + * | to this running transaction. + * | All other variants will wait for current one to finish and attach to + * | transaction N+1. + * | + * | To next stage: + * | Caller is chosen to commit transaction N, and all other trans handle + * | haven been released. + * V + * Transaction N [[TRANS_STATE_COMMIT_DOING]] + * | + * | The heavy lifting transaction work is started. + * | From running delayed refs (modifying extent tree) to creating pending + * | snapshots, running qgroups. + * | In short, modify supporting trees to reflect modifications of subvolume + * | trees. + * | + * | At this stage, all start_transaction() calls will wait for this + * | transaction to finish and attach to transaction N+1. + * | + * | To next stage: + * | Until all supporting trees are updated. + * V + * Transaction N [[TRANS_STATE_UNBLOCKED]] + * | Transaction N+1 + * | All needed trees are modified, thus we only [[TRANS_STATE_RUNNING]] + * | need to write them back to disk and update | + * | super blocks. | + * | | + * | At this stage, new transaction is allowed to | + * | start. | + * | All new start_transaction() calls will be | + * | attached to transid N+1. | + * | | + * | To next stage: | + * | Until all tree blocks and super blocks are | + * | written to block devices | + * V | + * Transaction N [[TRANS_STATE_COMPLETED]] V + * All tree blocks and super blocks are written. Transaction N+1 + * This transaction is finished and all its [[TRANS_STATE_COMMIT_START]] + * data structures will be cleaned up. | Life goes on + */ static const unsigned int btrfs_blocked_trans_types[TRANS_STATE_MAX] = { [TRANS_STATE_RUNNING] = 0U, - [TRANS_STATE_BLOCKED] = (__TRANS_USERSPACE | - __TRANS_START), - [TRANS_STATE_COMMIT_START] = (__TRANS_USERSPACE | - __TRANS_START | - __TRANS_ATTACH), - [TRANS_STATE_COMMIT_DOING] = (__TRANS_USERSPACE | - __TRANS_START | + [TRANS_STATE_COMMIT_PREP] = 0U, + [TRANS_STATE_COMMIT_START] = (__TRANS_START | __TRANS_ATTACH), + [TRANS_STATE_COMMIT_DOING] = (__TRANS_START | + __TRANS_ATTACH | + __TRANS_JOIN | + __TRANS_JOIN_NOSTART), + [TRANS_STATE_UNBLOCKED] = (__TRANS_START | __TRANS_ATTACH | - __TRANS_JOIN), - [TRANS_STATE_UNBLOCKED] = (__TRANS_USERSPACE | - __TRANS_START | + __TRANS_JOIN | + __TRANS_JOIN_NOLOCK | + __TRANS_JOIN_NOSTART), + [TRANS_STATE_SUPER_COMMITTED] = (__TRANS_START | __TRANS_ATTACH | __TRANS_JOIN | - __TRANS_JOIN_NOLOCK), - [TRANS_STATE_COMPLETED] = (__TRANS_USERSPACE | - __TRANS_START | + __TRANS_JOIN_NOLOCK | + __TRANS_JOIN_NOSTART), + [TRANS_STATE_COMPLETED] = (__TRANS_START | __TRANS_ATTACH | __TRANS_JOIN | - __TRANS_JOIN_NOLOCK), + __TRANS_JOIN_NOLOCK | + __TRANS_JOIN_NOSTART), }; void btrfs_put_transaction(struct btrfs_transaction *transaction) { - WARN_ON(refcount_read(&transaction->use_count) == 0); if (refcount_dec_and_test(&transaction->use_count)) { BUG_ON(!list_empty(&transaction->list)); - WARN_ON(!RB_EMPTY_ROOT(&transaction->delayed_refs.href_root)); + WARN_ON(!xa_empty(&transaction->delayed_refs.head_refs)); + WARN_ON(!xa_empty(&transaction->delayed_refs.dirty_extents)); if (transaction->delayed_refs.pending_csums) btrfs_err(transaction->fs_info, "pending csums is %llu", transaction->delayed_refs.pending_csums); - while (!list_empty(&transaction->pending_chunks)) { - struct extent_map *em; - - em = list_first_entry(&transaction->pending_chunks, - struct extent_map, list); - list_del_init(&em->list); - free_extent_map(em); - } /* * If any block groups are found in ->deleted_bgs then it's * because the transaction was aborted and a commit did not @@ -84,75 +156,66 @@ void btrfs_put_transaction(struct btrfs_transaction *transaction) * discard the physical locations of the block groups. */ while (!list_empty(&transaction->deleted_bgs)) { - struct btrfs_block_group_cache *cache; + struct btrfs_block_group *cache; cache = list_first_entry(&transaction->deleted_bgs, - struct btrfs_block_group_cache, + struct btrfs_block_group, bg_list); + /* + * Not strictly necessary to lock, as no other task will be using a + * block_group on the deleted_bgs list during a transaction abort. + */ + spin_lock(&transaction->fs_info->unused_bgs_lock); list_del_init(&cache->bg_list); - btrfs_put_block_group_trimming(cache); + spin_unlock(&transaction->fs_info->unused_bgs_lock); + btrfs_unfreeze_block_group(cache); btrfs_put_block_group(cache); } + WARN_ON(!list_empty(&transaction->dev_update_list)); kfree(transaction); } } -static void clear_btree_io_tree(struct extent_io_tree *tree) +static noinline void switch_commit_roots(struct btrfs_trans_handle *trans) { - spin_lock(&tree->lock); + struct btrfs_transaction *cur_trans = trans->transaction; + struct btrfs_fs_info *fs_info = trans->fs_info; + struct btrfs_root *root, *tmp; + /* - * Do a single barrier for the waitqueue_active check here, the state - * of the waitqueue should not change once clear_btree_io_tree is - * called. + * At this point no one can be using this transaction to modify any tree + * and no one can start another transaction to modify any tree either. */ - smp_mb(); - while (!RB_EMPTY_ROOT(&tree->state)) { - struct rb_node *node; - struct extent_state *state; - - node = rb_first(&tree->state); - state = rb_entry(node, struct extent_state, rb_node); - rb_erase(&state->rb_node, &tree->state); - RB_CLEAR_NODE(&state->rb_node); - /* - * btree io trees aren't supposed to have tasks waiting for - * changes in the flags of extent states ever. - */ - ASSERT(!waitqueue_active(&state->wq)); - free_extent_state(state); + ASSERT(cur_trans->state == TRANS_STATE_COMMIT_DOING, + "cur_trans->state=%d", cur_trans->state); - cond_resched_lock(&tree->lock); - } - spin_unlock(&tree->lock); -} + down_write(&fs_info->commit_root_sem); -static noinline void switch_commit_roots(struct btrfs_transaction *trans, - struct btrfs_fs_info *fs_info) -{ - struct btrfs_root *root, *tmp; + if (test_bit(BTRFS_FS_RELOC_RUNNING, &fs_info->flags)) + fs_info->last_reloc_trans = trans->transid; - down_write(&fs_info->commit_root_sem); - list_for_each_entry_safe(root, tmp, &trans->switch_commits, + list_for_each_entry_safe(root, tmp, &cur_trans->switch_commits, dirty_list) { list_del_init(&root->dirty_list); free_extent_buffer(root->commit_root); root->commit_root = btrfs_root_node(root); - if (is_fstree(root->objectid)) - btrfs_unpin_free_ino(root); - clear_btree_io_tree(&root->dirty_log_pages); + btrfs_extent_io_tree_release(&root->dirty_log_pages); + btrfs_qgroup_clean_swapped_blocks(root); } /* We can free old roots now. */ - spin_lock(&trans->dropped_roots_lock); - while (!list_empty(&trans->dropped_roots)) { - root = list_first_entry(&trans->dropped_roots, + spin_lock(&cur_trans->dropped_roots_lock); + while (!list_empty(&cur_trans->dropped_roots)) { + root = list_first_entry(&cur_trans->dropped_roots, struct btrfs_root, root_list); list_del_init(&root->root_list); - spin_unlock(&trans->dropped_roots_lock); + spin_unlock(&cur_trans->dropped_roots_lock); + btrfs_free_log(trans, root); btrfs_drop_and_free_fs_root(fs_info, root); - spin_lock(&trans->dropped_roots_lock); + spin_lock(&cur_trans->dropped_roots_lock); } - spin_unlock(&trans->dropped_roots_lock); + spin_unlock(&cur_trans->dropped_roots_lock); + up_write(&fs_info->commit_root_sem); } @@ -182,6 +245,25 @@ static inline int extwriter_counter_read(struct btrfs_transaction *trans) } /* + * To be called after doing the chunk btree updates right after allocating a new + * chunk (after btrfs_chunk_alloc_add_chunk_item() is called), when removing a + * chunk after all chunk btree updates and after finishing the second phase of + * chunk allocation (btrfs_create_pending_block_groups()) in case some block + * group had its chunk item insertion delayed to the second phase. + */ +void btrfs_trans_release_chunk_metadata(struct btrfs_trans_handle *trans) +{ + struct btrfs_fs_info *fs_info = trans->fs_info; + + if (!trans->chunk_bytes_reserved) + return; + + btrfs_block_rsv_release(fs_info, &fs_info->chunk_block_rsv, + trans->chunk_bytes_reserved, NULL); + trans->chunk_bytes_reserved = 0; +} + +/* * either allocate a new transaction or hop into the existing one */ static noinline int join_transaction(struct btrfs_fs_info *fs_info, @@ -192,16 +274,18 @@ static noinline int join_transaction(struct btrfs_fs_info *fs_info, spin_lock(&fs_info->trans_lock); loop: /* The file system has been taken offline. No new transactions. */ - if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) { + if (BTRFS_FS_ERROR(fs_info)) { spin_unlock(&fs_info->trans_lock); return -EROFS; } cur_trans = fs_info->running_transaction; if (cur_trans) { - if (cur_trans->aborted) { + if (TRANS_ABORTED(cur_trans)) { + const int abort_error = cur_trans->aborted; + spin_unlock(&fs_info->trans_lock); - return cur_trans->aborted; + return abort_error; } if (btrfs_blocked_trans_types[cur_trans->state] & type) { spin_unlock(&fs_info->trans_lock); @@ -211,15 +295,18 @@ loop: atomic_inc(&cur_trans->num_writers); extwriter_counter_inc(cur_trans, type); spin_unlock(&fs_info->trans_lock); + btrfs_lockdep_acquire(fs_info, btrfs_trans_num_writers); + btrfs_lockdep_acquire(fs_info, btrfs_trans_num_extwriters); return 0; } spin_unlock(&fs_info->trans_lock); /* - * If we are ATTACH, we just want to catch the current transaction, - * and commit it. If there is no transaction, just return ENOENT. + * If we are ATTACH or TRANS_JOIN_NOSTART, we just want to catch the + * current transaction, and commit it. If there is no transaction, just + * return ENOENT. */ - if (type == TRANS_ATTACH) + if (type == TRANS_ATTACH || type == TRANS_JOIN_NOSTART) return -ENOENT; /* @@ -232,41 +319,47 @@ loop: if (!cur_trans) return -ENOMEM; + btrfs_lockdep_acquire(fs_info, btrfs_trans_num_writers); + btrfs_lockdep_acquire(fs_info, btrfs_trans_num_extwriters); + spin_lock(&fs_info->trans_lock); if (fs_info->running_transaction) { /* * someone started a transaction after we unlocked. Make sure * to redo the checks above */ + btrfs_lockdep_release(fs_info, btrfs_trans_num_extwriters); + btrfs_lockdep_release(fs_info, btrfs_trans_num_writers); kfree(cur_trans); goto loop; - } else if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) { + } else if (BTRFS_FS_ERROR(fs_info)) { spin_unlock(&fs_info->trans_lock); + btrfs_lockdep_release(fs_info, btrfs_trans_num_extwriters); + btrfs_lockdep_release(fs_info, btrfs_trans_num_writers); kfree(cur_trans); return -EROFS; } cur_trans->fs_info = fs_info; + atomic_set(&cur_trans->pending_ordered, 0); + init_waitqueue_head(&cur_trans->pending_wait); atomic_set(&cur_trans->num_writers, 1); extwriter_counter_init(cur_trans, type); init_waitqueue_head(&cur_trans->writer_wait); init_waitqueue_head(&cur_trans->commit_wait); - init_waitqueue_head(&cur_trans->pending_wait); cur_trans->state = TRANS_STATE_RUNNING; /* * One for this trans handle, one so it will live on until we * commit the transaction. */ refcount_set(&cur_trans->use_count, 2); - atomic_set(&cur_trans->pending_ordered, 0); cur_trans->flags = 0; - cur_trans->start_time = get_seconds(); + cur_trans->start_time = ktime_get_seconds(); memset(&cur_trans->delayed_refs, 0, sizeof(cur_trans->delayed_refs)); - cur_trans->delayed_refs.href_root = RB_ROOT; - cur_trans->delayed_refs.dirty_extent_root = RB_ROOT; - atomic_set(&cur_trans->delayed_refs.num_entries, 0); + xa_init(&cur_trans->delayed_refs.head_refs); + xa_init(&cur_trans->delayed_refs.dirty_extents); /* * although the tree mod log is per file system and not per transaction, @@ -282,20 +375,21 @@ loop: spin_lock_init(&cur_trans->delayed_refs.lock); INIT_LIST_HEAD(&cur_trans->pending_snapshots); - INIT_LIST_HEAD(&cur_trans->pending_chunks); + INIT_LIST_HEAD(&cur_trans->dev_update_list); INIT_LIST_HEAD(&cur_trans->switch_commits); INIT_LIST_HEAD(&cur_trans->dirty_bgs); INIT_LIST_HEAD(&cur_trans->io_bgs); INIT_LIST_HEAD(&cur_trans->dropped_roots); mutex_init(&cur_trans->cache_write_mutex); - cur_trans->num_dirty_bgs = 0; spin_lock_init(&cur_trans->dirty_bgs_lock); INIT_LIST_HEAD(&cur_trans->deleted_bgs); spin_lock_init(&cur_trans->dropped_roots_lock); list_add_tail(&cur_trans->list, &fs_info->trans_list); - extent_io_tree_init(&cur_trans->dirty_pages, - fs_info->btree_inode); - fs_info->generation++; + btrfs_extent_io_tree_init(fs_info, &cur_trans->dirty_pages, + IO_TREE_TRANS_DIRTY_PAGES); + btrfs_extent_io_tree_init(fs_info, &cur_trans->pinned_extents, + IO_TREE_FS_PINNED_EXTENTS); + btrfs_set_fs_generation(fs_info, fs_info->generation + 1); cur_trans->transid = fs_info->generation; fs_info->running_transaction = cur_trans; cur_trans->aborted = 0; @@ -305,21 +399,21 @@ loop: } /* - * this does all the record keeping required to make sure that a reference - * counted root is properly recorded in a given transaction. This is required - * to make sure the old root from before we joined the transaction is deleted - * when the transaction commits + * This does all the record keeping required to make sure that a shareable root + * is properly recorded in a given transaction. This is required to make sure + * the old root from before we joined the transaction is deleted when the + * transaction commits. */ static int record_root_in_trans(struct btrfs_trans_handle *trans, struct btrfs_root *root, - int force) + bool force) { struct btrfs_fs_info *fs_info = root->fs_info; + int ret = 0; - if ((test_bit(BTRFS_ROOT_REF_COWS, &root->state) && - root->last_trans < trans->transid) || force) { - WARN_ON(root == fs_info->extent_root); - WARN_ON(root->commit_root != root->node); + if ((test_bit(BTRFS_ROOT_SHAREABLE, &root->state) && + btrfs_get_root_last_trans(root) < trans->transid) || force) { + WARN_ON(!force && root->commit_root != root->node); /* * see below for IN_TRANS_SETUP usage rules @@ -334,15 +428,15 @@ static int record_root_in_trans(struct btrfs_trans_handle *trans, smp_wmb(); spin_lock(&fs_info->fs_roots_radix_lock); - if (root->last_trans == trans->transid && !force) { + if (btrfs_get_root_last_trans(root) == trans->transid && !force) { spin_unlock(&fs_info->fs_roots_radix_lock); return 0; } radix_tree_tag_set(&fs_info->fs_roots_radix, - (unsigned long)root->root_key.objectid, + (unsigned long)btrfs_root_id(root), BTRFS_ROOT_TRANS_TAG); spin_unlock(&fs_info->fs_roots_radix_lock); - root->last_trans = trans->transid; + btrfs_set_root_last_trans(root, trans->transid); /* this is pretty tricky. We don't want to * take the relocation lock in btrfs_record_root_in_trans @@ -363,11 +457,11 @@ static int record_root_in_trans(struct btrfs_trans_handle *trans, * lock. smp_wmb() makes sure that all the writes above are * done before we pop in the zero below */ - btrfs_init_reloc_root(trans, root); + ret = btrfs_init_reloc_root(trans, root); smp_mb__before_atomic(); clear_bit(BTRFS_ROOT_IN_TRANS_SETUP, &root->state); } - return 0; + return ret; } @@ -385,7 +479,7 @@ void btrfs_add_dropped_root(struct btrfs_trans_handle *trans, /* Make sure we don't try to update the root at commit time */ spin_lock(&fs_info->fs_roots_radix_lock); radix_tree_tag_clear(&fs_info->fs_roots_radix, - (unsigned long)root->root_key.objectid, + (unsigned long)btrfs_root_id(root), BTRFS_ROOT_TRANS_TAG); spin_unlock(&fs_info->fs_roots_radix_lock); } @@ -394,8 +488,9 @@ int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans, struct btrfs_root *root) { struct btrfs_fs_info *fs_info = root->fs_info; + int ret; - if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state)) + if (!test_bit(BTRFS_ROOT_SHAREABLE, &root->state)) return 0; /* @@ -403,22 +498,22 @@ int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans, * and barriers */ smp_rmb(); - if (root->last_trans == trans->transid && + if (btrfs_get_root_last_trans(root) == trans->transid && !test_bit(BTRFS_ROOT_IN_TRANS_SETUP, &root->state)) return 0; mutex_lock(&fs_info->reloc_mutex); - record_root_in_trans(trans, root, 0); + ret = record_root_in_trans(trans, root, 0); mutex_unlock(&fs_info->reloc_mutex); - return 0; + return ret; } static inline int is_transaction_blocked(struct btrfs_transaction *trans) { - return (trans->state >= TRANS_STATE_BLOCKED && + return (trans->state >= TRANS_STATE_COMMIT_START && trans->state < TRANS_STATE_UNBLOCKED && - !trans->aborted); + !TRANS_ABORTED(trans)); } /* wait for commit against the current transaction to become unblocked @@ -435,28 +530,25 @@ static void wait_current_trans(struct btrfs_fs_info *fs_info) refcount_inc(&cur_trans->use_count); spin_unlock(&fs_info->trans_lock); + btrfs_might_wait_for_state(fs_info, BTRFS_LOCKDEP_TRANS_UNBLOCKED); wait_event(fs_info->transaction_wait, cur_trans->state >= TRANS_STATE_UNBLOCKED || - cur_trans->aborted); + TRANS_ABORTED(cur_trans)); btrfs_put_transaction(cur_trans); } else { spin_unlock(&fs_info->trans_lock); } } -static int may_wait_transaction(struct btrfs_fs_info *fs_info, int type) +static bool may_wait_transaction(struct btrfs_fs_info *fs_info, int type) { if (test_bit(BTRFS_FS_LOG_RECOVERING, &fs_info->flags)) - return 0; - - if (type == TRANS_USERSPACE) - return 1; + return false; - if (type == TRANS_START && - !atomic_read(&fs_info->open_ioctl_trans)) - return 1; + if (type == TRANS_START) + return true; - return 0; + return false; } static inline bool need_reserve_reloc_root(struct btrfs_root *root) @@ -464,39 +556,68 @@ static inline bool need_reserve_reloc_root(struct btrfs_root *root) struct btrfs_fs_info *fs_info = root->fs_info; if (!fs_info->reloc_ctl || - !test_bit(BTRFS_ROOT_REF_COWS, &root->state) || - root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID || + !test_bit(BTRFS_ROOT_SHAREABLE, &root->state) || + btrfs_root_id(root) == BTRFS_TREE_RELOC_OBJECTID || root->reloc_root) return false; return true; } +static int btrfs_reserve_trans_metadata(struct btrfs_fs_info *fs_info, + enum btrfs_reserve_flush_enum flush, + u64 num_bytes, + u64 *delayed_refs_bytes) +{ + struct btrfs_space_info *si = fs_info->trans_block_rsv.space_info; + u64 bytes = num_bytes + *delayed_refs_bytes; + int ret; + + /* + * We want to reserve all the bytes we may need all at once, so we only + * do 1 enospc flushing cycle per transaction start. + */ + ret = btrfs_reserve_metadata_bytes(si, bytes, flush); + + /* + * If we are an emergency flush, which can steal from the global block + * reserve, then attempt to not reserve space for the delayed refs, as + * we will consume space for them from the global block reserve. + */ + if (ret && flush == BTRFS_RESERVE_FLUSH_ALL_STEAL) { + bytes -= *delayed_refs_bytes; + *delayed_refs_bytes = 0; + ret = btrfs_reserve_metadata_bytes(si, bytes, flush); + } + + return ret; +} + static struct btrfs_trans_handle * start_transaction(struct btrfs_root *root, unsigned int num_items, unsigned int type, enum btrfs_reserve_flush_enum flush, bool enforce_qgroups) { struct btrfs_fs_info *fs_info = root->fs_info; - + struct btrfs_block_rsv *delayed_refs_rsv = &fs_info->delayed_refs_rsv; + struct btrfs_block_rsv *trans_rsv = &fs_info->trans_block_rsv; struct btrfs_trans_handle *h; struct btrfs_transaction *cur_trans; u64 num_bytes = 0; u64 qgroup_reserved = 0; + u64 delayed_refs_bytes = 0; bool reloc_reserved = false; + bool do_chunk_alloc = false; int ret; - /* Send isn't supposed to start transactions. */ - ASSERT(current->journal_info != BTRFS_SEND_TRANS_STUB); - - if (test_bit(BTRFS_FS_STATE_ERROR, &fs_info->fs_state)) + if (BTRFS_FS_ERROR(fs_info)) return ERR_PTR(-EROFS); if (current->journal_info) { WARN_ON(type & TRANS_EXTWRITERS); h = current->journal_info; - h->use_count++; - WARN_ON(h->use_count > 2); + refcount_inc(&h->use_count); + WARN_ON(refcount_read(&h->use_count) > 2); h->orig_rsv = h->block_rsv; h->block_rsv = NULL; goto got_it; @@ -508,12 +629,27 @@ start_transaction(struct btrfs_root *root, unsigned int num_items, */ if (num_items && root != fs_info->chunk_root) { qgroup_reserved = num_items * fs_info->nodesize; - ret = btrfs_qgroup_reserve_meta(root, qgroup_reserved, - enforce_qgroups); + /* + * Use prealloc for now, as there might be a currently running + * transaction that could free this reserved space prematurely + * by committing. + */ + ret = btrfs_qgroup_reserve_meta_prealloc(root, qgroup_reserved, + enforce_qgroups, false); if (ret) return ERR_PTR(ret); - num_bytes = btrfs_calc_trans_metadata_size(fs_info, num_items); + num_bytes = btrfs_calc_insert_metadata_size(fs_info, num_items); + /* + * If we plan to insert/update/delete "num_items" from a btree, + * we will also generate delayed refs for extent buffers in the + * respective btree paths, so reserve space for the delayed refs + * that will be generated by the caller as it modifies btrees. + * Try to reserve them to avoid excessive use of the global + * block reserve. + */ + delayed_refs_bytes = btrfs_calc_delayed_ref_bytes(fs_info, num_items); + /* * Do the reservation for the relocation root creation */ @@ -522,8 +658,25 @@ start_transaction(struct btrfs_root *root, unsigned int num_items, reloc_reserved = true; } - ret = btrfs_block_rsv_add(root, &fs_info->trans_block_rsv, - num_bytes, flush); + ret = btrfs_reserve_trans_metadata(fs_info, flush, num_bytes, + &delayed_refs_bytes); + if (ret) + goto reserve_fail; + + btrfs_block_rsv_add_bytes(trans_rsv, num_bytes, true); + + if (trans_rsv->space_info->force_alloc) + do_chunk_alloc = true; + } else if (num_items == 0 && flush == BTRFS_RESERVE_FLUSH_ALL && + !btrfs_block_rsv_full(delayed_refs_rsv)) { + /* + * Some people call with btrfs_start_transaction(root, 0) + * because they can be throttled, but have some other mechanism + * for reserving space. We still want these guys to refill the + * delayed block_rsv so just add 1 items worth of reservation + * here. + */ + ret = btrfs_delayed_refs_rsv_refill(fs_info, flush); if (ret) goto reserve_fail; } @@ -554,7 +707,8 @@ again: ret = join_transaction(fs_info, type); if (ret == -EBUSY) { wait_current_trans(fs_info); - if (unlikely(type == TRANS_ATTACH)) + if (unlikely(type == TRANS_ATTACH || + type == TRANS_JOIN_NOSTART)) ret = -ENOENT; } } while (ret == -EBUSY); @@ -566,16 +720,15 @@ again: h->transid = cur_trans->transid; h->transaction = cur_trans; - h->root = root; - h->use_count = 1; + refcount_set(&h->use_count, 1); h->fs_info = root->fs_info; h->type = type; - h->can_flush_pending_bgs = true; INIT_LIST_HEAD(&h->new_bgs); + btrfs_init_metadata_block_rsv(fs_info, &h->delayed_rsv, BTRFS_BLOCK_RSV_DELOPS); smp_mb(); - if (cur_trans->state >= TRANS_STATE_BLOCKED && + if (cur_trans->state >= TRANS_STATE_COMMIT_START && may_wait_transaction(fs_info, type)) { current->journal_info = h; btrfs_commit_transaction(h); @@ -585,16 +738,64 @@ again: if (num_bytes) { trace_btrfs_space_reservation(fs_info, "transaction", h->transid, num_bytes, 1); - h->block_rsv = &fs_info->trans_block_rsv; + h->block_rsv = trans_rsv; h->bytes_reserved = num_bytes; + if (delayed_refs_bytes > 0) { + trace_btrfs_space_reservation(fs_info, + "local_delayed_refs_rsv", + h->transid, + delayed_refs_bytes, 1); + h->delayed_refs_bytes_reserved = delayed_refs_bytes; + btrfs_block_rsv_add_bytes(&h->delayed_rsv, delayed_refs_bytes, true); + delayed_refs_bytes = 0; + } h->reloc_reserved = reloc_reserved; } got_it: - btrfs_record_root_in_trans(h, root); - - if (!current->journal_info && type != TRANS_USERSPACE) + if (!current->journal_info) current->journal_info = h; + + /* + * If the space_info is marked ALLOC_FORCE then we'll get upgraded to + * ALLOC_FORCE the first run through, and then we won't allocate for + * anybody else who races in later. We don't care about the return + * value here. + */ + if (do_chunk_alloc && num_bytes) { + struct btrfs_space_info *space_info = h->block_rsv->space_info; + u64 flags = space_info->flags; + + btrfs_chunk_alloc(h, space_info, btrfs_get_alloc_profile(fs_info, flags), + CHUNK_ALLOC_NO_FORCE); + } + + /* + * btrfs_record_root_in_trans() needs to alloc new extents, and may + * call btrfs_join_transaction() while we're also starting a + * transaction. + * + * Thus it need to be called after current->journal_info initialized, + * or we can deadlock. + */ + ret = btrfs_record_root_in_trans(h, root); + if (ret) { + /* + * The transaction handle is fully initialized and linked with + * other structures so it needs to be ended in case of errors, + * not just freed. + */ + btrfs_end_transaction(h); + goto reserve_fail; + } + /* + * Now that we have found a transaction to be a part of, convert the + * qgroup reservation