From 7eaceaccab5f40bbfda044629a6298616aeaed50 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Thu, 10 Mar 2011 08:52:07 +0100 Subject: block: remove per-queue plugging Code has been converted over to the new explicit on-stack plugging, and delay users have been converted to use the new API for that. So lets kill off the old plugging along with aops->sync_page(). Signed-off-by: Jens Axboe --- fs/btrfs/inode.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index fb9bd7832b6d..462e08e724b0 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -7218,7 +7218,6 @@ static const struct address_space_operations btrfs_aops = { .writepage = btrfs_writepage, .writepages = btrfs_writepages, .readpages = btrfs_readpages, - .sync_page = block_sync_page, .direct_IO = btrfs_direct_IO, .invalidatepage = btrfs_invalidatepage, .releasepage = btrfs_releasepage, -- cgit v1.2.2 From 57a45ced94fe48a701361d64230fc16eefa189dd Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 25 Jan 2011 16:30:38 -0500 Subject: Btrfs: change reserved_extents to an atomic_t We track delayed allocation per inodes via 2 counters, one is outstanding_extents and reserved_extents. Outstanding_extents is already an atomic_t, but reserved_extents is not and is protected by a spinlock. So convert this to an atomic_t and instead of using a spinlock, use atomic_cmpxchg when releasing delalloc bytes. This makes our inode 72 bytes smaller, and reduces locking overhead (albiet it was minimal to begin with). Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 9007bbd01dbf..d97b69afbbfb 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6632,9 +6632,8 @@ struct inode *btrfs_alloc_inode(struct super_block *sb) ei->index_cnt = (u64)-1; ei->last_unlink_trans = 0; - spin_lock_init(&ei->accounting_lock); atomic_set(&ei->outstanding_extents, 0); - ei->reserved_extents = 0; + atomic_set(&ei->reserved_extents, 0); ei->ordered_data_close = 0; ei->orphan_meta_reserved = 0; @@ -6670,7 +6669,7 @@ void btrfs_destroy_inode(struct inode *inode) WARN_ON(!list_empty(&inode->i_dentry)); WARN_ON(inode->i_data.nrpages); WARN_ON(atomic_read(&BTRFS_I(inode)->outstanding_extents)); - WARN_ON(BTRFS_I(inode)->reserved_extents); + WARN_ON(atomic_read(&BTRFS_I(inode)->reserved_extents)); /* * This can happen where we create an inode, but somebody else also -- cgit v1.2.2 From dc89e9824464e91fa0b06267864ceabe3186fd8b Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 28 Jan 2011 17:05:48 -0500 Subject: Btrfs: use a slab for the free space entries Since we alloc/free free space entries a whole lot, lets use a slab to keep track of them. This makes some of my tests slightly faster. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d97b69afbbfb..2d2e079713d7 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -50,6 +50,7 @@ #include "tree-log.h" #include "compression.h" #include "locking.h" +#include "free-space-cache.h" struct btrfs_iget_args { u64 ino; @@ -70,6 +71,7 @@ static struct kmem_cache *btrfs_inode_cachep; struct kmem_cache *btrfs_trans_handle_cachep; struct kmem_cache *btrfs_transaction_cachep; struct kmem_cache *btrfs_path_cachep; +struct kmem_cache *btrfs_free_space_cachep; #define S_SHIFT 12 static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { @@ -6761,6 +6763,8 @@ void btrfs_destroy_cachep(void) kmem_cache_destroy(btrfs_transaction_cachep); if (btrfs_path_cachep) kmem_cache_destroy(btrfs_path_cachep); + if (btrfs_free_space_cachep) + kmem_cache_destroy(btrfs_free_space_cachep); } int btrfs_init_cachep(void) @@ -6789,6 +6793,12 @@ int btrfs_init_cachep(void) if (!btrfs_path_cachep) goto fail; + btrfs_free_space_cachep = kmem_cache_create("btrfs_free_space_cache", + sizeof(struct btrfs_free_space), 0, + SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL); + if (!btrfs_free_space_cachep) + goto fail; + return 0; fail: btrfs_destroy_cachep(); -- cgit v1.2.2 From a41ad394a03b802497958d7c98a9dcf607266645 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 31 Jan 2011 15:30:16 -0500 Subject: Btrfs: convert to the new truncate sequence ->truncate() is going away, instead all of the work needs to be done in ->setattr(). So this converts us over to do this. It's fairly straightforward, just get rid of our .truncate inode operation and call btrfs_truncate() directly from btrfs_setsize. This works out better for us since truncate can technically return ENOSPC, and before we had no way of letting anybody know. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 80 +++++++++++++++++++++++++------------------------------- 1 file changed, 36 insertions(+), 44 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 2d2e079713d7..3662ffec17d9 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -84,7 +84,8 @@ static unsigned char btrfs_type_by_mode[S_IFMT >> S_SHIFT] = { [S_IFLNK >> S_SHIFT] = BTRFS_FT_SYMLINK, }; -static void btrfs_truncate(struct inode *inode); +static int btrfs_setsize(struct inode *inode, loff_t newsize); +static int btrfs_truncate(struct inode *inode); static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end); static noinline int cow_file_range(struct inode *inode, struct page *locked_page, @@ -2371,6 +2372,11 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) /* if we have links, this was a truncate, lets do that */ if (inode->i_nlink) { + if (!S_ISREG(inode->i_mode)) { + WARN_ON(1); + iput(inode); + continue; + } nr_truncate++; btrfs_truncate(inode); } else { @@ -3538,7 +3544,7 @@ out: return ret; } -int btrfs_cont_expand(struct inode *inode, loff_t size) +int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) { struct btrfs_trans_handle *trans; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -3546,7 +3552,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size) struct extent_map *em = NULL; struct extent_state *cached_state = NULL; u64 mask = root->sectorsize - 1; - u64 hole_start = (inode->i_size + mask) & ~mask; + u64 hole_start = (oldsize + mask) & ~mask; u64 block_end = (size + mask) & ~mask; u64 last_byte; u64 cur_offset; @@ -3617,27 +3623,17 @@ int btrfs_cont_expand(struct inode *inode, loff_t size) return err; } -static int btrfs_setattr_size(struct inode *inode, struct iattr *attr) +static int btrfs_setsize(struct inode *inode, loff_t newsize) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; + loff_t oldsize = i_size_read(inode); unsigned long nr; int ret; - if (attr->ia_size == inode->i_size) + if (newsize == oldsize) return 0; - if (attr->ia_size > inode->i_size) { - unsigned long limit; - limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur; - if (attr->ia_size > inode->i_sb->s_maxbytes) - return -EFBIG; - if (limit != RLIM_INFINITY && attr->ia_size > limit) { - send_sig(SIGXFSZ, current, 0); - return -EFBIG; - } - } - trans = btrfs_start_transaction(root, 5); if (IS_ERR(trans)) return PTR_ERR(trans); @@ -3651,16 +3647,16 @@ static int btrfs_setattr_size(struct inode *inode, struct iattr *attr) btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); - if (attr->ia_size > inode->i_size) { - ret = btrfs_cont_expand(inode, attr->ia_size); + if (newsize > oldsize) { + i_size_write(inode, newsize); + btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); + truncate_pagecache(inode, oldsize, newsize); + ret = btrfs_cont_expand(inode, oldsize, newsize); if (ret) { - btrfs_truncate(inode); + btrfs_setsize(inode, oldsize); return ret; } - i_size_write(inode, attr->ia_size); - btrfs_ordered_update_i_size(inode, inode->i_size, NULL); - trans = btrfs_start_transaction(root, 0); BUG_ON(IS_ERR(trans)); btrfs_set_trans_block_group(trans, inode); @@ -3676,22 +3672,22 @@ static int btrfs_setattr_size(struct inode *inode, struct iattr *attr) nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); - return 