search for: commit_root

This fixes bugzilla 57491. If we take a snapshot of a fs with a unlink ongoing and then try to send that root we will run into problems. When comparing with a parent root we will search the parents and the send roots commit_root, which if we''ve just created the snapshot will include the file that needs to be evicted by the orphan cleanup. So when we find a changed extent we will try and copy that info into the send stream, but when we lookup the inode we use the normal root, which no longer has the inode because...

...btrfs_lookup_leaf_ref(root, leaf->start); BUG_ON(!ref); @@ -7627,7 +7637,8 @@ static noinline int init_reloc_tree(stru return 0; root_item = kmalloc(sizeof(*root_item), GFP_NOFS); - BUG_ON(!root_item); + if (!root_item) + return -ENOMEM; ret = btrfs_copy_root(trans, root, root->commit_root, &eb, BTRFS_TREE_RELOC_OBJECTID); @@ -7653,7 +7664,7 @@ static noinline int init_reloc_tree(stru reloc_root = btrfs_read_fs_root_no_radix(root->fs_info->tree_root, &root_key); - BUG_ON(!reloc_root); + BUG_ON(IS_ERR(reloc_root)); reloc_root->last_trans = tran...

...btrfs/ctree.h index 8f4b81d..f290b98 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1088,9 +1088,6 @@ struct btrfs_fs_info { struct btrfs_root { struct extent_buffer *node; - /* the node lock is held while changing the node pointer */ - spinlock_t node_lock; - struct extent_buffer *commit_root; struct btrfs_root *log_root; struct btrfs_root *reloc_root; diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 228cf36..64b2896 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1064,7 +1064,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, INIT_LIST...

...bt &root->root_key, &root->root_item); BUG_ON(ret); - btrfs_write_dirty_block_groups(trans, root); - ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); + ret = btrfs_write_dirty_block_groups(trans, root); BUG_ON(ret); } free_extent_buffer(root->commit_root); @@ -495,9 +491,6 @@ static noinline int commit_cowonly_roots root = list_entry(next, struct btrfs_root, dirty_list); update_cowonly_root(trans, root); - - ret = btrfs_run_delayed_refs(trans, root, (unsigned long)-1); - BUG_ON(ret); } return 0; } -- To unsubscribe from this list: se...

...ast_snapshot(&root->root_item) || + btrfs_header_flag(buf, BTRFS_HEADER_FLAG_RELOC))) + __btrfs_update_ref(trans, root, buf, cow); + else clean_tree_block(trans, root, buf); - } if (buf == root->node) { root->node = cow; extent_buffer_get(cow); - if (buf != root->commit_root) { - btrfs_free_extent(trans, root, buf->start, - buf->len, buf->start, - root->root_key.objectid, - btrfs_header_generation(buf), - level, 1); - } + + btrfs_free_extent(trans, root, buf->start, buf->len, + 0, root->root_key.objectid, level, 0);...

This is a first draft of a subvolume quota implementation. It is possible to limit subvolumes and any group of subvolumes and also to track the amount of space that will get freed when deleting snapshots. The current version is functionally incomplete, with the main missing feature being the initial scan and rescan of an existing filesystem. I put some effort into writing an introduction into

...ock_level = -1; - - if (cow && (p->keep_locks || p->lowest_level)) - write_lock_level = BTRFS_MAX_LEVEL; - - min_write_lock_level = write_lock_level; - -again: - /* - * we try very hard to do read locks on the root - */ - root_lock = BTRFS_READ_LOCK; - level = 0; - if (p->search_commit_root) { - /* - * the commit roots are read only - * so we always do read locks - */ - b = root->commit_root; - extent_buffer_get(b); - level = btrfs_header_level(b); - if (!p->skip_locking) - btrfs_tree_read_lock(b); - } else { - if (p->skip_locking) { - b = btrfs_root_node(root...

...c Tue Jun 24 21:04:39 2008 -0400 @@ -701,8 +701,10 @@ static int __setup_root(u32 nodesize, u3 struct btrfs_fs_info *fs_info, u64 objectid) { + printk(KERN_ERR "setting up root %p\n", root); root->node = NULL; root->inode = NULL; + root->orphan_dir = NULL; root->commit_root = NULL; root->sectorsize = sectorsize; root->nodesize = nodesize; @@ -720,6 +722,7 @@ static int __setup_root(u32 nodesize, u3 root->in_sysfs = 0; INIT_LIST_HEAD(&root->dirty_list); + INIT_LIST_HEAD(&root->orphan_list); memset(&root->root_key, 0, sizeof(roo...

...gt;delayed_ref_updates; @@ -4895,6 +4896,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root, int update_ref) update); } } + btrfs_release_path(root, path); BUG_ON(err); @@ -6027,7 +6029,7 @@ int btrfs_drop_dead_reloc_roots(struct btrfs_root *root) BUG_ON(reloc_root->commit_root != NULL); while (1) { trans = btrfs_join_transaction(root, 1); - BUG_ON(!trans); + BUG_ON(IS_ERR(trans)); mutex_lock(&root->fs_info->drop_mutex); ret = btrfs_drop_snapshot(trans, reloc_root); @@ -6329,7 +6331,7 @@ static noinline int relocate_one_extent(struct btrfs_r...

...ans); ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); + btrfs_fixable_bug_on(ret); } location.objectid = BTRFS_DATA_RELOC_TREE_OBJECTID; @@ -7241,7 +7241,7 @@ static noinline int init_reloc_tree(struct btrfs_trans_handle *trans, ret = btrfs_copy_root(trans, root, root->commit_root, &eb, BTRFS_TREE_RELOC_OBJECTID); - BUG_ON(ret); + btrfs_fixable_bug_on(ret); root_key.objectid = BTRFS_TREE_RELOC_OBJECTID; root_key.offset = root->root_key.objectid; @@ -7258,7 +7258,7 @@ static noinline int init_reloc_tree(struct btrfs_trans_handle *trans, ret = btrfs_...

...stripesize, struct btrfs_root *root, - struct btrfs_fs_info *fs_info, - u64 objectid) +static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, + u32 stripesize, struct btrfs_root *root, + struct btrfs_fs_info *fs_info, + u64 objectid) { root->node = NULL; root->commit_root = NULL; @@ -1116,8 +1116,6 @@ static int __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, INIT_LIST_HEAD(&root->anon_super.s_list); INIT_LIST_HEAD(&root->anon_super.s_instances); init_rwsem(&root->anon_super.s_umount); - - return 0; } static int find_and_setup...

...adding orphans to an inode''s root, we start a transaction for that root that when ended in several places such as for example extent-tree.c:btrfs_remove_block_group(), inode.c:btrfs_unlink() and inode.c:btrfs_evict_node(), doesn''t result in a commit, that is, inode.c:btrfs_orphan_commit_root() doesn''t get called (via transaction.c:commit_fs_roots()). The respective inode will end up being called by inode.c:btrfs_evict_node() (via the VFS). So my likely dirty hack, after some debugging, of freeing the orphan block rsv in btrfs_evict_inode() if its not being used by other tasks...

I''ve been working to try to improve the write-ahead log''s performance, and I found that the bottleneck addresses in the checksum items, especially when we want to make a random write on a large file, e.g a 4G file. Then a idea for this suggested by Chris is to use sub transaction ids and just to log the part of inode that had changed since either the last log commit or the last

Hi all - Here''s my current error handling patchset, against 3.1-rc8. Almost all of this patchset is preparing for actual error handling. Before we start in on that work, I''m trying to reduce the surface we need to worry about. It turns out that there is a ton of code that returns an error code but never actually reports an error. The patchset has grown to 65 patches. 46 of them