from prealloc to pertrans. A different transaction + * can't race in and free our pertrans out from under us. + */ + if (qgroup_reserved) + btrfs_qgroup_convert_reserved_meta(root, qgroup_reserved); + return h; join_fail: @@ -603,10 +804,11 @@ join_fail: kmem_cache_free(btrfs_trans_handle_cachep, h); alloc_fail: if (num_bytes) - btrfs_block_rsv_release(fs_info, &fs_info->trans_block_rsv, - num_bytes); + btrfs_block_rsv_release(fs_info, trans_rsv, num_bytes, NULL); + if (delayed_refs_bytes) + btrfs_space_info_free_bytes_may_use(trans_rsv->space_info, delayed_refs_bytes); reserve_fail: - btrfs_qgroup_free_meta(root, qgroup_reserved); + btrfs_qgroup_free_meta_prealloc(root, qgroup_reserved); return ERR_PTR(ret); } @@ -619,51 +821,10 @@ struct btrfs_trans_handle *btrfs_start_transaction(struct btrfs_root *root, struct btrfs_trans_handle *btrfs_start_transaction_fallback_global_rsv( struct btrfs_root *root, - unsigned int num_items, - int min_factor) -{ - struct btrfs_fs_info *fs_info = root->fs_info; - struct btrfs_trans_handle *trans; - u64 num_bytes; - int ret; - - /* - * We have two callers: unlink and block group removal. The - * former should succeed even if we will temporarily exceed - * quota and the latter operates on the extent root so - * qgroup enforcement is ignored anyway. - */ - trans = start_transaction(root, num_items, TRANS_START, - BTRFS_RESERVE_FLUSH_ALL, false); - if (!IS_ERR(trans) || PTR_ERR(trans) != -ENOSPC) - return trans; - - trans = btrfs_start_transaction(root, 0); - if (IS_ERR(trans)) - return trans; - - num_bytes = btrfs_calc_trans_metadata_size(fs_info, num_items); - ret = btrfs_cond_migrate_bytes(fs_info, &fs_info->trans_block_rsv, - num_bytes, min_factor); - if (ret) { - btrfs_end_transaction(trans); - return ERR_PTR(ret); - } - - trans->block_rsv = &fs_info->trans_block_rsv; - trans->bytes_reserved = num_bytes; - trace_btrfs_space_reservation(fs_info, "transaction", - trans->transid, num_bytes, 1); - - return trans; -} - -struct btrfs_trans_handle *btrfs_start_transaction_lflush( - struct btrfs_root *root, unsigned int num_items) { return start_transaction(root, num_items, TRANS_START, - BTRFS_RESERVE_FLUSH_LIMIT, true); + BTRFS_RESERVE_FLUSH_ALL_STEAL, false); } struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root) @@ -672,20 +833,27 @@ struct btrfs_trans_handle *btrfs_join_transaction(struct btrfs_root *root) true); } -struct btrfs_trans_handle *btrfs_join_transaction_nolock(struct btrfs_root *root) +struct btrfs_trans_handle *btrfs_join_transaction_spacecache(struct btrfs_root *root) { return start_transaction(root, 0, TRANS_JOIN_NOLOCK, BTRFS_RESERVE_NO_FLUSH, true); } -struct btrfs_trans_handle *btrfs_start_ioctl_transaction(struct btrfs_root *root) +/* + * Similar to regular join but it never starts a transaction when none is + * running or when there's a running one at a state >= TRANS_STATE_UNBLOCKED. + * This is similar to btrfs_attach_transaction() but it allows the join to + * happen if the transaction commit already started but it's not yet in the + * "doing" phase (the state is < TRANS_STATE_COMMIT_DOING). + */ +struct btrfs_trans_handle *btrfs_join_transaction_nostart(struct btrfs_root *root) { - return start_transaction(root, 0, TRANS_USERSPACE, + return start_transaction(root, 0, TRANS_JOIN_NOSTART, BTRFS_RESERVE_NO_FLUSH, true); } /* - * btrfs_attach_transaction() - catch the running transaction + * Catch the running transaction. * * It is used when we want to commit the current the transaction, but * don't want to start a new one. @@ -704,9 +872,9 @@ struct btrfs_trans_handle *btrfs_attach_transaction(struct btrfs_root *root) } /* - * btrfs_attach_transaction_barrier() - catch the running transaction + * Catch the running transaction. * - * It is similar to the above function, the differentia is this one + * It is similar to the above function, the difference is this one * will wait for all the inactive transactions until they fully * complete. */ @@ -717,16 +885,61 @@ btrfs_attach_transaction_barrier(struct btrfs_root *root) trans = start_transaction(root, 0, TRANS_ATTACH, BTRFS_RESERVE_NO_FLUSH, true); - if (IS_ERR(trans) && PTR_ERR(trans) == -ENOENT) - btrfs_wait_for_commit(root->fs_info, 0); + if (trans == ERR_PTR(-ENOENT)) { + int ret; + + ret = btrfs_wait_for_commit(root->fs_info, 0); + if (ret) + return ERR_PTR(ret); + } return trans; } -/* wait for a transaction commit to be fully complete */ -static noinline void wait_for_commit(struct btrfs_transaction *commit) +/* Wait for a transaction commit to reach at least the given state. */ +static noinline void wait_for_commit(struct btrfs_transaction *commit, + const enum btrfs_trans_state min_state) { - wait_event(commit->commit_wait, commit->state == TRANS_STATE_COMPLETED); + struct btrfs_fs_info *fs_info = commit->fs_info; + u64 transid = commit->transid; + bool put = false; + + /* + * At the moment this function is called with min_state either being + * TRANS_STATE_COMPLETED or TRANS_STATE_SUPER_COMMITTED. + */ + if (min_state == TRANS_STATE_COMPLETED) + btrfs_might_wait_for_state(fs_info, BTRFS_LOCKDEP_TRANS_COMPLETED); + else + btrfs_might_wait_for_state(fs_info, BTRFS_LOCKDEP_TRANS_SUPER_COMMITTED); + + while (1) { + wait_event(commit->commit_wait, commit->state >= min_state); + if (put) + btrfs_put_transaction(commit); + + if (min_state < TRANS_STATE_COMPLETED) + break; + + /* + * A transaction isn't really completed until all of the + * previous transactions are completed, but with fsync we can + * end up with SUPER_COMMITTED transactions before a COMPLETED + * transaction. Wait for those. + */ + + spin_lock(&fs_info->trans_lock); + commit = list_first_entry_or_null(&fs_info->trans_list, + struct btrfs_transaction, + list); + if (!commit || commit->transid > transid) { + spin_unlock(&fs_info->trans_lock); + break; + } + refcount_inc(&commit->use_count); + put = true; + spin_unlock(&fs_info->trans_lock); + } } int btrfs_wait_for_commit(struct btrfs_fs_info *fs_info, u64 transid) @@ -735,7 +948,7 @@ int btrfs_wait_for_commit(struct btrfs_fs_info *fs_info, u64 transid) int ret = 0; if (transid) { - if (transid <= fs_info->last_trans_committed) + if (transid <= btrfs_get_last_trans_committed(fs_info)) goto out; /* find specified transaction */ @@ -759,7 +972,7 @@ int btrfs_wait_for_commit(struct btrfs_fs_info *fs_info, u64 transid) * raced with btrfs_commit_transaction */ if (!cur_trans) { - if (transid > fs_info->last_trans_committed) + if (transid > btrfs_get_last_trans_committed(fs_info)) ret = -EINVAL; goto out; } @@ -781,7 +994,8 @@ int btrfs_wait_for_commit(struct btrfs_fs_info *fs_info, u64 transid) goto out; /* nothing committing|committed */ } - wait_for_commit(cur_trans); + wait_for_commit(cur_trans, TRANS_STATE_COMPLETED); + ret = cur_trans->aborted; btrfs_put_transaction(cur_trans); out: return ret; @@ -789,42 +1003,60 @@ out: void btrfs_throttle(struct btrfs_fs_info *fs_info) { - if (!atomic_read(&fs_info->open_ioctl_trans)) - wait_current_trans(fs_info); + wait_current_trans(fs_info); } -static int should_end_transaction(struct btrfs_trans_handle *trans) +bool btrfs_should_end_transaction(struct btrfs_trans_handle *trans) { - struct btrfs_fs_info *fs_info = trans->fs_info; + struct btrfs_transaction *cur_trans = trans->transaction; + + if (cur_trans->state >= TRANS_STATE_COMMIT_START || + test_bit(BTRFS_DELAYED_REFS_FLUSHING, &cur_trans->delayed_refs.flags)) + return true; - if (fs_info->global_block_rsv.space_info->full && - btrfs_check_space_for_delayed_refs(trans, fs_info)) - return 1; + if (btrfs_check_space_for_delayed_refs(trans->fs_info)) + return true; - return !!btrfs_block_rsv_check(&fs_info->global_block_rsv, 5); + return !!btrfs_block_rsv_check(&trans->fs_info->global_block_rsv, 50); } -int btrfs_should_end_transaction(struct btrfs_trans_handle *trans) +static void btrfs_trans_release_metadata(struct btrfs_trans_handle *trans) + { - struct btrfs_transaction *cur_trans = trans->transaction; struct btrfs_fs_info *fs_info = trans->fs_info; - int updates; - int err; - smp_mb(); - if (cur_trans->state >= TRANS_STATE_BLOCKED || - cur_trans->delayed_refs.flushing) - return 1; - - updates = trans->delayed_ref_updates; - trans->delayed_ref_updates = 0; - if (updates) { - err = btrfs_run_delayed_refs(trans, fs_info, updates * 2); - if (err) /* Error code will also eval true */ - return err; + if (!trans->block_rsv) { + ASSERT(trans->bytes_reserved == 0, + "trans->bytes_reserved=%llu", trans->bytes_reserved); + ASSERT(trans->delayed_refs_bytes_reserved == 0, + "trans->delayed_refs_bytes_reserved=%llu", + trans->delayed_refs_bytes_reserved); + return; + } + + if (!trans->bytes_reserved) { + ASSERT(trans->delayed_refs_bytes_reserved == 0, + "trans->delayed_refs_bytes_reserved=%llu", + trans->delayed_refs_bytes_reserved); + return; } - return should_end_transaction(trans); + ASSERT(trans->block_rsv == &fs_info->trans_block_rsv); + trace_btrfs_space_reservation(fs_info, "transaction", + trans->transid, trans->bytes_reserved, 0); + btrfs_block_rsv_release(fs_info, trans->block_rsv, + trans->bytes_reserved, NULL); + trans->bytes_reserved = 0; + + if (!trans->delayed_refs_bytes_reserved) + return; + + trace_btrfs_space_reservation(fs_info, "local_delayed_refs_rsv", + trans->transid, + trans->delayed_refs_bytes_reserved, 0); + btrfs_block_rsv_release(fs_info, &trans->delayed_rsv, + trans->delayed_refs_bytes_reserved, NULL); + trans->delayed_refs_bytes_reserved = 0; } static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, @@ -832,63 +1064,21 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, { struct btrfs_fs_info *info = trans->fs_info; struct btrfs_transaction *cur_trans = trans->transaction; - u64 transid = trans->transid; - unsigned long cur = trans->delayed_ref_updates; - int lock = (trans->type != TRANS_JOIN_NOLOCK); - int err = 0; - int must_run_delayed_refs = 0; - - if (trans->use_count > 1) { - trans->use_count--; + int ret = 0; + + if (refcount_read(&trans->use_count) > 1) { + refcount_dec(&trans->use_count); trans->block_rsv = trans->orig_rsv; return 0; } - btrfs_trans_release_metadata(trans, info); - trans->block_rsv = NULL; - - if (!list_empty(&trans->new_bgs)) - btrfs_create_pending_block_groups(trans, info); - - trans->delayed_ref_updates = 0; - if (!trans->sync) { - must_run_delayed_refs = - btrfs_should_throttle_delayed_refs(trans, info); - cur = max_t(unsigned long, cur, 32); - - /* - * don't make the caller wait if they are from a NOLOCK - * or ATTACH transaction, it will deadlock with commit - */ - if (must_run_delayed_refs == 1 && - (trans->type & (__TRANS_JOIN_NOLOCK | __TRANS_ATTACH))) - must_run_delayed_refs = 2; - } - - btrfs_trans_release_metadata(trans, info); + btrfs_trans_release_metadata(trans); trans->block_rsv = NULL; - if (!list_empty(&trans->new_bgs)) - btrfs_create_pending_block_groups(trans, info); + btrfs_create_pending_block_groups(trans); btrfs_trans_release_chunk_metadata(trans); - if (lock && !atomic_read(&info->open_ioctl_trans) && - should_end_transaction(trans) && - READ_ONCE(cur_trans->state) == TRANS_STATE_RUNNING) { - spin_lock(&info->trans_lock); - if (cur_trans->state == TRANS_STATE_RUNNING) - cur_trans->state = TRANS_STATE_BLOCKED; - spin_unlock(&info->trans_lock); - } - - if (lock && READ_ONCE(cur_trans->state) == TRANS_STATE_BLOCKED) { - if (throttle) - return btrfs_commit_transaction(trans); - else - wake_up_process(info->transaction_kthread); - } - if (trans->type & __TRANS_FREEZABLE) sb_end_intwrite(info->sb); @@ -897,12 +1087,11 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, atomic_dec(&cur_trans->num_writers); extwriter_counter_dec(cur_trans, trans->type); - /* - * Make sure counter is updated before we wake up waiters. - */ - smp_mb(); - if (waitqueue_active(&cur_trans->writer_wait)) - wake_up(&cur_trans->writer_wait); + cond_wake_up(&cur_trans->writer_wait); + + btrfs_lockdep_release(info, btrfs_trans_num_extwriters); + btrfs_lockdep_release(info, btrfs_trans_num_writers); + btrfs_put_transaction(cur_trans); if (current->journal_info == trans) @@ -911,18 +1100,16 @@ static int __btrfs_end_transaction(struct btrfs_trans_handle *trans, if (throttle) btrfs_run_delayed_iputs(info); - if (trans->aborted || - test_bit(BTRFS_FS_STATE_ERROR, &info->fs_state)) { + if (TRANS_ABORTED(trans) || BTRFS_FS_ERROR(info)) { wake_up_process(info->transaction_kthread); - err = -EIO; + if (TRANS_ABORTED(trans)) + ret = trans->aborted; + else + ret = -EROFS; } kmem_cache_free(btrfs_trans_handle_cachep, trans); - if (must_run_delayed_refs) { - btrfs_async_run_delayed_refs(info, cur, transid, - must_run_delayed_refs == 1); - } - return err; + return ret; } int btrfs_end_transaction(struct btrfs_trans_handle *trans) @@ -943,20 +1130,19 @@ int btrfs_end_transaction_throttle(struct btrfs_trans_handle *trans) int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info, struct extent_io_tree *dirty_pages, int mark) { - int err = 0; - int werr = 0; + int ret = 0; struct address_space *mapping = fs_info->btree_inode->i_mapping; struct extent_state *cached_state = NULL; u64 start = 0; u64 end; - while (!find_first_extent_bit(dirty_pages, start, &start, &end, - mark, &cached_state)) { + while (btrfs_find_first_extent_bit(dirty_pages, start, &start, &end, + mark, &cached_state)) { bool wait_writeback = false; - err = convert_extent_bit(dirty_pages, start, end, - EXTENT_NEED_WAIT, - mark, &cached_state); + ret = btrfs_convert_extent_bit(dirty_pages, start, end, + EXTENT_NEED_WAIT, + mark, &cached_state); /* * convert_extent_bit can return -ENOMEM, which is most of the * time a temporary error. So when it happens, ignore the error @@ -968,24 +1154,24 @@ int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info, * superblock that points to btree nodes/leafs for which * writeback hasn't finished yet (and without errors). * We cleanup any entries left in the io tree when committing - * the transaction (through clear_btree_io_tree()). + * the transaction (through extent_io_tree_release()). */ - if (err == -ENOMEM) { - err = 0; + if (ret == -ENOMEM) { + ret = 0; wait_writeback = true; } - if (!err) - err = filemap_fdatawrite_range(mapping, start, end); - if (err) - werr = err; - else if (wait_writeback) - werr = filemap_fdatawait_range(mapping, start, end); - free_extent_state(cached_state); + if (!ret) + ret = filemap_fdatawrite_range(mapping, start, end); + if (!ret && wait_writeback) + btrfs_btree_wait_writeback_range(fs_info, start, end); + btrfs_free_extent_state(cached_state); + if (ret) + break; cached_state = NULL; cond_resched(); start = end + 1; } - return werr; + return ret; } /* @@ -997,55 +1183,50 @@ int btrfs_write_marked_extents(struct btrfs_fs_info *fs_info, static int __btrfs_wait_marked_extents(struct btrfs_fs_info *fs_info, struct extent_io_tree *dirty_pages) { - int err = 0; - int werr = 0; - struct address_space *mapping = fs_info->btree_inode->i_mapping; struct extent_state *cached_state = NULL; u64 start = 0; u64 end; + int ret = 0; - while (!find_first_extent_bit(dirty_pages, start, &start, &end, - EXTENT_NEED_WAIT, &cached_state)) { + while (btrfs_find_first_extent_bit(dirty_pages, start, &start, &end, + EXTENT_NEED_WAIT, &cached_state)) { /* * Ignore -ENOMEM errors returned by clear_extent_bit(). * When committing the transaction, we'll remove any entries * left in the io tree. For a log commit, we don't remove them * after committing the log because the tree can be accessed * concurrently - we do it only at transaction commit time when - * it's safe to do it (through clear_btree_io_tree()). + * it's safe to do it (through extent_io_tree_release()). */ - err = clear_extent_bit(dirty_pages, start, end, - EXTENT_NEED_WAIT, - 0, 0, &cached_state, GFP_NOFS); - if (err == -ENOMEM) - err = 0; - if (!err) - err = filemap_fdatawait_range(mapping, start, end); - if (err) - werr = err; - free_extent_state(cached_state); + ret = btrfs_clear_extent_bit(dirty_pages, start, end, + EXTENT_NEED_WAIT, &cached_state); + if (ret == -ENOMEM) + ret = 0; + if (!ret) + btrfs_btree_wait_writeback_range(fs_info, start, end); + btrfs_free_extent_state(cached_state); + if (ret) + break; cached_state = NULL; cond_resched(); start = end + 1; } - if (err) - werr = err; - return werr; + return ret; } -int btrfs_wait_extents(struct btrfs_fs_info *fs_info, +static int btrfs_wait_extents(struct btrfs_fs_info *fs_info, struct extent_io_tree *dirty_pages) { bool errors = false; - int err; + int ret; - err = __btrfs_wait_marked_extents(fs_info, dirty_pages); + ret = __btrfs_wait_marked_extents(fs_info, dirty_pages); if (test_and_clear_bit(BTRFS_FS_BTREE_ERR, &fs_info->flags)) errors = true; - if (errors && !err) - err = -EIO; - return err; + if (errors && !ret) + ret = -EIO; + return ret; } int btrfs_wait_tree_log_extents(struct btrfs_root *log_root, int mark) @@ -1053,59 +1234,53 @@ int btrfs_wait_tree_log_extents(struct btrfs_root *log_root, int mark) struct btrfs_fs_info *fs_info = log_root->fs_info; struct extent_io_tree *dirty_pages = &log_root->dirty_log_pages; bool errors = false; - int err; + int ret; - ASSERT(log_root->root_key.objectid == BTRFS_TREE_LOG_OBJECTID); + ASSERT(btrfs_root_id(log_root) == BTRFS_TREE_LOG_OBJECTID, + "root_id(log_root)=%llu", btrfs_root_id(log_root)); - err = __btrfs_wait_marked_extents(fs_info, dirty_pages); - if ((mark & EXTENT_DIRTY) && + ret = __btrfs_wait_marked_extents(fs_info, dirty_pages); + if ((mark & EXTENT_DIRTY_LOG1) && test_and_clear_bit(BTRFS_FS_LOG1_ERR, &fs_info->flags)) errors = true; - if ((mark & EXTENT_NEW) && + if ((mark & EXTENT_DIRTY_LOG2) && test_and_clear_bit(BTRFS_FS_LOG2_ERR, &fs_info->flags)) errors = true; - if (errors && !err) - err = -EIO; - return err; + if (errors && !ret) + ret = -EIO; + return ret; } /* - * when btree blocks are allocated, they have some corresponding bits set for - * them in one of two extent_io trees. This is used to make sure all of - * those extents are on disk for transaction or log commit + * When btree blocks are allocated the corresponding extents are marked dirty. + * This function ensures such extents are persisted on disk for transaction or + * log commit. + * + * @trans: transaction whose dirty pages we'd like to write */ -static int btrfs_write_and_wait_marked_extents(struct btrfs_fs_info *fs_info, - struct extent_io_tree *dirty_pages, int mark) +static int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans) { int ret; int ret2; + struct extent_io_tree *dirty_pages = &trans->transaction->dirty_pages; + struct btrfs_fs_info *fs_info = trans->fs_info; struct blk_plug plug; blk_start_plug(&plug); - ret = btrfs_write_marked_extents(fs_info, dirty_pages, mark); + ret = btrfs_write_marked_extents(fs_info, dirty_pages, EXTENT_DIRTY); blk_finish_plug(&plug); ret2 = btrfs_wait_extents(fs_info, dirty_pages); + btrfs_extent_io_tree_release(&trans->transaction->dirty_pages); + if (ret) return ret; - if (ret2) + else if (ret2) return ret2; - return 0; -} - -static int btrfs_write_and_wait_transaction(struct btrfs_trans_handle *trans, - struct btrfs_fs_info *fs_info) -{ - int ret; - - ret = btrfs_write_and_wait_marked_extents(fs_info, - &trans->transaction->dirty_pages, - EXTENT_DIRTY); - clear_btree_io_tree(&trans->transaction->dirty_pages); - - return ret; + else + return 0; } /* @@ -1155,70 +1330,76 @@ static int update_cowonly_root(struct btrfs_trans_handle *trans, * failures will cause the file system to go offline. We still need * to clean up the delayed refs. */ -static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans, - struct btrfs_fs_info *fs_info) +static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans) { + struct btrfs_fs_info *fs_info = trans->fs_info; struct list_head *dirty_bgs = &trans->transaction->dirty_bgs; struct list_head *io_bgs = &trans->transaction->io_bgs; - struct list_head *next; struct extent_buffer *eb; int ret; + /* + * At this point no one can be using this transaction to modify any tree + * and no one can start another transaction to modify any tree either. + */ + ASSERT(trans->transaction->state == TRANS_STATE_COMMIT_DOING, + "trans->transaction->state=%d", trans->transaction->state); + eb = btrfs_lock_root_node(fs_info->tree_root); ret = btrfs_cow_block(trans, fs_info->tree_root, eb, NULL, - 0, &eb); + 0, &eb, BTRFS_NESTING_COW); btrfs_tree_unlock(eb); free_extent_buffer(eb); if (ret) return ret; - ret = btrfs_run_delayed_refs(trans, fs_info, (unsigned long)-1); - if (ret) - return ret; - - ret = btrfs_run_dev_stats(trans, fs_info); + ret = btrfs_run_dev_stats(trans); if (ret) return ret; - ret = btrfs_run_dev_replace(trans, fs_info); + ret = btrfs_run_dev_replace(trans); if (ret) return ret; - ret = btrfs_run_qgroups(trans, fs_info); + ret = btrfs_run_qgroups(trans); if (ret) return ret; - ret = btrfs_setup_space_cache(trans, fs_info); + ret = btrfs_setup_space_cache(trans); if (ret) return ret; - /* run_qgroups might have added some more refs */ - ret = btrfs_run_delayed_refs(trans, fs_info, (unsigned long)-1); - if (ret) - return ret; again: while (!list_empty(&fs_info->dirty_cowonly_roots)) { struct btrfs_root *root; - next = fs_info->dirty_cowonly_roots.next; - list_del_init(next); - root = list_entry(next, struct btrfs_root, dirty_list); + + root = list_first_entry(&fs_info->dirty_cowonly_roots, + struct btrfs_root, dirty_list); clear_bit(BTRFS_ROOT_DIRTY, &root->state); + list_move_tail(&root->dirty_list, + &trans->transaction->switch_commits); - if (root != fs_info->extent_root) - list_add_tail(&root->dirty_list, - &trans->transaction->switch_commits); ret = update_cowonly_root(trans, root); if (ret) return ret; - ret = btrfs_run_delayed_refs(trans, fs_info, (unsigned long)-1); - if (ret) - return ret; } + /* Now flush any delayed refs generated by updating all of the roots */ + ret = btrfs_run_delayed_refs(trans, U64_MAX); + if (ret) + return ret; + while (!list_empty(dirty_bgs) || !list_empty(io_bgs)) { - ret = btrfs_write_dirty_block_groups(trans, fs_info); + ret = btrfs_write_dirty_block_groups(trans); if (ret) return ret; - ret = btrfs_run_delayed_refs(trans, fs_info, (unsigned long)-1); + + /* + * We're writing the dirty block groups, which could generate + * delayed refs, which could generate more dirty block groups, + * so we want to keep this flushing in this loop to make sure + * everything gets run. + */ + ret = btrfs_run_delayed_refs(trans, U64_MAX); if (ret) return ret; } @@ -1226,14 +1407,40 @@ again: if (!list_empty(&fs_info->dirty_cowonly_roots)) goto again; - list_add_tail(&fs_info->extent_root->dirty_list, - &trans->transaction->switch_commits); - btrfs_after_dev_replace_commit(fs_info); + /* Update dev-replace pointer once everything is committed */ + fs_info->dev_replace.committed_cursor_left = + fs_info->dev_replace.cursor_left_last_write_of_item; return 0; } /* + * If we had a pending drop we need to see if there are any others left in our + * dead roots list, and if not clear our bit and wake any waiters. + */ +void btrfs_maybe_wake_unfinished_drop(struct btrfs_fs_info *fs_info) +{ + /* + * We put the drop in progress roots at the front of the list, so if the + * first entry doesn't have UNFINISHED_DROP set we can wake everybody + * up. + */ + spin_lock(&fs_info->trans_lock); + if (!list_empty(&fs_info->dead_roots)) { + struct btrfs_root *root = list_first_entry(&fs_info->dead_roots, + struct btrfs_root, + root_list); + if (test_bit(BTRFS_ROOT_UNFINISHED_DROP, &root->state)) { + spin_unlock(&fs_info->trans_lock); + return; + } + } + spin_unlock(&fs_info->trans_lock); + + btrfs_wake_unfinished_drop(fs_info); +} + +/* * dead roots are old snapshots that need to be deleted. This allocates * a dirty root struct and adds it into the list of dead roots that need to * be deleted @@ -1243,21 +1450,35 @@ void btrfs_add_dead_root(struct btrfs_root *root) struct btrfs_fs_info *fs_info = root->fs_info; spin_lock(&fs_info->trans_lock); - if (list_empty(&root->root_list)) - list_add_tail(&root->root_list, &fs_info->dead_roots); + if (list_empty(&root->root_list)) { + btrfs_grab_root(root); + + /* We want to process the partially complete drops first. */ + if (test_bit(BTRFS_ROOT_UNFINISHED_DROP, &root->state)) + list_add(&root->root_list, &fs_info->dead_roots); + else + list_add_tail(&root->root_list, &fs_info->dead_roots); + } spin_unlock(&fs_info->trans_lock); } /* - * update