0; - } + } else { - /* - * We're truncating a file that used to have good data down to - * zero. Make sure it gets into the ordered flush list so that - * any new writes get down to disk quickly. - */ - if (attr->ia_size == 0) - BTRFS_I(inode)->ordered_data_close = 1; + /* + * We're truncating a file that used to have good data down to + * zero. Make sure it gets into the ordered flush list so that + * any new writes get down to disk quickly. + */ + if (newsize == 0) + BTRFS_I(inode)->ordered_data_close = 1; - /* we don't support swapfiles, so vmtruncate shouldn't fail */ - ret = vmtruncate(inode, attr->ia_size); - BUG_ON(ret); + /* we don't support swapfiles, so vmtruncate shouldn't fail */ + truncate_setsize(inode, newsize); + ret = btrfs_truncate(inode); + } - return 0; + return ret; } static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) @@ -3708,7 +3704,7 @@ static int btrfs_setattr(struct dentry *dentry, struct iattr *attr) return err; if (S_ISREG(inode->i_mode) && (attr->ia_valid & ATTR_SIZE)) { - err = btrfs_setattr_size(inode, attr); + err = btrfs_setsize(inode, attr->ia_size); if (err) return err; } @@ -6478,7 +6474,7 @@ out: return ret; } -static void btrfs_truncate(struct inode *inode) +static int btrfs_truncate(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; int ret; @@ -6486,14 +6482,9 @@ static void btrfs_truncate(struct inode *inode) unsigned long nr; u64 mask = root->sectorsize - 1; - if (!S_ISREG(inode->i_mode)) { - WARN_ON(1); - return; - } - ret = btrfs_truncate_page(inode->i_mapping, inode->i_size); if (ret) - return; + return ret; btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1); btrfs_ordered_update_i_size(inode, inode->i_size, NULL); @@ -6568,6 +6559,8 @@ static void btrfs_truncate(struct inode *inode) ret = btrfs_end_transaction_throttle(trans, root); BUG_ON(ret); btrfs_btree_balance_dirty(root, nr); + + return ret; } /* @@ -7367,7 +7360,6 @@ static const struct address_space_operations btrfs_symlink_aops = { }; static const struct inode_operations btrfs_file_inode_operations = { - .truncate = btrfs_truncate, .getattr = btrfs_getattr, .setattr = btrfs_setattr, .setxattr = btrfs_setxattr, -- cgit v1.2.2 From 3893e33b0bebee2f67d96b6c15259dc884523c20 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 31 Jan 2011 16:03:11 -0500 Subject: Btrfs: cleanup error handling in the truncate path Now that we can handle having errors in the truncate path lets make sure we return errors instead of doing BUG_ON() and such. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 64 +++++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 45 insertions(+), 19 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3662ffec17d9..d83025063ee7 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3597,13 +3597,15 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) err = btrfs_drop_extents(trans, inode, cur_offset, cur_offset + hole_size, &hint_byte, 1); - BUG_ON(err); + if (err) + break; err = btrfs_insert_file_extent(trans, root, inode->i_ino, cur_offset, 0, 0, hole_size, 0, hole_size, 0, 0, 0); - BUG_ON(err); + if (err) + break; btrfs_drop_extent_cache(inode, hole_start, last_byte - 1, 0); @@ -3641,7 +3643,10 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) btrfs_set_trans_block_group(trans, inode); ret = btrfs_orphan_add(trans, inode); - BUG_ON(ret); + if (ret) { + btrfs_end_transaction(trans, root); + return ret; + } nr = trans->blocks_used; btrfs_end_transaction(trans, root); @@ -3658,17 +3663,24 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) } trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); + btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; BUG_ON(!trans->block_rsv); + /* + * If this fails just leave the orphan item so that it can get + * cleaned up next time we mount. + */ ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); - if (inode->i_nlink > 0) { - ret = btrfs_orphan_del(trans, inode); - BUG_ON(ret); + if (ret) { + btrfs_end_transaction(trans, root); + return ret; } + if (inode->i_nlink > 0) + ret = btrfs_orphan_del(trans, inode); nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); @@ -6478,6 +6490,7 @@ static int btrfs_truncate(struct inode *inode) { struct btrfs_root *root = BTRFS_I(inode)->root; int ret; + int err = 0; struct btrfs_trans_handle *trans; unsigned long nr; u64 mask = root->sectorsize - 1; @@ -6490,7 +6503,8 @@ static int btrfs_truncate(struct inode *inode) btrfs_ordered_update_i_size(inode, inode->i_size, NULL); trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; @@ -6517,29 +6531,38 @@ static int btrfs_truncate(struct inode *inode) while (1) { if (!trans) { trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) + return PTR_ERR(trans); btrfs_set_trans_block_group(trans, inode); trans->block_rsv = root->orphan_block_rsv; } ret = btrfs_block_rsv_check(trans, root, root->orphan_block_rsv, 0, 5); - if (ret) { - BUG_ON(ret != -EAGAIN); + if (ret == -EAGAIN) { ret = btrfs_commit_transaction(trans, root); - BUG_ON(ret); + if (ret) + return ret; trans = NULL; continue; + } else if (ret) { + err = ret; + break; } ret = btrfs_truncate_inode_items(trans, root, inode, inode->i_size, BTRFS_EXTENT_DATA_KEY); - if (ret != -EAGAIN) + if (ret != -EAGAIN) { + err = ret; break; + } ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + if (ret) { + err = ret; + break; + } nr = trans->blocks_used; btrfs_end_transaction(trans, root); @@ -6549,18 +6572,21 @@ static int btrfs_truncate(struct inode *inode) if (ret == 0 && inode->i_nlink > 0) { ret = btrfs_orphan_del(trans, inode); - BUG_ON(ret); + if (ret) + err = ret; } ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + if (ret && !err) + err = ret; nr = trans->blocks_used; ret = btrfs_end_transaction_throttle(trans, root); - BUG_ON(ret); + if (ret && !err) + err = ret; btrfs_btree_balance_dirty(root, nr); - return ret; + return err; } /* -- cgit v1.2.2 From 66b4ffd110f9b48b8d8c1319ee446b53b8d073bf Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 31 Jan 2011 16:22:42 -0500 Subject: Btrfs: handle errors in btrfs_orphan_cleanup If we cannot truncate an inode for some reason we will never delete the orphan item associated with that inode, which means that we will loop forever in btrfs_orphan_cleanup. Instead of doing this just return error so we fail to mount. It sucks, but hey it's better than hanging. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 47 +++++++++++++++++++++++++++++++---------------- 1 file changed, 31 insertions(+), 16 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index d83025063ee7..0600265cb9b0 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2284,7 +2284,7 @@ int btrfs_orphan_del(struct btrfs_trans_handle *trans, struct inode *inode) * this cleans up any orphans that may be left on the list from the last use * of this root. */ -void btrfs_orphan_cleanup(struct btrfs_root *root) +int btrfs_orphan_cleanup(struct btrfs_root *root) { struct btrfs_path *path; struct extent_buffer *leaf; @@ -2294,10 +2294,13 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) int ret = 0, nr_unlink = 0, nr_truncate = 0; if (cmpxchg(&root->orphan_cleanup_state, 0, ORPHAN_CLEANUP_STARTED)) - return; + return 0; path = btrfs_alloc_path(); - BUG_ON(!path); + if (!path) { + ret = -ENOMEM; + goto out; + } path->reada = -1; key.objectid = BTRFS_ORPHAN_OBJECTID; @@ -2306,11 +2309,8 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) while (1) { ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); - if (ret < 0) { - printk(KERN_ERR "Error searching slot for orphan: %d" - "\n", ret); - break; - } + if (ret < 0) + goto out; /* * if ret == 0 means we found what we were searching for, which @@ -2318,6 +2318,7 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) * find