all the cowonly tree roots on disk + * Update each subvolume root and its relocation root, if it exists, in the tree + * of tree roots. Also free log roots if they exist. */ -static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, - struct btrfs_fs_info *fs_info) +static noinline int commit_fs_roots(struct btrfs_trans_handle *trans) { + struct btrfs_fs_info *fs_info = trans->fs_info; struct btrfs_root *gang[8]; int i; int ret; - int err = 0; + + /* + * At this point no one can be using this transaction to modify any tree + * and no one can start another transaction to modify any tree either. + */ + ASSERT(trans->transaction->state == TRANS_STATE_COMMIT_DOING, + "trans->transaction->state=%d", trans->transaction->state); spin_lock(&fs_info->fs_roots_radix_lock); while (1) { @@ -1269,16 +1490,32 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, break; for (i = 0; i < ret; i++) { struct btrfs_root *root = gang[i]; + int ret2; + + /* + * At this point we can neither have tasks logging inodes + * from a root nor trying to commit a log tree. + */ + ASSERT(atomic_read(&root->log_writers) == 0, + "atomic_read(&root->log_writers)=%d", + atomic_read(&root->log_writers)); + ASSERT(atomic_read(&root->log_commit[0]) == 0, + "atomic_read(&root->log_commit[0])=%d", + atomic_read(&root->log_commit[0])); + ASSERT(atomic_read(&root->log_commit[1]) == 0, + "atomic_read(&root->log_commit[1])=%d", + atomic_read(&root->log_commit[1])); + radix_tree_tag_clear(&fs_info->fs_roots_radix, - (unsigned long)root->root_key.objectid, + (unsigned long)btrfs_root_id(root), BTRFS_ROOT_TRANS_TAG); + btrfs_qgroup_free_meta_all_pertrans(root); spin_unlock(&fs_info->fs_roots_radix_lock); btrfs_free_log(trans, root); - btrfs_update_reloc_root(trans, root); - btrfs_orphan_commit_root(trans, root); - - btrfs_save_ino_cache(root, trans); + ret2 = btrfs_update_reloc_root(trans, root); + if (ret2) + return ret2; /* see comments in should_cow_block() */ clear_bit(BTRFS_ROOT_FORCE_COW, &root->state); @@ -1291,54 +1528,16 @@ static noinline int commit_fs_roots(struct btrfs_trans_handle *trans, root->node); } - err = btrfs_update_root(trans, fs_info->tree_root, + ret2 = btrfs_update_root(trans, fs_info->tree_root, &root->root_key, &root->root_item); + if (ret2) + return ret2; spin_lock(&fs_info->fs_roots_radix_lock); - if (err) - break; - btrfs_qgroup_free_meta_all(root); } } spin_unlock(&fs_info->fs_roots_radix_lock); - return err; -} - -/* - * defrag a given btree. - * Every leaf in the btree is read and defragged. - */ -int btrfs_defrag_root(struct btrfs_root *root) -{ - struct btrfs_fs_info *info = root->fs_info; - struct btrfs_trans_handle *trans; - int ret; - - if (test_and_set_bit(BTRFS_ROOT_DEFRAG_RUNNING, &root->state)) - return 0; - - while (1) { - trans = btrfs_start_transaction(root, 0); - if (IS_ERR(trans)) - return PTR_ERR(trans); - - ret = btrfs_defrag_leaves(trans, root); - - btrfs_end_transaction(trans); - btrfs_btree_balance_dirty(info); - cond_resched(); - - if (btrfs_fs_closing(info) || ret != -EAGAIN) - break; - - if (btrfs_defrag_cancelled(info)) { - btrfs_debug(info, "defrag_root cancelled"); - ret = -EAGAIN; - break; - } - } - clear_bit(BTRFS_ROOT_DEFRAG_RUNNING, &root->state); - return ret; + return 0; } /* @@ -1358,30 +1557,49 @@ static int qgroup_account_snapshot(struct btrfs_trans_handle *trans, int ret; /* - * Save some performance in the case that qgroups are not - * enabled. If this check races with the ioctl, rescan will - * kick in anyway. + * Save some performance in the case that qgroups are not enabled. If + * this check races with the ioctl, rescan will kick in anyway. */ - if (!test_bit(BTRFS_FS_QUOTA_ENABLED, &fs_info->flags)) + if (!btrfs_qgroup_full_accounting(fs_info)) return 0; /* - * We are going to commit transaction, see btrfs_commit_transaction() - * comment for reason locking tree_log_mutex + * Ensure dirty @src will be committed. Or, after coming + * commit_fs_roots() and switch_commit_roots(), any dirty but not + * recorded root will never be updated again, causing an outdated root + * item. */ - mutex_lock(&fs_info->tree_log_mutex); + ret = record_root_in_trans(trans, src, 1); + if (ret) + return ret; - ret = commit_fs_roots(trans, fs_info); + /* + * btrfs_qgroup_inherit relies on a consistent view of the usage for the + * src root, so we must run the delayed refs here. + * + * However this isn't particularly fool proof, because there's no + * synchronization keeping us from changing the tree after this point + * before we do the qgroup_inherit, or even from making changes while + * we're doing the qgroup_inherit. But that's a problem for the future, + * for now flush the delayed refs to narrow the race window where the + * qgroup counters could end up wrong. + */ + ret = btrfs_run_delayed_refs(trans, U64_MAX); + if (unlikely(ret)) { + btrfs_abort_transaction(trans, ret); + return ret; + } + + ret = commit_fs_roots(trans); if (ret) goto out; - ret = btrfs_qgroup_account_extents(trans, fs_info); + ret = btrfs_qgroup_account_extents(trans); if (ret < 0) goto out; /* Now qgroup are all updated, we can inherit it to new qgroups */ - ret = btrfs_qgroup_inherit(trans, fs_info, - src->root_key.objectid, dst_objectid, - inherit); + ret = btrfs_qgroup_inherit(trans, btrfs_root_id(src), dst_objectid, + btrfs_root_id(parent), inherit); if (ret < 0) goto out; @@ -1397,18 +1615,16 @@ static int qgroup_account_snapshot(struct btrfs_trans_handle *trans, * like chunk and root tree, as they won't affect qgroup. * And we don't write super to avoid half committed status. */ - ret = commit_cowonly_roots(trans, fs_info); + ret = commit_cowonly_roots(trans); if (ret) goto out; - switch_commit_roots(trans->transaction, fs_info); - ret = btrfs_write_and_wait_transaction(trans, fs_info); + switch_commit_roots(trans); + ret = btrfs_write_and_wait_transaction(trans); if (ret) btrfs_handle_fs_error(fs_info, ret, "Error while writing out transaction for qgroup"); out: - mutex_unlock(&fs_info->tree_log_mutex); - /* * Force parent root to be updated, as we recorded it before so its * last_trans == cur_transid. @@ -1416,7 +1632,7 @@ out: * insert_dir_item() */ if (!ret) - record_root_in_trans(trans, parent, 1); + ret = record_root_in_trans(trans, parent, 1); return ret; } @@ -1430,28 +1646,29 @@ out: * the creation of the pending snapshots, just return 0. */ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, - struct btrfs_fs_info *fs_info, struct btrfs_pending_snapshot *pending) { + + struct btrfs_fs_info *fs_info = trans->fs_info; struct btrfs_key key; struct btrfs_root_item *new_root_item; struct btrfs_root *tree_root = fs_info->tree_root; struct btrfs_root *root = pending->root; struct btrfs_root *parent_root; struct btrfs_block_rsv *rsv; - struct inode *parent_inode; - struct btrfs_path *path; + struct btrfs_inode *parent_inode = pending->dir; + BTRFS_PATH_AUTO_FREE(path); struct btrfs_dir_item *dir_item; - struct dentry *dentry; struct extent_buffer *tmp; struct extent_buffer *old; - struct timespec cur_time; + struct timespec64 cur_time; int ret = 0; u64 to_reserve = 0; u64 index = 0; u64 objectid; u64 root_flags; - uuid_le new_uuid; + unsigned int nofs_flags; + struct fscrypt_name fname; ASSERT(pending->path); path = pending->path; @@ -1459,9 +1676,22 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, ASSERT(pending->root_item); new_root_item = pending->root_item; - pending->error = btrfs_find_free_objectid(tree_root, &objectid); + /* + * We're inside a transaction and must make sure that any potential + * allocations with GFP_KERNEL in fscrypt won't recurse back to + * filesystem. + */ + nofs_flags = memalloc_nofs_save(); + pending->error = fscrypt_setup_filename(&parent_inode->vfs_inode, + &pending->dentry->d_name, 0, + &fname); + memalloc_nofs_restore(nofs_flags); if (pending->error) - goto no_free_objectid; + goto free_pending; + + pending->error = btrfs_get_free_objectid(tree_root, &objectid); + if (pending->error) + goto free_fname; /* * Make qgroup to skip current new snapshot's qgroupid, as it is @@ -1472,7 +1702,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, btrfs_reloc_pre_snapshot(pending, &to_reserve); if (to_reserve > 0) { - pending->error = btrfs_block_rsv_add(root, + pending->error = btrfs_block_rsv_add(fs_info, &pending->block_rsv, to_reserve, BTRFS_RESERVE_NO_FLUSH); @@ -1480,34 +1710,31 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, goto clear_skip_qgroup; } - key.objectid = objectid; - key.offset = (u64)-1; - key.type = BTRFS_ROOT_ITEM_KEY; - rsv = trans->block_rsv; trans->block_rsv = &pending->block_rsv; trans->bytes_reserved = trans->block_rsv->reserved; trace_btrfs_space_reservation(fs_info, "transaction", trans->transid, trans->bytes_reserved, 1); - dentry = pending->dentry; - parent_inode = pending->dir; - parent_root = BTRFS_I(parent_inode)->root; - record_root_in_trans(trans, parent_root, 0); - - cur_time = current_time(parent_inode); + parent_root = parent_inode->root; + ret = record_root_in_trans(trans, parent_root, 0); + if (ret) + goto fail; + cur_time = current_time(&parent_inode->vfs_inode); /* * insert the directory item */ - ret = btrfs_set_inode_index(BTRFS_I(parent_inode), &index); - BUG_ON(ret); /* -ENOMEM */ + ret = btrfs_set_inode_index(parent_inode, &index); + if (unlikely(ret)) { + btrfs_abort_transaction(trans, ret); + goto fail; + } /* check if there is a file/dir which has the same name. */ dir_item = btrfs_lookup_dir_item(NULL, parent_root, path, - btrfs_ino(BTRFS_I(parent_inode)), - dentry->d_name.name, - dentry->d_name.len, 0); + btrfs_ino(parent_inode), + &fname.disk_name, 0); if (dir_item != NULL && !IS_ERR(dir_item)) { pending->error = -EEXIST; goto dir_item_existed; @@ -1518,19 +1745,31 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, } btrfs_release_path(path); + ret = btrfs_create_qgroup(trans, objectid); + if (ret && ret != -EEXIST) { + if (unlikely(ret != -ENOTCONN || btrfs_qgroup_enabled(fs_info))) { + btrfs_abort_transaction(trans, ret); + goto fail; + } + } + /* * pull in the delayed directory update * and the delayed inode item * otherwise we corrupt the FS during * snapshot */ - ret = btrfs_run_delayed_items(trans, fs_info); - if (ret) { /* Transaction aborted */ + ret = btrfs_run_delayed_items(trans); + if (unlikely(ret)) { btrfs_abort_transaction(trans, ret); goto fail; } - record_root_in_trans(trans, root, 0); + ret = record_root_in_trans(trans, root, 0); + if (unlikely(ret)) { + btrfs_abort_transaction(trans, ret); + goto fail; + } btrfs_set_root_last_snapshot(&root->root_item, trans->transid); memcpy(new_root_item, &root->root_item, sizeof(*new_root_item)); btrfs_check_and_init_root_item(new_root_item); @@ -1544,8 +1783,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, btrfs_set_root_generation_v2(new_root_item, trans->transid); - uuid_le_gen(&new_uuid); - memcpy(new_root_item->uuid, new_uuid.b, BTRFS_UUID_SIZE); + generate_random_guid(new_root_item->uuid); memcpy(new_root_item->parent_uuid, root->root_item.uuid, BTRFS_UUID_SIZE); if (!(root_flags & BTRFS_ROOT_SUBVOL_RDONLY)) { @@ -1561,35 +1799,36 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, btrfs_set_root_otransid(new_root_item, trans->transid); old = btrfs_lock_root_node(root); - ret = btrfs_cow_block(trans, root, old, NULL, 0, &old); - if (ret) { + ret = btrfs_cow_block(trans, root, old, NULL, 0, &old, + BTRFS_NESTING_COW); + if (unlikely(ret)) { btrfs_tree_unlock(old); free_extent_buffer(old); btrfs_abort_transaction(trans, ret); goto fail; } - btrfs_set_lock_blocking(old); - ret = btrfs_copy_root(trans, root, old, &tmp, objectid); /* clean up in any case */ btrfs_tree_unlock(old); free_extent_buffer(old); - if (ret) { + if (unlikely(ret)) { btrfs_abort_transaction(trans, ret); goto fail; } /* see comments in should_cow_block() */ set_bit(BTRFS_ROOT_FORCE_COW, &root->state); - smp_wmb(); + smp_mb__after_atomic(); btrfs_set_root_node(new_root_item, tmp); /* record when the snapshot was created in key.offset */ + key.objectid = objectid; + key.type = BTRFS_ROOT_ITEM_KEY; key.offset = trans->transid; ret = btrfs_insert_root(trans, tree_root, &key, new_root_item); btrfs_tree_unlock(tmp); free_extent_buffer(tmp); - if (ret) { + if (unlikely(ret)) { btrfs_abort_transaction(trans, ret); goto fail; } @@ -1597,31 +1836,26 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, /* * insert root back/forward references */ - ret = btrfs_add_root_ref(trans, fs_info, objectid, - parent_root->root_key.objectid, - btrfs_ino(BTRFS_I(parent_inode)), index, - dentry->d_name.name, dentry->d_name.len); - if (ret) { + ret = btrfs_add_root_ref(trans, objectid, + btrfs_root_id(parent_root), + btrfs_ino(parent_inode), index, + &fname.disk_name); + if (unlikely(ret)) { btrfs_abort_transaction(trans, ret); goto fail; } key.offset = (u64)-1; - pending->snap = btrfs_read_fs_root_no_name(fs_info, &key); + pending->snap = btrfs_get_new_fs_root(fs_info, objectid, &pending->anon_dev); if (IS_ERR(pending->snap)) { ret = PTR_ERR(pending->snap); + pending->snap = NULL; btrfs_abort_transaction(trans, ret); goto fail; } ret = btrfs_reloc_post_snapshot(trans, pending); - if (ret) { - btrfs_abort_transaction(trans, ret); - goto fail; - } - - ret = btrfs_run_delayed_refs(trans, fs_info, (unsigned long)-1); - if (ret) { + if (unlikely(ret)) { btrfs_abort_transaction(trans, ret); goto fail; } @@ -1632,54 +1866,49 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans, * To co-operate with that hack, we do hack again. * Or snapshot will be greatly slowed down by a subtree qgroup rescan */ - ret = qgroup_account_snapshot(trans, root, parent_root, - pending->inherit, objectid); + if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_FULL) + ret = qgroup_account_snapshot(trans, root, parent_root, + pending->inherit, objectid); + else if (btrfs_qgroup_mode(fs_info) == BTRFS_QGROUP_MODE_SIMPLE) + ret = btrfs_qgroup_inherit(trans, btrfs_root_id(root), objectid, + btrfs_root_id(parent_root), pending->inherit); if (ret < 0) goto fail; - ret = btrfs_insert_dir_item(trans, parent_root, - dentry->d_name.name, dentry->d_name.len, - BTRFS_I(parent_inode), &key, - BTRFS_FT_DIR, index); - /* We have check then name at the beginning, so it is impossible. */ - BUG_ON(ret == -EEXIST || ret == -EOVERFLOW); - if (ret) { + ret = btrfs_insert_dir_item(trans, &fname.disk_name, + parent_inode, &key, BTRFS_FT_DIR, + index); + if (unlikely(ret)) { btrfs_abort_transaction(trans, ret); goto fail; } - btrfs_i_size_write(BTRFS_I(parent_inode), parent_inode->i_size + - dentry->d_name.len * 2); - parent_inode->i_mtime = parent_inode->i_ctime = - current_time(parent_inode); - ret = btrfs_update_inode_fallback(trans, parent_root, parent_inode); - if (ret) { + btrfs_i_size_write(parent_inode, parent_inode->vfs_inode.i_size + + fname.disk_name.len * 2); + inode_set_mtime_to_ts(&parent_inode->vfs_inode, + inode_set_ctime_current(&parent_inode->vfs_inode)); + ret = btrfs_update_inode_fallback(trans, parent_inode); + if (unlikely(ret)) { btrfs_abort_transaction(trans, ret); goto fail; } - ret = btrfs_uuid_tree_add(trans, fs_info, new_uuid.b, - BTRFS_UUID_KEY_SUBVOL, objectid); - if (ret) { + ret = btrfs_uuid_tree_add(trans, new_root_item->uuid, + BTRFS_UUID_KEY_SUBVOL, + objectid); + if (unlikely(ret)) { btrfs_abort_transaction(trans, ret); goto fail; } if (!btrfs_is_empty_uuid(new_root_item->received_uuid)) { - ret = btrfs_uuid_tree_add(trans, fs_info, - new_root_item->received_uuid, + ret = btrfs_uuid_tree_add(trans, new_root_item->received_uuid, BTRFS_UUID_KEY_RECEIVED_SUBVOL, objectid); - if (ret && ret != -EEXIST) { + if (unlikely(ret && ret != -EEXIST)) { btrfs_abort_transaction(trans, ret); goto fail; } } - ret = btrfs_run_delayed_refs(trans, fs_info, (unsigned long)-1); - if (ret) { - btrfs_abort_transaction(trans, ret); - goto fail; - } - fail: pending->error = ret; dir_item_existed: @@ -1687,10 +1916,11 @@ dir_item_existed: trans->bytes_reserved = 0; clear_skip_qgroup: btrfs_clear_skip_qgroup(trans); -no_free_objectid: +free_fname: + fscrypt_free_filename(&fname); +free_pending: kfree(new_root_item); pending->root_item = NULL; - btrfs_free_path(path); pending->path = NULL; return ret; @@ -1699,8 +1929,7 @@ no_free_objectid: /* * create all the snapshots we've scheduled for creation */ -static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans, - struct btrfs_fs_info *fs_info) +static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans) { struct btrfs_pending_snapshot *pending, *next; struct list_head *head = &trans->transaction->pending_snapshots; @@ -1708,7 +1937,7 @@ static noinline int create_pending_snapshots(struct btrfs_trans_handle *trans, list_for_each_entry_safe(pending, next, head, list) { list_del(&pending->list); - ret = create_pending_snapshot(trans, fs_info, pending); + ret = create_pending_snapshot(trans, pending); if (ret) break; } @@ -1733,23 +1962,12 @@ static void update_super_roots(struct btrfs_fs_info *fs_info) super->root_level = root_item->level; if (btrfs_test_opt(fs_info, SPACE_CACHE)) super->cache_generation = root_item->generation; + else if (test_bit(BTRFS_FS_CLEANUP_SPACE_CACHE_V1, &fs_info->flags)) + super->cache_generation = 0; if (test_bit(BTRFS_FS_UPDATE_UUID_TREE_GEN, &fs_info->flags)) super->uuid_tree_generation = root_item->generation; } -int btrfs_transaction_in_commit(struct btrfs_fs_info *info) -{ - struct btrfs_transaction *trans; - int ret = 0; - - spin_lock(&info->trans_lock); - trans = info->running_transaction; - if (trans) - ret = (trans->state >= TRANS_STATE_COMMIT_START); - spin_unlock(&info->trans_lock); - return ret; -} - int btrfs_transaction_blocked(struct btrfs_fs_info *info) { struct btrfs_transaction *trans; @@ -1763,74 +1981,14 @@ int btrfs_transaction_blocked(struct btrfs_fs_info *info) return ret; } -/* - * wait for the current transaction commit to start and block subsequent - * transaction joins - */ -static void wait_current_trans_commit_start(struct btrfs_fs_info *fs_info, - struct btrfs_transaction *trans) -{ - wait_event(fs_info->transaction_blocked_wait, - trans->state >= TRANS_STATE_COMMIT_START || trans->aborted); -} - -/* - * wait for the current transaction to start and then become unblocked. - * caller holds ref. - */ -static void wait_current_trans_commit_start_and_unblock( - struct btrfs_fs_info *fs_info, - struct btrfs_transaction *trans) -{ - wait_event(fs_info->transaction_wait, - trans->state >= TRANS_STATE_UNBLOCKED || trans->aborted); -} - -/* - * commit transactions asynchronously. once btrfs_commit_transaction_async - * returns, any subsequent transaction will not be allowed to join. - */ -struct btrfs_async_commit { - struct btrfs_trans_handle *newtrans; - struct work_struct work; -}; - -static void do_async_commit(struct work_struct *work) -{ - struct btrfs_async_commit *ac = - container_of(work, struct btrfs_async_commit, work); - - /* - * We've got freeze protection passed with the transaction. - * Tell lockdep about it. - */ - if (ac->newtrans->type & __TRANS_FREEZABLE) - __sb_writers_acquired(ac->newtrans->fs_info->sb, SB_FREEZE_FS); - - current->journal_info = ac->newtrans; - - btrfs_commit_transaction(ac->newtrans); - kfree(ac); -} - -int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, - int wait_for_unblock) +void btrfs_commit_transaction_async(struct btrfs_trans_handle *trans) { struct btrfs_fs_info *fs_info = trans->fs_info; - struct btrfs_async_commit *ac; struct btrfs_transaction *cur_trans; - ac = kmalloc(sizeof(*ac), GFP_NOFS); - if (!ac) - return -ENOMEM; - - INIT_WORK(&ac->work, do_async_commit); - ac->newtrans = btrfs_join_transaction(trans->root); - if (IS_ERR(ac->newtrans)) { - int err = PTR_ERR(ac->newtrans); - kfree(ac); - return err; - } + /* Kick the transaction kthread. */ + set_bit(BTRFS_FS_COMMIT_TRANS, &fs_info->flags); + wake_up_process(fs_info->transaction_kthread); /* take transaction reference */ cur_trans = trans->transaction; @@ -1839,36 +1997,41 @@ int btrfs_commit_transaction_async(struct btrfs_trans_handle *trans, btrfs_end_transaction(trans); /* - * Tell lockdep we've released the freeze rwsem, since the - * async commit thread will be the one to unlock it. + * Wait for the current transaction commit to start and block + * subsequent transaction joins */ - if (ac->newtrans->type & __TRANS_FREEZABLE) - __sb_writers_release(fs_info->sb, SB_FREEZE_FS); + btrfs_might_wait_for_state(fs_info, BTRFS_LOCKDEP_TRANS_COMMIT_PREP); + wait_event(fs_info->transaction_blocked_wait, + cur_trans->state >= TRANS_STATE_COMMIT_START || + TRANS_ABORTED(cur_trans)); + btrfs_put_transaction(cur_trans); +} - schedule_work(&ac->work); +/* + * If there is a running transaction commit it or if it's already committing, + * wait for its commit to complete. Does not start and commit a new transaction + * if there isn't any running. + */ +int btrfs_commit_current_transaction(struct btrfs_root *root) +{ + struct btrfs_trans_handle *trans; - /* wait for transaction to start and unblock */ - if (wait_for_unblock) - wait_current_trans_commit_start_and_unblock(fs_info, cur_trans); - else - wait_current_trans_commit_start(fs_info, cur_trans); + trans = btrfs_attach_transaction_barrier(root); + if (IS_ERR(trans)) { + int ret = PTR_ERR(trans); - if (current->journal_info == trans) - current->journal_info = NULL; + return (ret == -ENOENT) ? 0 : ret; + } - btrfs_put_transaction(cur_trans); - return 0; + return btrfs_commit_transaction(trans); } - -static void cleanup_transaction(struct btrfs_trans_handle *trans, - struct btrfs_root *root, int err) +static void cleanup_transaction(struct btrfs_trans_handle *trans, int err) { - struct btrfs_fs_info *fs_info = root->fs_info; + struct btrfs_fs_info *fs_info = trans->fs_info; struct btrfs_transaction *cur_trans = trans->transaction; - DEFINE_WAIT(wait); - WARN_ON(trans->use_count > 1); + WARN_ON(refcount_read(&trans->use_count) > 1); btrfs_abort_transaction(trans, err); @@ -1881,18 +2044,34 @@ static void cleanup_transaction(struct btrfs_trans_handle *trans, */ BUG_ON(list_empty(&cur_trans->list)); - list_del_init(&cur_trans->list); if (cur_trans == fs_info->running_transaction) { cur_trans->state = TRANS_STATE_COMMIT_DOING; spin_unlock(&fs_info->trans_lock); + + /* + * The thread has already released the lockdep map as reader + * already in btrfs_commit_transaction(). + */ + btrfs_might_wait_for_event(fs_info, btrfs_trans_num_writers); wait_event(cur_trans->writer_wait, atomic_read(&cur_trans->num_writers) == 1); spin_lock(&fs_info->trans_lock); } + + /* + * Now that we know no one else is still using the transaction we can + * remove the transaction from the list of transactions. This avoids + * the transaction kthread from cleaning up the transaction while some + * other task is still using it, which could result in a use-after-free + * on things like log trees, as it forces the transaction kthread to + * wait for this transaction to be cleaned up by us. + */ + list_del_init(&cur_trans->list); + spin_unlock(&fs_info->trans_lock); - btrfs_cleanup_one_transaction(trans->transaction, fs_info); + btrfs_cleanup_one_transaction(trans->transaction); spin_lock(&fs_info->trans_lock); if (cur_trans == fs_info->running_transaction) @@ -1904,33 +2083,115 @@ static void cleanup_transaction(struct btrfs_trans_handle *trans, btrfs_put_transaction(cur_trans); btrfs_put_transaction(cur_trans); - trace_btrfs_transaction_commit(root); + trace_btrfs_transaction_commit(fs_info); if (current->journal_info == trans) current->journal_info = NULL; - btrfs_scrub_cancel(fs_info); + + /* + * If relocation is running, we can't cancel scrub because that will + * result in a deadlock. Before relocating a block group, relocation + * pauses scrub, then starts and commits a transaction before unpausing + * scrub. If the transaction commit is being done by the relocation + * task or triggered by another task and the relocation task is waiting + * for the commit, and we end up here due to an error in the commit + * path, then calling btrfs_scrub_cancel() will deadlock, as we are + * asking for scrub to stop while having it asked to be paused higher + * above in relocation code. + */ + if (!test_bit(BTRFS_FS_RELOC_RUNNING, &fs_info->flags)) + btrfs_scrub_cancel(fs_info); kmem_cache_free(btrfs_trans_handle_cachep, trans); } +/* + * Release reserved delayed ref space of all pending block groups of the + * transaction and remove them from the list + */ +static void btrfs_cleanup_pending_block_groups(struct btrfs_trans_handle *trans) +{ + struct btrfs_fs_info *fs_info = trans->fs_info; + struct btrfs_block_group *block_group, *tmp; + + list_for_each_entry_safe(block_group, tmp, &trans->new_bgs, bg_list) { + btrfs_dec_delayed_refs_rsv_bg_inserts(fs_info); + /* + * Not strictly necessary to lock, as no other task will be using a + * block_group on the new_bgs list during a transaction abort. + */ + spin_lock(&fs_info->unused_bgs_lock); + list_del_init(&block_group->bg_list); + btrfs_put_block_group(block_group); + spin_unlock(&fs_info->unused_bgs_lock); + } +} + static inline int btrfs_start_delalloc_flush(struct btrfs_fs_info *fs_info) { + /* + * We use try_to_writeback_inodes_sb() here because if we used + * btrfs_start_delalloc_roots we would deadlock with fs freeze. + * Currently are holding the fs freeze lock, if we do an async flush + * we'll do btrfs_join_transaction() and deadlock because we need to + * wait for the fs freeze lock. Using the direct flushing we benefit + * from already being in a transaction and our join_transaction doesn't + * have to re-take the fs freeze lock. + * + * Note that try_to_writeback_inodes_sb() will only trigger writeback + * if it can read lock sb->s_umount. It will always be able to lock it, + * except when the filesystem is being unmounted or being frozen, but in + * those cases sync_filesystem() is called, which results in calling + * writeback_inodes_sb() while holding a write lock on sb->s_umount. + * Note that we don't call writeback_inodes_sb() directly, because it + * will emit a warning if sb->s_umount is not locked. + */ if (btrfs_test_opt(fs_info, FLUSHONCOMMIT)) - return btrfs_start_delalloc_roots(fs_info, 1, -1); + try_to_writeback_inodes_sb(fs_info->sb, WB_REASON_SYNC); return 0; } static inline void btrfs_wait_delalloc_flush(struct btrfs_fs_info *fs_info) { if (btrfs_test_opt(fs_info, FLUSHONCOMMIT)) - btrfs_wait_ordered_roots(fs_info, U64_MAX, 0, (u64)-1); + btrfs_wait_ordered_roots(fs_info, U64_MAX, NULL); } -static inline void -btrfs_wait_pending_ordered(struct btrfs_transaction *cur_trans) +/* + * Add a pending snapshot associated with the given transaction handle to the + * respective handle. This must be called after the transaction commit started + * and while holding fs_info->trans_lock. + * This serves to guarantee a caller of btrfs_commit_transaction() that it can + * safely free the pending snapshot pointer in case btrfs_commit_transaction() + * returns an error. + */ +static void add_pending_snapshot(struct btrfs_trans_handle *trans) { - wait_event(cur_trans->pending_wait, - atomic_read(&cur_trans->pending_ordered) == 0); + struct btrfs_transaction *cur_trans = trans->transaction; + + if (!trans->pending_snapshot) + return; + + lockdep_assert_held(&trans->fs_info->trans_lock); + ASSERT(cur_trans->state >= TRANS_STATE_COMMIT_PREP, + "cur_trans->state=%d", cur_trans->state); + + list_add(&trans->pending_snapshot->list, &cur_trans->pending_snapshots); +} + +static void update_commit_stats(struct btrfs_fs_info *fs_info) +{ + ktime_t now = ktime_get_ns(); + ktime_t interval = now - fs_info->commit_stats.critical_section_start_time; + + ASSERT(fs_info->commit_stats.critical_section_start_time); + + fs_info->commit_stats.commit_count++; + fs_info->commit_stats.last_commit_dur = interval; + fs_info->commit_stats.max_commit_dur = + max_t(u64, fs_info->commit_stats.max_commit_dur, interval); + fs_info->commit_stats.total_commit_dur += interval; + fs_info->commit_stats.critical_section_start_time = 0; } int btrfs_commit_transaction(struct btrfs_trans_handle *trans) @@ -1940,43 +2201,38 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) struct btrfs_transaction *prev_trans = NULL; int ret; + ASSERT(refcount_read(&trans->use_count) == 1, + "refcount_read(&trans->use_count)=%d", refcount_read(&trans->use_count)); + btrfs_trans_state_lockdep_acquire(fs_info, BTRFS_LOCKDEP_TRANS_COMMIT_PREP); + + clear_bit(BTRFS_FS_NEED_TRANS_COMMIT, &fs_info->flags); + /* Stop the commit early if ->aborted is set */ - if (unlikely(READ_ONCE(cur_trans->aborted))) { + if (TRANS_ABORTED(cur_trans)) { ret = cur_trans->aborted; - btrfs_end_transaction(trans); - return ret; + goto lockdep_trans_commit_start_release; } - /* make a pass through all the delayed refs we have so far - * any runnings procs may add more while we are here - */ - ret = btrfs_run_delayed_refs(trans, fs_info, 0); - if (ret) { - btrfs_end_transaction(trans); - return ret; - } - - btrfs_trans_release_metadata(trans, fs_info); + btrfs_trans_release_metadata(trans); trans->block_rsv = NULL; - cur_trans = trans->transaction; - /* - * set the flushing flag so procs in this transaction have to - * start sending their work down. + * We only want one transaction commit doing the flushing so we do not + * waste a bunch of time on lock contention on the extent root node. */ - cur_trans->delayed_refs.flushing = 1; - smp_wmb(); - - if (!list_empty(&trans->new_bgs)) - btrfs_create_pending_block_groups(trans, fs_info); - - ret = btrfs_run_delayed_refs(trans, fs_info, 0); - if (ret) { - btrfs_end_transaction(trans); - return ret; + if (!test_and_set_bit(BTRFS_DELAYED_REFS_FLUSHING, + &cur_trans->delayed_refs.flags)) { + /* + * Make a pass through all the delayed refs we have so far. + * Any running threads may add more while we are here. + */ + ret = btrfs_run_delayed_refs(trans, 0); + if (ret) + goto lockdep_trans_commit_start_release; } + btrfs_create_pending_block_groups(trans); + if (!test_bit(BTRFS_TRANS_DIRTY_BG_RUN, &cur_trans->flags)) { int run_it = 0; @@ -1999,23 +2255,31 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) run_it = 1; mutex_unlock(&fs_info->ro_block_group_mutex); - if (run_it) - ret = btrfs_start_dirty_block_groups(trans, fs_info); - } - if (ret) { - btrfs_end_transaction(trans); - return ret; + if (run_it) { + ret = btrfs_start_dirty_block_groups(trans); + if (ret) + goto lockdep_trans_commit_start_release; + } } spin_lock(&fs_info->trans_lock); - if (cur_trans->state >= TRANS_STATE_COMMIT_START) { + if (cur_trans->state >= TRANS_STATE_COMMIT_PREP) { + enum btrfs_trans_state want_state = TRANS_STATE_COMPLETED; + + add_pending_snapshot(trans); + spin_unlock(&fs_info->trans_lock); refcount_inc(&cur_trans->use_count); - ret = btrfs_end_transaction(trans); - wait_for_commit(cur_trans); + if (trans->in_fsync) + want_state = TRANS_STATE_SUPER_COMMITTED; - if (unlikely(cur_trans->aborted)) + btrfs_trans_state_lockdep_release(fs_info, + BTRFS_LOCKDEP_TRANS_COMMIT_PREP); + ret = btrfs_end_transaction(trans); + wait_for_commit(cur_trans, want_state); + + if (TRANS_ABORTED(cur_trans)) ret = cur_trans->aborted; btrfs_put_transaction(cur_trans); @@ -2023,50 +2287,90 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) return ret; } - cur_trans->state = TRANS_STATE_COMMIT_START; + cur_trans->state = TRANS_STATE_COMMIT_PREP; wake_up(&fs_info->transaction_blocked_wait); + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_COMMIT_PREP); + + if (!list_is_first(&cur_trans->list, &fs_info->trans_list)) { + enum btrfs_trans_state want_state = TRANS_STATE_COMPLETED; - if (cur_trans->list.prev != &fs_info->trans_list) { - prev_trans = list_entry(cur_trans->list.prev, - struct btrfs_transaction, list); - if (prev_trans->state != TRANS_STATE_COMPLETED) { + if (trans->in_fsync) + want_state = TRANS_STATE_SUPER_COMMITTED; + + prev_trans = list_prev_entry(cur_trans, list); + if (prev_trans->state < want_state) { refcount_inc(&prev_trans->use_count); spin_unlock(&fs_info->trans_lock); - wait_for_commit(prev_trans); - ret = prev_trans->aborted; + wait_for_commit(prev_trans, want_state); + + ret = READ_ONCE(prev_trans->aborted); btrfs_put_transaction(prev_trans); if (ret) - goto cleanup_transaction; - } else { - spin_unlock(&fs_info->trans_lock); + goto lockdep_release; + spin_lock(&fs_info->trans_lock); } } else { - spin_unlock(&fs_info->trans_lock); + /* + * The previous transaction was aborted and was already removed + * from the list of transactions at fs_info->trans_list. So we + * abort to prevent writing a new superblock that reflects a + * corrupt state (pointing to trees with unwritten nodes/leafs). + */ + if (BTRFS_FS_ERROR(fs_info)) { + spin_unlock(&fs_info->trans_lock); + ret = -EROFS; + goto lockdep_release; + } } + cur_trans->state = TRANS_STATE_COMMIT_START; + wake_up(&fs_info->transaction_blocked_wait); + spin_unlock(&fs_info->trans_lock); + + /* + * Get the time spent on the work done by the commit thread and not + * the time spent waiting on a previous commit + */ + fs_info->commit_stats.critical_section_start_time = ktime_get_ns(); extwriter_counter_dec(cur_trans, trans->type); ret = btrfs_start_delalloc_flush(fs_info); if (ret) - goto cleanup_transaction; + goto lockdep_release; - ret = btrfs_run_delayed_items(trans, fs_info); + ret = btrfs_run_delayed_items(trans); if (ret) - goto cleanup_transaction; + goto lockdep_release; + /* + * The thread has started/joined the transaction thus it holds the + * lockdep map as a reader. It has to release it before acquiring the + * lockdep map as a writer. + */ + btrfs_lockdep_release(fs_info, btrfs_trans_num_extwriters); + btrfs_might_wait_for_event(fs_info, btrfs_trans_num_extwriters); wait_event(cur_trans->writer_wait, extwriter_counter_read(cur_trans) == 0); /* some pending stuffs might be added after the previous flush. */ - ret = btrfs_run_delayed_items(trans, fs_info); - if (ret) + ret = btrfs_run_delayed_items(trans); + if (ret) { + btrfs_lockdep_release(fs_info, btrfs_trans_num_writers); goto cleanup_transaction; + } btrfs_wait_delalloc_flush(fs_info); - btrfs_wait_pending_ordered(cur_trans); + /* + * Wait for all ordered extents started by a fast fsync that joined this + * transaction. Otherwise if this transaction commits before the ordered + * extents complete we lose logged data after a power failure. + */ + btrfs_might_wait_for_event(fs_info, btrfs_trans_pending_ordered); + wait_event(cur_trans->pending_wait, + atomic_read(&cur_trans->pending_ordered) == 0); btrfs_scrub_pause(fs_info); /* @@ -2075,14 +2379,40 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) * COMMIT_DOING so make sure to wait for num_writers to == 1 again. */ spin_lock(&fs_info->trans_lock); + add_pending_snapshot(trans); cur_trans->state = TRANS_STATE_COMMIT_DOING; spin_unlock(&fs_info->trans_lock); + + /* + * The thread has started/joined the transaction thus it holds the + * lockdep map as a reader. It has to release it before acquiring the + * lockdep map as a writer. + */ + btrfs_lockdep_release(fs_info, btrfs_trans_num_writers); + btrfs_might_wait_for_event(fs_info, btrfs_trans_num_writers); wait_event(cur_trans->writer_wait, atomic_read(&cur_trans->num_writers) == 1); - /* ->aborted might be set after the previous check, so check it */ - if (unlikely(READ_ONCE(cur_trans->aborted))) { + /* + * Make lockdep happy by acquiring the state locks after + * btrfs_trans_num_writers is released. If we acquired the state locks + * before releasing the btrfs_trans_num_writers lock then lockdep would + * complain because we did not follow the reverse order unlocking rule. + */ + btrfs_trans_state_lockdep_acquire(fs_info, BTRFS_LOCKDEP_TRANS_COMPLETED); + btrfs_trans_state_lockdep_acquire(fs_info, BTRFS_LOCKDEP_TRANS_SUPER_COMMITTED); + btrfs_trans_state_lockdep_acquire(fs_info, BTRFS_LOCKDEP_TRANS_UNBLOCKED); + + /* + * We've started the commit, clear the flag in case we were triggered to + * do an async commit but somebody else started before the transaction + * kthread could do the work. + */ + clear_bit(BTRFS_FS_COMMIT_TRANS, &fs_info->flags); + + if (TRANS_ABORTED(cur_trans)) { ret = cur_trans->aborted; + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_UNBLOCKED); goto scrub_continue; } /* @@ -2097,11 +2427,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) * deal with them in create_pending_snapshot(), which is the * core function of the snapshot creation. */ - ret = create_pending_snapshots(trans, fs_info); - if (ret) { - mutex_unlock(&fs_info->reloc_mutex); - goto scrub_continue; - } + ret = create_pending_snapshots(trans); + if (ret) + goto unlock_reloc; /* * We insert the dir indexes of the snapshots and update the inode @@ -2110,20 +2438,16 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) * them. * * We needn't worry that this operation will corrupt the snapshots, - * because all the tree which are snapshoted will be forced to COW + * because all the tree which are snapshotted will be forced to COW * the nodes and leaves. */ - ret = btrfs_run_delayed_items(trans, fs_info); - if (ret) { - mutex_unlock(&fs_info->reloc_mutex); - goto scrub_continue; - } + ret = btrfs_run_delayed_items(trans); + if (ret) + goto unlock_reloc; - ret = btrfs_run_delayed_refs(trans, fs_info, (unsigned long)-1); - if (ret) { - mutex_unlock(&fs_info->reloc_mutex); - goto scrub_continue; - } + ret = btrfs_run_delayed_refs(trans, U64_MAX); + if (ret) + goto unlock_reloc; /* * make sure none of the code above managed to slip in a @@ -2133,33 +2457,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) WARN_ON(cur_trans != trans->transaction); - /* btrfs_commit_tree_roots is responsible for getting the - * various roots consistent with each other. Every pointer - * in the tree of tree roots has to point to the most up to date - * root for every subvolume and other tree. So, we have to keep - * the tree logging code from jumping in and changing any - * of the trees. - * - * At this point in the commit, there can't be any tree-log - * writers, but a little lower down we drop the trans mutex - * and let new people in. By holding the tree_log_mutex - * from now until after the super is written, we avoid races - * with the tree-log code. - */ - mutex_lock(&fs_info->tree_log_mutex); - - ret = commit_fs_roots(trans, fs_info); - if (ret) { - mutex_unlock(&fs_info->tree_log_mutex); - mutex_unlock(&fs_info->reloc_mutex); - goto scrub_continue; - } - - /* - * Since the transaction is done, we can apply the pending changes - * before the next transaction. - */ - btrfs_apply_pending_changes(fs_info); + ret = commit_fs_roots(trans); + if (ret) + goto unlock_reloc; /* commit_fs_roots gets rid of all the tree log roots, it is now * safe to free the root of tree log roots @@ -2167,47 +2467,26 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) btrfs_free_log_root_tree(trans, fs_info); /* - * commit_fs_roots() can call btrfs_save_ino_cache(), which generates - * new delayed refs. Must handle them or qgroup can be wrong. - */ - ret = btrfs_run_delayed_refs(trans, fs_info, (unsigned long)-1); - if (ret) { - mutex_unlock(&fs_info->tree_log_mutex); - mutex_unlock(&fs_info->reloc_mutex); - goto scrub_continue; - } - - /* * Since fs roots are all committed, we can get a quite accurate * new_roots. So let's do quota accounting. */ - ret = btrfs_qgroup_account_extents(trans, fs_info); - if (ret < 0) { - mutex_unlock(&fs_info->tree_log_mutex); - mutex_unlock(&fs_info->reloc_mutex); - goto scrub_continue; - } + ret = btrfs_qgroup_account_extents(trans); + if (ret < 0) + goto unlock_reloc; - ret = commit_cowonly_roots(trans, fs_info); - if (ret) { - mutex_unlock(&fs_info->tree_log_mutex); - mutex_unlock(&fs_info->reloc_mutex); - goto scrub_continue; - } + ret = commit_cowonly_roots(trans); + if (ret) + goto unlock_reloc; /* * The tasks which save the space cache and inode cache may also * update ->aborted, check it. */ - if (unlikely(READ_ONCE(cur_trans->aborted))) { + if (TRANS_ABORTED(cur_trans)) { ret = cur_trans->aborted; - mutex_unlock(&fs_info->tree_log_mutex); - mutex_unlock(&fs_info->reloc_mutex); - goto scrub_continue; + goto unlock_reloc; } - btrfs_prepare_extent_commit(fs_info); - cur_trans = fs_info->running_transaction; btrfs_set_root_node(&fs_info->tree_root->root_item, @@ -2220,7 +2499,14 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) list_add_tail(&fs_info->chunk_root->dirty_list, &cur_trans->switch_commits); - switch_commit_roots(cur_trans, fs_info); + if (btrfs_fs_incompat(fs_info, EXTENT_TREE_V2)) { + btrfs_set_root_node(&fs_info->block_group_root->root_item, + fs_info->block_group_root->node); + list_add_tail(&fs_info->block_group_root->dirty_list, + &cur_trans->switch_commits); + } + + switch_commit_roots(trans); ASSERT(list_empty(&cur_trans->dirty_bgs)); ASSERT(list_empty(&cur_trans->io_bgs)); @@ -2231,14 +2517,23 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) memcpy(fs_info->super_for_commit, fs_info->super_copy, sizeof(*fs_info->super_copy)); - btrfs_update_commit_device_size(fs_info); - btrfs_update_commit_device_bytes_used(fs_info, cur_trans); + btrfs_commit_device_sizes(cur_trans); clear_bit(BTRFS_FS_LOG1_ERR, &fs_info->flags); clear_bit(BTRFS_FS_LOG2_ERR, &fs_info->flags); btrfs_trans_release_chunk_metadata(trans); + /* + * Before changing the transaction state to TRANS_STATE_UNBLOCKED and + * setting fs_info->running_transaction to NULL, lock tree_log_mutex to + * make sure that before we commit our superblock, no other task can + * start a new transaction and commit a log tree before we commit our + * superblock. Anyone trying to commit a log tree locks this mutex before + * writing its superblock. + */ + mutex_lock(&fs_info->tree_log_mutex); + spin_lock(&fs_info->trans_lock); cur_trans->state = TRANS_STATE_UNBLOCKED; fs_info->running_transaction = NULL; @@ -2246,8 +2541,14 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) mutex_unlock(&fs_info->reloc_mutex); wake_up(&fs_info->transaction_wait); + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_UNBLOCKED); - ret = btrfs_write_and_wait_transaction(trans, fs_info); + /* If we have features changed, wake up the cleaner to update sysfs. */ + if (test_bit(BTRFS_FS_FEATURE_CHANGED, &fs_info->flags) && + fs_info->cleaner_kthread) + wake_up_process(fs_info->cleaner_kthread); + + ret = btrfs_write_and_wait_transaction(trans); if (ret) { btrfs_handle_fs_error(fs_info, ret, "Error while writing out transaction"); @@ -2256,29 +2557,38 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) } ret = write_all_supers(fs_info, 0); - if (ret) { - mutex_unlock(&fs_info->tree_log_mutex); - goto scrub_continue; - } - /* * the super is written, we can safely allow the tree-loggers * to go about their business */ mutex_unlock(&fs_info->tree_log_mutex); + if (ret) + goto scrub_continue; - btrfs_finish_extent_commit(trans, fs_info); + update_commit_stats(fs_info); + /* + * We needn't acquire the lock here because there is no other task + * which can change it. + */ + cur_trans->state = TRANS_STATE_SUPER_COMMITTED; + wake_up(&cur_trans->commit_wait); + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_SUPER_COMMITTED); + + ret = btrfs_finish_extent_commit(trans); + if (ret) + goto scrub_continue; if (test_bit(BTRFS_TRANS_HAVE_FREE_BGS, &cur_trans->flags)) btrfs_clear_space_info_full(fs_info); - fs_info->last_trans_committed = cur_trans->transid; + btrfs_set_last_trans_committed(fs_info, cur_trans->transid); /* * We needn't acquire the lock here because there is no other task * which can change it. */ cur_trans->state = TRANS_STATE_COMPLETED; wake_up(&cur_trans->commit_wait); + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_COMPLETED); spin_lock(&fs_info->trans_lock); list_del_init(&cur_trans->list); @@ -2290,7 +2600,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) if (trans->type & __TRANS_FREEZABLE) sb_end_intwrite(fs_info->sb); - trace_btrfs_transaction_commit(trans->root); + trace_btrfs_transaction_commit(fs_info); btrfs_scrub_continue(fs_info); @@ -2299,29 +2609,36 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans) kmem_cache_free(btrfs_trans_handle_cachep, trans); - /* - * If fs has been frozen, we can not handle delayed iputs, otherwise - * it'll result in deadlock about SB_FREEZE_FS. - */ - if (current != fs_info->transaction_kthread && - current != fs_info->cleaner_kthread && - !test_bit(BTRFS_FS_FROZEN, &fs_info->flags)) - btrfs_run_delayed_iputs(fs_info); - return ret; +unlock_reloc: + mutex_unlock(&fs_info->reloc_mutex); + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_UNBLOCKED); scrub_continue: + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_SUPER_COMMITTED); + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_COMPLETED); btrfs_scrub_continue(fs_info); cleanup_transaction: - btrfs_trans_release_metadata(trans, fs_info); + btrfs_trans_release_metadata(trans); + btrfs_cleanup_pending_block_groups(trans); btrfs_trans_release_chunk_metadata(trans); trans->block_rsv = NULL; btrfs_warn(fs_info, "Skipping commit of aborted transaction."); if (current->journal_info == trans) current->journal_info = NULL; - cleanup_transaction(trans, trans->root, ret); + cleanup_transaction(trans, ret); return ret; + +lockdep_release: + btrfs_lockdep_release(fs_info, btrfs_trans_num_extwriters); + btrfs_lockdep_release(fs_info, btrfs_trans_num_writers); + goto cleanup_transaction; + +lockdep_trans_commit_start_release: + btrfs_trans_state_lockdep_release(fs_info, BTRFS_LOCKDEP_TRANS_COMMIT_PREP); + btrfs_end_transaction(trans); + return ret; } /* @@ -2334,10 +2651,10 @@ cleanup_transaction: * because btrfs_commit_super will poke cleaner thread and it will process it a * few seconds later. */ -int btrfs_clean_one_deleted_snapshot(struct btrfs_root *root) +int btrfs_clean_one_deleted_snapshot(struct btrfs_fs_info *fs_info) { + struct btrfs_root *root; int ret; - struct btrfs_fs_info *fs_info = root->fs_info; spin_lock(&fs_info->trans_lock); if (list_empty(&fs_info->dead_roots)) { @@ -2349,44 +2666,58 @@ int btrfs_clean_one_deleted_snapshot(struct btrfs_root *root) list_del_init(&root->root_list); spin_unlock(&fs_info->trans_lock); - btrfs_debug(fs_info, "cleaner removing %llu", root->objectid); + btrfs_debug(fs_info, "cleaner removing %llu", btrfs_root_id(root)); btrfs_kill_all_delayed_nodes(root); if (btrfs_header_backref_rev(root->node) < BTRFS_MIXED_BACKREF_REV) - ret = btrfs_drop_snapshot(root, NULL, 0, 0); + ret = btrfs_drop_snapshot(root, false, false); else - ret = btrfs_drop_snapshot(root, NULL, 1, 0); + ret = btrfs_drop_snapshot(root, true, false); + btrfs_put_root(root); return (ret < 0) ? 0 : 1; } -void btrfs_apply_pending_changes(struct btrfs_fs_info *fs_info) +/* + * We only mark the transaction aborted and then set the file system read-only. + * This will prevent new transactions from starting or trying to join this + * one. + * + * This means that error recovery at the call site is limited to freeing + * any local memory allocations and passing the error code up without + * further cleanup. The transaction should complete as it normally would + * in the call path but will return -EIO. + * + * We'll complete the cleanup in btrfs_end_transaction and + * btrfs_commit_transaction. + */ +void __cold __btrfs_abort_transaction(struct btrfs_trans_handle *trans, + const char *function, + unsigned int line, int error, bool first_hit) { - unsigned long prev; - unsigned long bit; - - prev = xchg(&fs_info->pending_changes, 0); - if (!prev) - return; - - bit = 1 << BTRFS_PENDING_SET_INODE_MAP_CACHE; - if (prev & bit) - btrfs_set_opt(fs_info->mount_opt, INODE_MAP_CACHE); - prev &= ~bit; + struct btrfs_fs_info *fs_info = trans->fs_info; - bit = 1 << BTRFS_PENDING_CLEAR_INODE_MAP_CACHE; - if (prev & bit) - btrfs_clear_opt(fs_info->mount_opt, INODE_MAP_CACHE); - prev &= ~bit; + WRITE_ONCE(trans->aborted, error); + WRITE_ONCE(trans->transaction->aborted, error); + if (first_hit && error == -ENOSPC) + btrfs_dump_space_info_for_trans_abort(fs_info); + /* Wake up anybody who may be waiting on this transaction */ + wake_up(&fs_info->transaction_wait); + wake_up(&fs_info->transaction_blocked_wait); + __btrfs_handle_fs_error(fs_info, function, line, error, NULL); +} - bit = 1 << BTRFS_PENDING_COMMIT; - if (prev & bit) - btrfs_debug(fs_info, "pending commit done"); - prev &= ~bit; +int __init btrfs_transaction_init(void) +{ + btrfs_trans_handle_cachep = KMEM_CACHE(btrfs_trans_handle, SLAB_TEMPORARY); + if (!btrfs_trans_handle_cachep) + return -ENOMEM; + return 0; +} - if (prev) - btrfs_warn(fs_info, - "unknown pending changes left 0x%lx, ignoring", prev); +void __cold btrfs_transaction_exit(void) +{ + kmem_cache_destroy(btrfs_trans_handle_cachep); } |