the key and see if we have stuff that matches */ if (ret > 0) { + ret = 0; if (path->slots[0] == 0) break; path->slots[0]--; @@ -2345,7 +2346,10 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) found_key.type = BTRFS_INODE_ITEM_KEY; found_key.offset = 0; inode = btrfs_iget(root->fs_info->sb, &found_key, root, NULL); - BUG_ON(IS_ERR(inode)); + if (IS_ERR(inode)) { + ret = PTR_ERR(inode); + goto out; + } /* * add this inode to the orphan list so btrfs_orphan_del does @@ -2363,7 +2367,10 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) */ if (is_bad_inode(inode)) { trans = btrfs_start_transaction(root, 0); - BUG_ON(IS_ERR(trans)); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out; + } btrfs_orphan_del(trans, inode); btrfs_end_transaction(trans, root); iput(inode); @@ -2378,16 +2385,16 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) continue; } nr_truncate++; - btrfs_truncate(inode); + ret = btrfs_truncate(inode); } else { nr_unlink++; } /* this will do delete_inode and everything for us */ iput(inode); + if (ret) + goto out; } - btrfs_free_path(path); - root->orphan_cleanup_state = ORPHAN_CLEANUP_DONE; if (root->orphan_block_rsv) @@ -2396,14 +2403,20 @@ void btrfs_orphan_cleanup(struct btrfs_root *root) if (root->orphan_block_rsv || root->orphan_item_inserted) { trans = btrfs_join_transaction(root, 1); - BUG_ON(IS_ERR(trans)); - btrfs_end_transaction(trans, root); + if (!IS_ERR(trans)) + btrfs_end_transaction(trans, root); } if (nr_unlink) printk(KERN_INFO "btrfs: unlinked %d orphans\n", nr_unlink); if (nr_truncate) printk(KERN_INFO "btrfs: truncated %d orphans\n", nr_truncate); + +out: + if (ret) + printk(KERN_CRIT "btrfs: could not do orphan cleanup %d\n", ret); + btrfs_free_path(path); + return ret; } /* @@ -4156,8 +4169,10 @@ struct inode *btrfs_lookup_dentry(struct inode *dir, struct dentry *dentry) if (!IS_ERR(inode) && root != sub_root) { down_read(&root->fs_info->cleanup_work_sem); if (!(inode->i_sb->s_flags & MS_RDONLY)) - btrfs_orphan_cleanup(sub_root); + ret = btrfs_orphan_cleanup(sub_root); up_read(&root->fs_info->cleanup_work_sem); + if (ret) + inode = ERR_PTR(ret); } return inode; -- cgit v1.2.2 From ded5db9de78f963979e1605f859de67626f54693 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 4 Mar 2011 14:09:46 -0500 Subject: Btrfs: make sure to remove the orphan item from the in-memory list This fixes a problem where if truncate fails the inode will still be on the in memory orphan list. This is will make us complain when the inode gets destroyed because it's still on the orphan list. So if we fail just remove us from the in memory list and carry on. Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 0600265cb9b0..3bd0ff63bf30 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6589,6 +6589,12 @@ static int btrfs_truncate(struct inode *inode) ret = btrfs_orphan_del(trans, inode); if (ret) err = ret; + } else if (ret && inode->i_nlink > 0) { + /* + * Failed to do the truncate, remove us from the in memory + * orphan list. + */ + ret = btrfs_orphan_del(NULL, inode); } ret = btrfs_update_inode(trans, root, inode); -- cgit v1.2.2 From f0cd846e9221811d87047f1428cf5226e7236efe Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 4 Mar 2011 14:37:08 -0500 Subject: Btrfs: only add orphan items when truncating We don't need an orphan item when expanding files, we just need them for truncating them, so only add the orphan item in btrfs_truncate instead of in btrfs_setsize. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 45 ++++++++++++++++++--------------------------- 1 file changed, 18 insertions(+), 27 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 3bd0ff63bf30..206b60362cec 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3649,22 +3649,6 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) if (newsize == oldsize) return 0; - trans = btrfs_start_transaction(root, 5); - if (IS_ERR(trans)) - return PTR_ERR(trans); - - btrfs_set_trans_block_group(trans, inode); - - ret = btrfs_orphan_add(trans, inode); - if (ret) { - btrfs_end_transaction(trans, root); - return ret; - } - - nr = trans->blocks_used; - btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); - if (newsize > oldsize) { i_size_write(inode, newsize); btrfs_ordered_update_i_size(inode, i_size_read(inode), NULL); @@ -3675,25 +3659,15 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) return ret; } - trans = btrfs_start_transaction(root, 0); + trans = btrfs_start_transaction(root, 1); if (IS_ERR(trans)) return PTR_ERR(trans); - btrfs_set_trans_block_group(trans, inode); - trans->block_rsv = root->orphan_block_rsv; - BUG_ON(!trans->block_rsv); - - /* - * If this fails just leave the orphan item so that it can get - * cleaned up next time we mount. - */ ret = btrfs_update_inode(trans, root, inode); if (ret) { btrfs_end_transaction(trans, root); return ret; } - if (inode->i_nlink > 0) - ret = btrfs_orphan_del(trans, inode); nr = trans->blocks_used; btrfs_end_transaction(trans, root); btrfs_btree_balance_dirty(root, nr); @@ -6517,6 +6491,23 @@ static int btrfs_truncate(struct inode *inode) btrfs_wait_ordered_range(inode, inode->i_size & (~mask), (u64)-1); btrfs_ordered_update_i_size(inode, inode->i_size, NULL); + trans = btrfs_start_transaction(root, 5); + if (IS_ERR(trans)) + return PTR_ERR(trans); + + btrfs_set_trans_block_group(trans, inode); + + ret = btrfs_orphan_add(trans, inode); + if (ret) { + btrfs_end_transaction(trans, root); + return ret; + } + + nr = trans->blocks_used; + btrfs_end_transaction(trans, root); + btrfs_btree_balance_dirty(root, nr); + + /* Now start a transaction for the truncate */ trans = btrfs_start_transaction(root, 0); if (IS_ERR(trans)) return PTR_ERR(trans); -- cgit v1.2.2 From 930f028abe39dfd0849b53131d19c4b67aacbe67 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 4 Mar 2011 14:41:41 -0500 Subject: Btrfs: use mark_inode_dirty when expanding the file Mark_inode_dirty will call btrfs_dirty_inode which will take care of updating the inode. This makes setsize a little cleaner since we don't have to start a transaction and update the inode in there, we can just call mark_inode_dirty. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 16 +--------------- 1 file changed, 1 insertion(+), 15 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 206b60362cec..64d57e032b4e 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3640,10 +3640,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) static int btrfs_setsize(struct inode *inode, loff_t newsize) { - struct btrfs_root *root = BTRFS_I(inode)->root; - struct btrfs_trans_handle *trans; loff_t oldsize = i_size_read(inode); - unsigned long nr; int ret; if (newsize == oldsize) @@ -3659,18 +3656,7 @@ static int btrfs_setsize(struct inode *inode, loff_t newsize) return ret; } - trans = btrfs_start_transaction(root, 1); - if (IS_ERR(trans)) - return PTR_ERR(trans); - - ret = btrfs_update_inode(trans, root, inode); - if (ret) { - btrfs_end_transaction(trans, root); - return ret; - } - nr = trans->blocks_used; - btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + mark_inode_dirty(inode); } else { /* -- cgit v1.2.2 From 695a0d0da09e75c4475bbb303def159023ef72ca Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 4 Mar 2011 15:46:53 -0500 Subject: Btrfs: add a comment explaining what btrfs_cont_expand does Everytime I have to deal with btrfs_cont_expand I stare at it for 20 minutes trying to remember what exactly it does and why the hell we need it. So add a comment to save future-Josef some time. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 64d57e032b4e..888dbdb3b128 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3557,6 +3557,12 @@ out: return ret; } +/* + * This function puts in dummy file extents for the area we're creating a hole + * for. So if we are truncating this file to a larger size we need to insert + * these file extents so that btrfs_get_extent will return a EXTENT_MAP_HOLE for + * the range between oldsize and size + */ int btrfs_cont_expand(struct inode *inode, loff_t oldsize, loff_t size) { struct btrfs_trans_handle *trans; -- cgit v1.2.2 From 22a94d44bd6876a90630338229da6c0436d46593 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 16 Mar 2011 16:47:17 -0400 Subject: Btrfs: add checks to verify dir items are correct We need to make sure the dir items we get are valid dir items. So any time we try and read one check it with verify_dir_item, which will do various sanity checks to make sure it looks sane. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 888dbdb3b128..e010000d4bc9 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4272,6 +4272,9 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, while (di_cur < di_total) { struct btrfs_key location; + if (verify_dir_item(root, leaf, di)) + break; + name_len = btrfs_dir_name_len(leaf, di); if (name_len <= sizeof(tmp_name)) { name_ptr = tmp_name; -- cgit v1.2.2 From 98bc3149fad639c8f50c7110b961a2a2fe085eed Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 22 Mar 2011 11:00:46 -0400 Subject: Btrfs: don't allocate dip->csums when doing writes When doing direct writes we store the checksums in the ordered sum stuff in the ordered extent for writing them when the write completes, so we don't even use the dip->csums array. So if we're writing, don't bother allocating dip->csums since we won't use it anyway. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e010000d4bc9..570cd44fe91b 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5944,6 +5944,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, int nr_pages = 0; u32 *csums = dip->csums; int ret = 0; + int write = rw & REQ_WRITE; bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); if (!bio) @@ -5980,7 +5981,8 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, goto out_err; } - if (!skip_sum) + /* Write's use the ordered csums */ + if (!write && !skip_sum) csums = csums + nr_pages; start_sector += submit_len >> 9; file_offset += submit_len; @@ -6048,7 +6050,8 @@ static void btrfs_submit_direct(int rw, struct bio *bio, struct inode *inode, } dip->csums = NULL; - if (!skip_sum) { + /* Write's use the ordered csum stuff, so we don't need dip->csums */ + if (!write && !skip_sum) { dip->csums = kmalloc(sizeof(u32) * bio->bi_vcnt, GFP_NOFS); if (!dip->csums) { kfree(dip); -- cgit v1.2.2 From c0da7aa1a2d8fcafe271a7077599253c8ed94bb2 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 22 Mar 2011 11:05:07 -0400 Subject: Btrfs: mark the bio with an error if we have a failure in dio I noticed that dio_end_io calls the appropriate endio function with an error, but the endio functions don't actually do anything with that error, they assume that if there was an error then the bio will not be uptodate. So if we had checksum failures we would never pass back EIO. So if there is an error in our endio functions make sure to clear the uptodate flag on the bio. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 570cd44fe91b..e9813bd7d556 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5744,6 +5744,10 @@ static void btrfs_endio_direct_read(struct bio *bio, int err) kfree(dip->csums); kfree(dip); + + /* If we had a csum failure make sure to clear the uptodate flag */ + if (err) + clear_bit(BIO_UPTODATE, &bio->bi_flags); dio_end_io(bio, err); } @@ -5845,6 +5849,10 @@ out_done: kfree(dip->csums); kfree(dip); + + /* If we had an error make sure to clear the uptodate flag */ + if (err) + clear_bit(BIO_UPTODATE, &bio->bi_flags); dio_end_io(bio, err); } -- cgit v1.2.2 From 1abe9b8a138c9988ba8f7bfded6453649a31541f Mon Sep 17 00:00:00 2001 From: liubo Date: Thu, 24 Mar 2011 11:18:59 +0000 Subject: Btrfs: add initial tracepoint support for btrfs Tracepoints can provide insight into why btrfs hits bugs and be greatly helpful for debugging, e.g dd-7822 [000] 2121.641088: btrfs_inode_request: root = 5(FS_TREE), gen = 4, ino = 256, blocks = 8, disk_i_size = 0, last_trans = 8, logged_trans = 0 dd-7822 [000] 2121.641100: btrfs_inode_new: root = 5(FS_TREE), gen = 8, ino = 257, blocks = 0, disk_i_size = 0, last_trans = 0, logged_trans = 0 btrfs-transacti-7804 [001] 2146.935420: btrfs_cow_block: root = 2(EXTENT_TREE), refs = 2, orig_buf = 29368320 (orig_level = 0), cow_buf = 29388800 (cow_level = 0) btrfs-transacti-7804 [001] 2146.935473: btrfs_cow_block: root = 1(ROOT_TREE), refs = 2, orig_buf = 29364224 (orig_level = 0), cow_buf = 29392896 (cow_level = 0) btrfs-transacti-7804 [001] 2146.972221: btrfs_transaction_commit: root = 1(ROOT_TREE), gen = 8 flush-btrfs-2-7821 [001] 2155.824210: btrfs_chunk_alloc: root = 3(CHUNK_TREE), offset = 1103101952, size = 1073741824, num_stripes = 1, sub_stripes = 0, type = DATA flush-btrfs-2-7821 [001] 2155.824241: btrfs_cow_block: root = 2(EXTENT_TREE), refs = 2, orig_buf = 29388800 (orig_level = 0), cow_buf = 29396992 (cow_level = 0) flush-btrfs-2-7821 [001] 2155.824255: btrfs_cow_block: root = 4(DEV_TREE), refs = 2, orig_buf = 29372416 (orig_level = 0), cow_buf = 29401088 (cow_level = 0) flush-btrfs-2-7821 [000] 2155.824329: btrfs_cow_block: root = 3(CHUNK_TREE), refs = 2, orig_buf = 20971520 (orig_level = 0), cow_buf = 20975616 (cow_level = 0) btrfs-endio-wri-7800 [001] 2155.898019: btrfs_cow_block: root = 5(FS_TREE), refs = 2, orig_buf = 29384704 (orig_level = 0), cow_buf = 29405184 (cow_level = 0) btrfs-endio-wri-7800 [001] 2155.898043: btrfs_cow_block: root = 7(CSUM_TREE), refs = 2, orig_buf = 29376512 (orig_level = 0), cow_buf = 29409280 (cow_level = 0) Here is what I have added: 1) ordere_extent: btrfs_ordered_extent_add btrfs_ordered_extent_remove btrfs_ordered_extent_start btrfs_ordered_extent_put These provide critical information to understand how ordered_extents are updated. 2) extent_map: btrfs_get_extent extent_map is used in both read and write cases, and it is useful for tracking how btrfs specific IO is running. 3) writepage: __extent_writepage btrfs_writepage_end_io_hook Pages are cirtical resourses and produce a lot of corner cases during writeback, so it is valuable to know how page is written to disk. 4) inode: btrfs_inode_new btrfs_inode_request btrfs_inode_evict These can show where and when a inode is created, when a inode is evicted. 5) sync: btrfs_sync_file btrfs_sync_fs These show sync arguments. 6) transaction: btrfs_transaction_commit In transaction based filesystem, it will be useful to know the generation and who does commit. 7) back reference and cow: btrfs_delayed_tree_ref btrfs_delayed_data_ref btrfs_delayed_ref_head btrfs_cow_block Btrfs natively supports back references, these tracepoints are helpful on understanding btrfs's COW mechanism. 8) chunk: btrfs_chunk_alloc btrfs_chunk_free Chunk is a link between physical offset and logical offset, and stands for space infomation in btrfs, and these are helpful on tracing space things. 9) reserved_extent: btrfs_reserved_extent_alloc btrfs_reserved_extent_free These can show how btrfs uses its space. Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e9813bd7d556..eaa271484199 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1787,6 +1787,8 @@ out: static int btrfs_writepage_end_io_hook(struct page *page, u64 start, u64 end, struct extent_state *state, int uptodate) { + trace_btrfs_writepage_end_io_hook(page, start, end, uptodate); + ClearPagePrivate2(page); return btrfs_finish_ordered_io(page->mapping->host, start, end); } @@ -3718,6 +3720,8 @@ void btrfs_evict_inode(struct inode *inode) unsigned long nr; int ret; + trace_btrfs_inode_evict(inode); + truncate_inode_pages(&inode->i_data, 0); if (inode->i_nlink && (btrfs_root_refs(&root->root_item) != 0 || root == root->fs_info->tree_root)) @@ -4510,6 +4514,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, return ERR_PTR(-ENOMEM); if (dir) { + trace_btrfs_inode_request(dir); + ret = btrfs_set_inode_index(dir, index); if (ret) { iput(inode); @@ -4584,6 +4590,9 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, insert_inode_hash(inode); inode_tree_add(inode); + + trace_btrfs_inode_new(inode); + return inode; fail: if (dir) @@ -5261,6 +5270,9 @@ insert: } write_unlock(&em_tree->lock); out: + + trace_btrfs_get_extent(root, em); + if (path) btrfs_free_path(path); if (trans) { -- cgit v1.2.2 From 75e7cb7fe0c391561bd3af36515be3f3c64a04c6 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Tue, 22 Mar 2011 10:12:20 +0000 Subject: Btrfs: Per file/directory controls for COW and compression Data compression and data cow are controlled across the entire FS by mount options right now. ioctls are needed to set this on a per file or per directory basis. This has been proposed previously, but VFS developers wanted us to use generic ioctls rather than btrfs-specific ones. According to Chris's comment, there should be just one true compression method(probably LZO) stored in the super. However, before this, we would wait for that one method is stable enough to be adopted into the super. So I list it as a long term goal, and just store it in ram today. After applying this patch, we can use the generic "FS_IOC_SETFLAGS" ioctl to control file and directory's datacow and compression attribute. NOTE: - The compression type is selected by such rules: If we mount btrfs with compress options, ie, zlib/lzo, the type is it. Otherwise, we'll use the default compress type (zlib today). v1->v2: - rebase to the latest btrfs. v2->v3: - fix a problem, i.e. when a file is set NOCOW via mount option, then this NOCOW will be screwed by inheritance from parent directory. Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 31 ++++++++++++++++++++++++++++--- 1 file changed, 28 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index eaa271484199..7a7a202b82ab 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -384,7 +384,8 @@ again: */ if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS) && (btrfs_test_opt(root, COMPRESS) || - (BTRFS_I(inode)->force_compress))) { + (BTRFS_I(inode)->force_compress) || + (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS))) { WARN_ON(pages); pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); @@ -1256,7 +1257,8 @@ static int run_delalloc_range(struct inode *inode, struct page *locked_page, ret = run_delalloc_nocow(inode, locked_page, start, end, page_started, 0, nr_written); else if (!btrfs_test_opt(root, COMPRESS) && - !(BTRFS_I(inode)->force_compress)) + !(BTRFS_I(inode)->force_compress) && + !(BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS)) ret = cow_file_range(inode, locked_page, start, end, page_started, nr_written, 1); else @@ -4584,7 +4586,8 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, if ((mode & S_IFREG)) { if (btrfs_test_opt(root, NODATASUM)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATASUM; - if (btrfs_test_opt(root, NODATACOW)) + if (btrfs_test_opt(root, NODATACOW) || + (BTRFS_I(dir)->flags & BTRFS_INODE_NODATACOW)) BTRFS_I(inode)->flags |= BTRFS_INODE_NODATACOW; } @@ -6866,6 +6869,26 @@ static int btrfs_getattr(struct vfsmount *mnt, return 0; } +/* + * If a file is moved, it will inherit the cow and compression flags of the new + * directory. + */ +static void fixup_inode_flags(struct inode *dir, struct inode *inode) +{ + struct btrfs_inode *b_dir = BTRFS_I(dir); + struct btrfs_inode *b_inode = BTRFS_I(inode); + + if (b_dir->flags & BTRFS_INODE_NODATACOW) + b_inode->flags |= BTRFS_INODE_NODATACOW; + else + b_inode->flags &= ~BTRFS_INODE_NODATACOW; + + if (b_dir->flags & BTRFS_INODE_COMPRESS) + b_inode->flags |= BTRFS_INODE_COMPRESS; + else + b_inode->flags &= ~BTRFS_INODE_COMPRESS; +} + static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry) { @@ -6999,6 +7022,8 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, } } + fixup_inode_flags(new_dir, old_inode); + ret = btrfs_add_link(trans, new_dir, old_inode, new_dentry->d_name.name, new_dentry->d_name.len, 0, index); -- cgit v1.2.2 From 3ab3564f018b9b265d0258e4a231794bacd5ad85 Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 22 Mar 2011 17:20:26 +0000 Subject: btrfs: return EXDEV when linking from different subvolumes btrfs_link returns EPERM if a cross-subvolume link is attempted. However, in this case I believe EXDEV to be the more appropriate value. >From the link(2) man page: EXDEV oldpath and newpath are not on the same mounted file system. (Linux permits a file system to be mounted at multiple points, but link() does not work across different mount points, even if the same file system is mounted on both.) This matters because an application may have different behaviors based on return codes. Signed-off-by: Mark Fasheh Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 7a7a202b82ab..67fd6e9552d3 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4817,7 +4817,7 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, /* do not allow sys_link's with other subvols of the same device */ if (root->objectid != BTRFS_I(inode)->root->objectid) - return -EPERM; + return -EXDEV; btrfs_inc_nlink(inode); inode->i_ctime = CURRENT_TIME; -- cgit v1.2.2 From dac97e516c617f9c797f64b0224050b70aea30c7 Mon Sep 17 00:00:00 2001 From: Yoshinori Sano Date: Tue, 15 Feb 2011 12:01:42 +0000 Subject: Btrfs: fix uncheck memory allocations To make Btrfs code more robust, several return value checks where memory allocation can fail are introduced. I use BUG_ON where I don't know how to handle the error properly, which increases the number of using the notorious BUG_ON, though. Signed-off-by: Yoshinori Sano Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 67fd6e9552d3..f739b256967e 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -290,6 +290,7 @@ static noinline int add_async_extent(struct async_cow *cow, struct async_extent *async_extent; async_extent = kmalloc(sizeof(*async_extent), GFP_NOFS); + BUG_ON(!async_extent); async_extent->start = start; async_extent->ram_size = ram_size; async_extent->compressed_size = compressed_size; @@ -388,6 +389,7 @@ again: (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS))) { WARN_ON(pages); pages = kzalloc(sizeof(struct page *) * nr_pages, GFP_NOFS); + BUG_ON(!pages); if (BTRFS_I(inode)->force_compress) compress_type = BTRFS_I(inode)->force_compress; -- cgit v1.2.2 From c2db1073fdf9757e6fd8b4a59d15b6ecc7a2af8a Mon Sep 17 00:00:00 2001 From: Tsutomu Itoh Date: Tue, 1 Mar 2011 06:48:31 +0000 Subject: Btrfs: check return value of btrfs_alloc_path() Adding the check on the return value of btrfs_alloc_path() to several places. And, some of callers are modified by this change. Signed-off-by: Tsutomu Itoh Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index f739b256967e..04e9fffee8cc 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1467,8 +1467,11 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, if (bio_flags & EXTENT_BIO_COMPRESSED) { return btrfs_submit_compressed_read(inode, bio, mirror_num, bio_flags); - } else if (!skip_sum) - btrfs_lookup_bio_sums(root, inode, bio, NULL); + } else if (!skip_sum) { + ret = btrfs_lookup_bio_sums(root, inode, bio, NULL); + if (ret) + return ret; + } goto mapit; } else if (!skip_sum) { /* csum items have already been cloned */ @@ -1903,10 +1906,10 @@ static int btrfs_io_failed_hook(struct bio *failed_bio, else rw = READ; - BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio, + ret = BTRFS_I(inode)->io_tree.ops->submit_bio_hook(inode, rw, bio, failrec->last_mirror, failrec->bio_flags, 0); - return 0; + return ret; } /* @@ -5943,9 +5946,12 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, __btrfs_submit_bio_start_direct_io, __btrfs_submit_bio_done); goto err; - } else if (!skip_sum) - btrfs_lookup_bio_sums_dio(root, inode, bio, + } else if (!skip_sum) { + ret = btrfs_lookup_bio_sums_dio(root, inode, bio, file_offset, csums); + if (ret) + goto err; + } ret = btrfs_map_bio(root, rw, bio, 0, 1); err: -- cgit v1.2.2 From 92986796d84ef939e304099dece32572a755b280 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 4 Mar 2011 17:14:37 +0000 Subject: btrfs: don't mess with i_nlink of unlocked inode in rename() old_inode is not locked; it's not safe to play with its link count. Instead of bumping it and calling btrfs_unlink_inode(), add a variant of the latter that does not do btrfs_drop_nlink()/ btrfs_update_inode(), call it instead of btrfs_inc_nlink()/ btrfs_unlink_inode() and do btrfs_update_inode() ourselves. Signed-off-by: Al Viro Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 36 +++++++++++++++++++++++++----------- 1 file changed, 25 insertions(+), 11 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 04e9fffee8cc..4822b3132784 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2664,10 +2664,10 @@ failed: * recovery code. It remove a link in a directory with a given name, and * also drops the back refs in the inode to the directory */ -int btrfs_unlink_inode(struct btrfs_trans_handle *trans, - struct btrfs_root *root, - struct inode *dir, struct inode *inode, - const char *name, int name_len) +static int __btrfs_unlink_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct inode *dir, struct inode *inode, + const char *name, int name_len) { struct btrfs_path *path; int ret = 0; @@ -2739,12 +2739,25 @@ err: btrfs_i_size_write(dir, dir->i_size - name_len * 2); inode->i_ctime = dir->i_mtime = dir->i_ctime = CURRENT_TIME; btrfs_update_inode(trans, root, dir); - btrfs_drop_nlink(inode); - ret = btrfs_update_inode(trans, root, inode); out: return ret; } +int btrfs_unlink_inode(struct btrfs_trans_handle *trans, + struct btrfs_root *root, + struct inode *dir, struct inode *inode, + const char *name, int name_len) +{ + int ret; + ret = __btrfs_unlink_inode(trans, root, dir, inode, name, name_len); + if (!ret) { + btrfs_drop_nlink(inode); + ret = btrfs_update_inode(trans, root, inode); + } + return ret; +} + + /* helper to check if there is any shared block in the path */ static int check_path_shared(struct btrfs_root *root, struct btrfs_path *path) @@ -6999,11 +7012,12 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, old_dentry->d_name.name, old_dentry->d_name.len); } else { - btrfs_inc_nlink(old_dentry->d_inode); - ret = btrfs_unlink_inode(trans, root, old_dir, - old_dentry->d_inode, - old_dentry->d_name.name, - old_dentry->d_name.len); + ret = __btrfs_unlink_inode(trans, root, old_dir, + old_dentry->d_inode, + old_dentry->d_name.name, + old_dentry->d_name.len); + if (!ret) + ret = btrfs_update_inode(trans, root, old_inode); } BUG_ON(ret); -- cgit v1.2.2 From c055e99eea6e4f614267632fac546e7896c0227b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 4 Mar 2011 17:15:18 +0000 Subject: btrfs: check link counter overflow in link(2) Signed-off-by: Al Viro Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 4822b3132784..04babaf31a33 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4837,6 +4837,9 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, if (root->objectid != BTRFS_I(inode)->root->objectid) return -EXDEV; + if (inode->i_nlink == ~0U) + return -EMLINK; + btrfs_inc_nlink(inode); inode->i_ctime = CURRENT_TIME; -- cgit v1.2.2 From 1561deda687eef0e95065f1268d680ddc5976ee7 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Sun, 27 Mar 2011 16:07:36 +0800 Subject: btrfs: fix possible deadlock by clearing __GFP_FS flag Using the GFP_HIGHUSER_MOVABLE flag to allocate the metadata's page may cause deadlock. Task1 open() ... btrfs_search_slot() ... btrfs_cow_block() ... alloc_page() wait for reclaiming shrink_slab() ... shrink_icache_memory() ... btrfs_evict_inode() ... btrfs_search_slot() If the path is locked by task1, the deadlock happens. So the btree's page cache is different with the file's page cache, it can not allocate pages by GFP_HIGHUSER_MOVABLE flag, we must clear __GFP_FS flag in GFP_HIGHUSER_MOVABLE flag. Reported-by: Itaru Kitayama Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 04babaf31a33..06274186b290 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2536,6 +2536,8 @@ static void btrfs_read_locked_inode(struct inode *inode) BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item); alloc_group_block = btrfs_inode_block_group(leaf, inode_item); + if (location.objectid == BTRFS_FREE_SPACE_OBJECTID) + inode->i_mapping->flags &= ~__GFP_FS; /* * try to precache a NULL acl entry for files that don't have @@ -4084,7 +4086,6 @@ struct inode *btrfs_iget(struct super_block *s, struct btrfs_key *location, BTRFS_I(inode)->root = root; memcpy(&BTRFS_I(inode)->location, location, sizeof(*location)); btrfs_read_locked_inode(inode); - inode_tree_add(inode); unlock_new_inode(inode); if (new) -- cgit v1.2.2 From 25985edcedea6396277003854657b5f3cb31a628 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Wed, 30 Mar 2011 22:57:33 -0300 Subject: Fix common misspellings Fixes generated by 'codespell' and manually reviewed. Signed-off-by: Lucas De Marchi --- fs/btrfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 93c28a1d6bdc..80920bce01ab 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2324,7 +2324,7 @@ int btrfs_orphan_cleanup(struct btrfs_root *root) /* * if ret == 0 means we found what we were searching for, which - * is weird, but possible, so only screw with path if we didnt + * is weird, but possible, so only screw with path if we didn't * find the key and see if we have stuff that matches */ if (ret > 0) { -- cgit v1.2.2 From fe3f566cd19bb6d787c92b2e202c85f929abf3ac Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Mon, 28 Mar 2011 08:30:38 +0000 Subject: Btrfs: Fix oops for defrag with compression turned on When we defrag a file, whose size can be fit into an inline extent, with compression enabled, the compress type is set to be fs_info->compress_type, which is 0 if the btrfs filesystem is mounted without compress option. This leads to oops. Reported-by: Daniel Blueman Signed-off-by: Li Zefan Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 06274186b290..62ae9d5da806 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -111,6 +111,7 @@ static int btrfs_init_inode_security(struct btrfs_trans_handle *trans, static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *inode, u64 start, size_t size, size_t compressed_size, + int compress_type, struct page **compressed_pages) { struct btrfs_key key; @@ -125,12 +126,9 @@ static noinline int insert_inline_extent(struct btrfs_trans_handle *trans, size_t cur_size = size; size_t datasize; unsigned long offset; - int compress_type = BTRFS_COMPRESS_NONE; - if (compressed_size && compressed_pages) { - compress_type = root->fs_info->compress_type; + if (compressed_size && compressed_pages) cur_size = compressed_size; - } path = btrfs_alloc_path(); if (!path) @@ -220,7 +218,7 @@ fail: static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans, struct btrfs_root *root, struct inode *inode, u64 start, u64 end, - size_t compressed_size, + size_t compressed_size, int compress_type, struct page **compressed_pages) { u64 isize = i_size_read(inode); @@ -253,7 +251,7 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans, inline_len = min_t(u64, isize, actual_end); ret = insert_inline_extent(trans, root, inode, start, inline_len, compressed_size, - compressed_pages); + compress_type, compressed_pages); BUG_ON(ret); btrfs_delalloc_release_metadata(inode, end + 1 - start); btrfs_drop_extent_cache(inode, start, aligned_end - 1, 0); @@ -432,12 +430,13 @@ again: * to make an uncompressed inline extent. */ ret = cow_file_range_inline(trans, root, inode, - start, end, 0, NULL); + start, end, 0, 0, NULL); } else { /* try making a compressed inline extent */ ret = cow_file_range_inline(trans, root, inode, start, end, - total_compressed, pages); + total_compressed, + compress_type, pages); } if (ret == 0) { /* @@ -791,7 +790,7 @@ static noinline int cow_file_range(struct inode *inode, if (start == 0) { /* lets try to make an inline extent */ ret = cow_file_range_inline(trans, root, inode, - start, end, 0, NULL); + start, end, 0, 0, NULL); if (ret == 0) { extent_clear_unlock_delalloc(inode, &BTRFS_I(inode)->io_tree, -- cgit v1.2.2 From b44c59a80ded004e1a82712e5f9e17b131c03221 Mon Sep 17 00:00:00 2001 From: Johann Lombardi Date: Thu, 31 Mar 2011 13:23:47 +0000 Subject: Btrfs: fix subvol_sem leak in btrfs_rename() btrfs_rename() does not release the subvol_sem if the transaction failed to start. Signed-off-by: Johann Lombardi Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 62ae9d5da806..1ca3e68586cf 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6961,8 +6961,10 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, * should cover the worst case number of items we'll modify. */ trans = btrfs_start_transaction(root, 20); - if (IS_ERR(trans)) - return PTR_ERR(trans); + if (IS_ERR(trans)) { + ret = PTR_ERR(trans); + goto out_notrans; + } btrfs_set_trans_block_group(trans, new_dir); @@ -7062,7 +7064,7 @@ static int btrfs_rename(struct inode *old_dir, struct dentry *old_dentry, } out_fail: btrfs_end_transaction_throttle(trans, root); - +out_notrans: if (old_inode->i_ino == BTRFS_FIRST_FREE_OBJECTID) up_read(&root->fs_info->subvol_sem); -- cgit v1.2.2 From adae52b94e18afa1f84fab67df2a8a872c2f5533 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Thu, 31 Mar 2011 09:43:23 +0000 Subject: btrfs: clear __GFP_FS flag in the space cache inode the object id of the space cache inode's key is allocated from the relative root, just like the regular file. So we can't identify space cache inode by checking the object id of the inode's key, and we have to clear __GFP_FS flag at the time we look up the space cache inode. Signed-off-by: Miao Xie Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 1ca3e68586cf..57a03f6eb224 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2535,8 +2535,6 @@ static void btrfs_read_locked_inode(struct inode *inode) BTRFS_I(inode)->flags = btrfs_inode_flags(leaf, inode_item); alloc_group_block = btrfs_inode_block_group(leaf, inode_item); - if (location.objectid == BTRFS_FREE_SPACE_OBJECTID) - inode->i_mapping->flags &= ~__GFP_FS; /* * try to precache a NULL acl entry for files that don't have -- cgit v1.2.2 From c9ddec74aa950a220cc4caa5215cfc5d886050b7 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 28 Mar 2011 13:43:25 +0000 Subject: Btrfs: don't warn in btrfs_add_orphan When I moved the orphan adding to btrfs_truncate I missed the fact that during orphan cleanup we just add the orphan items to the orphan list without going through btrfs_orphan_add, which results in lots of warnings on mount if you have any orphan items that need to be truncated. Just remove this warning since it's ok, this will allow all of the normal space accounting take place. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 57a03f6eb224..cc6022842e0c 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2220,8 +2220,6 @@ int btrfs_orphan_add(struct btrfs_trans_handle *trans, struct inode *inode) insert = 1; #endif insert = 1; - } else { - WARN_ON(!BTRFS_I(inode)->orphan_meta_reserved); } if (!BTRFS_I(inode)->orphan_meta_reserved) { -- cgit v1.2.2 From 12ddb96cb6752218d8a1aeb696ec9b0ca7adb42f Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 5 Apr 2011 13:02:27 -0400 Subject: Btrfs: map the inode item when doing fill_inode_item Instead of calling kmap_atomic for every thing we set in the inode item, map the entire inode item at the start and unmap it at the end. This makes a sequential dd of 400mb O_DIRECT something like 1% faster. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index cc6022842e0c..da2680263d9f 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -2589,6 +2589,13 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, struct btrfs_inode_item *item, struct inode *inode) { + if (!leaf->map_token) + map_private_extent_buffer(leaf, (unsigned long)item, + sizeof(struct btrfs_inode_item), + &leaf->map_token, &leaf->kaddr, + &leaf->map_start, &leaf->map_len, + KM_USER1); + btrfs_set_inode_uid(leaf, item, inode->i_uid); btrfs_set_inode_gid(leaf, item, inode->i_gid); btrfs_set_inode_size(leaf, item, BTRFS_I(inode)->disk_i_size); @@ -2617,6 +2624,11 @@ static void fill_inode_item(struct btrfs_trans_handle *trans, btrfs_set_inode_rdev(leaf, item, inode->i_rdev); btrfs_set_inode_flags(leaf, item, BTRFS_I(inode)->flags); btrfs_set_inode_block_group(leaf, item, BTRFS_I(inode)->block_group); + + if (leaf->map_token) { + unmap_extent_buffer(leaf, leaf->map_token, KM_USER1); + leaf->map_token = NULL; + } } /* -- cgit v1.2.2 From 1ef30be142d2cc60e2687ef267de864cf31be995 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Tue, 5 Apr 2011 19:25:36 -0400 Subject: Btrfs: do not call btrfs_update_inode in endio if nothing changed In the DIO code we often don't update the i_disk_size because the i_size isn't updated until after the DIO is completed, so basically we are allocating a path, doing a search, and updating the inode item for no reason since nothing changed. btrfs_ordered_update_i_size will return 1 if it didn't update i_disk_size, so only run btrfs_update_inode if btrfs_ordered_update_i_size returns 0. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index da2680263d9f..4a238d676e5b 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1769,9 +1769,12 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end) add_pending_csums(trans, inode, ordered_extent->file_offset, &ordered_extent->list); - btrfs_ordered_update_i_size(inode, 0, ordered_extent); - ret = btrfs_update_inode(trans, root, inode); - BUG_ON(ret); + ret = btrfs_ordered_update_i_size(inode, 0, ordered_extent); + if (!ret) { + ret = btrfs_update_inode(trans, root, inode); + BUG_ON(ret); + } + ret = 0; out: if (nolock) { if (trans) @@ -5865,8 +5868,10 @@ again: } add_pending_csums(trans, inode, ordered->file_offset, &ordered->list); - btrfs_ordered_update_i_size(inode, 0, ordered); - btrfs_update_inode(trans, root, inode); + ret = btrfs_ordered_update_i_size(inode, 0, ordered); + if (!ret) + btrfs_update_inode(trans, root, inode); + ret = 0; out_unlock: unlock_extent_cached(&BTRFS_I(inode)->io_tree, ordered->file_offset, ordered->file_offset + ordered->len - 1, -- cgit v1.2.2 From 02f57c7aedef1a537f4b16db7061cdd8efa3bb4e Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 6 Apr 2011 14:25:44 -0400 Subject: Btrfs: don't split dio bios if we don't have to We have been unconditionally allocating a new bio and re-adding all pages from our original bio to the new bio. This is needed if our original bio is larger than our stripe size, but if it is smaller than the stripe size then there is no need to do this. So check the map length and if we are under that then go ahead and submit the original bio. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 20 +++++++++++++------- 1 file changed, 13 insertions(+), 7 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 4a238d676e5b..149c77fd1eb5 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6006,13 +6006,6 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, int ret = 0; int write = rw & REQ_WRITE; - bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); - if (!bio) - return -ENOMEM; - bio->bi_private = dip; - bio->bi_end_io = btrfs_end_dio_bio; - atomic_inc(&dip->pending_bios); - map_length = orig_bio->bi_size; ret = btrfs_map_block(map_tree, READ, start_sector << 9, &map_length, NULL, 0); @@ -6021,6 +6014,18 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, return -EIO; } + if (map_length >= orig_bio->bi_size) { + bio = orig_bio; + goto submit; + } + + bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); + if (!bio) + return -ENOMEM; + bio->bi_private = dip; + bio->bi_end_io = btrfs_end_dio_bio; + atomic_inc(&dip->pending_bios); + while (bvec <= (orig_bio->bi_io_vec + orig_bio->bi_vcnt - 1)) { if (unlikely(map_length < submit_len + bvec->bv_len || bio_add_page(bio, bvec->bv_page, bvec->bv_len, @@ -6071,6 +6076,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, } } +submit: ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, csums); if (!ret) -- cgit v1.2.2 From 1ae399382512b3e4d6c923e53da9e45935577040 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 6 Apr 2011 14:41:34 -0400 Subject: Btrfs: do not use async submit for small DIO io's When looking at our DIO performance Chris said that for small IO's doing the async submit stuff tends to be more overhead than it's worth. With this on top of my other fixes I get about a 17-20% speedup doing a sequential dd with 4k IO's. Basically if we don't have to split the bio for the map length it's small enough to be directly submitted, otherwise go back to the async submit. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 149c77fd1eb5..2bb76c6157a4 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5957,7 +5957,7 @@ static struct bio *btrfs_dio_bio_alloc(struct block_device *bdev, static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, int rw, u64 file_offset, int skip_sum, - u32 *csums) + u32 *csums, int async_submit) { int write = rw & REQ_WRITE; struct btrfs_root *root = BTRFS_I(inode)->root; @@ -5968,13 +5968,24 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, if (ret) goto err; - if (write && !skip_sum) { + if (skip_sum) + goto map; + + if (write && async_submit) { ret = btrfs_wq_submit_bio(root->fs_info, inode, rw, bio, 0, 0, file_offset, __btrfs_submit_bio_start_direct_io, __btrfs_submit_bio_done); goto err; + } else if (write) { + /* + * If we aren't doing async submit, calculate the csum of the + * bio now. + */ + ret = btrfs_csum_one_bio(root, inode, bio, file_offset, 1); + if (ret) + goto err; } else if (!skip_sum) { ret = btrfs_lookup_bio_sums_dio(root, inode, bio, file_offset, csums); @@ -5982,7 +5993,8 @@ static inline int __btrfs_submit_dio_bio(struct bio *bio, struct inode *inode, goto err; } - ret = btrfs_map_bio(root, rw, bio, 0, 1); +map: + ret = btrfs_map_bio(root, rw, bio, 0, async_submit); err: bio_put(bio); return ret; @@ -6004,6 +6016,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, int nr_pages = 0; u32 *csums = dip->csums; int ret = 0; + int async_submit = 0; int write = rw & REQ_WRITE; map_length = orig_bio->bi_size; @@ -6019,6 +6032,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, goto submit; } + async_submit = 1; bio = btrfs_dio_bio_alloc(orig_bio->bi_bdev, start_sector, GFP_NOFS); if (!bio) return -ENOMEM; @@ -6039,7 +6053,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, atomic_inc(&dip->pending_bios); ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, - csums); + csums, async_submit); if (ret) { bio_put(bio); atomic_dec(&dip->pending_bios); @@ -6078,7 +6092,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, submit: ret = __btrfs_submit_dio_bio(bio, inode, rw, file_offset, skip_sum, - csums); + csums, async_submit); if (!ret) return 0; -- cgit v1.2.2 From 16d299ac7446b5a75c5683a9ae11d7907d444c86 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 6 Apr 2011 14:53:07 -0400 Subject: Btrfs: reuse the extent_map we found when calling btrfs_get_extent In btrfs_get_block_direct we call btrfs_get_extent to lookup the extent for the range that we are looking for. If we don't find an extent, btrfs_get_extent will insert a extent_map for that area and mark it as a hole. So it does the job of allocating a new extent map and inserting it into the io tree. But if we're creating a new extent we free it up and redo all of that work. So instead pass the em to btrfs_new_extent_direct(), and if it will work just allocate the disk space and set it up properly and bypass the freeing/allocating of a new extent map and the expensive operation of inserting the thing into the io_tree. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 36 ++++++++++++++++++++++++++++-------- 1 file changed, 28 insertions(+), 8 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 2bb76c6157a4..24310c9cb14f 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5445,17 +5445,30 @@ out: } static struct extent_map *btrfs_new_extent_direct(struct inode *inode, + struct extent_map *em, u64 start, u64 len) { struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_trans_handle *trans; - struct extent_map *em; struct extent_map_tree *em_tree = &BTRFS_I(inode)->extent_tree; struct btrfs_key ins; u64 alloc_hint; int ret; + bool insert = false; - btrfs_drop_extent_cache(inode, start, start + len - 1, 0); + /* + * Ok if the extent map we looked up is a hole and is for the exact + * range we want, there is no reason to allocate a new one, however if + * it is not right then we need to free this one and drop the cache for + * our range. + */ + if (em->block_start != EXTENT_MAP_HOLE || em->start != start || + em->len != len) { + free_extent_map(em); + em = NULL; + insert = true; + btrfs_drop_extent_cache(inode, start, start + len - 1, 0); + } trans = btrfs_join_transaction(root, 0); if (IS_ERR(trans)) @@ -5471,10 +5484,12 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, goto out; } - em = alloc_extent_map(GFP_NOFS); if (!em) { - em = ERR_PTR(-ENOMEM); - goto out; + em = alloc_extent_map(GFP_NOFS); + if (!em) { + em = ERR_PTR(-ENOMEM); + goto out; + } } em->start = start; @@ -5484,9 +5499,15 @@ static struct extent_map *btrfs_new_extent_direct(struct inode *inode, em->block_start = ins.objectid; em->block_len = ins.offset; em->bdev = root->fs_info->fs_devices->latest_bdev; + + /* + * We need to do this because if we're using the original em we searched + * for, we could have EXTENT_FLAG_VACANCY set, and we don't want that. + */ + em->flags = 0; set_bit(EXTENT_FLAG_PINNED, &em->flags); - while (1) { + while (insert) { write_lock(&em_tree->lock); ret = add_extent_mapping(em_tree, em); write_unlock(&em_tree->lock); @@ -5704,8 +5725,7 @@ must_cow: * it above */ len = bh_result->b_size; - free_extent_map(em); - em = btrfs_new_extent_direct(inode, start, len); + em = btrfs_new_extent_direct(inode, em, start, len); if (IS_ERR(em)) return PTR_ERR(em); len = min(len, em->len - (start - em->start)); -- cgit v1.2.2 From 93a54bc4c28a125978cddbe2db9e347391e3522d Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 6 Apr 2011 15:11:44 -0400 Subject: Btrfs: check for duplicate iov_base's when doing dio reads Apparently it is ok to submit a read to an IDE device with the same target page for different offsets. This is what Windows does under qemu. The problem is under DIO we expect them to be different buffers for checksumming reasons, and so this sort of thing will result in checksum errors, when in reality the file is fine. So when reading, check to make sure that all iov bases are different, and if they aren't fall back to buffered mode, since that will work out right. Thanks, Signed-off-by: Josef Bacik --- fs/btrfs/inode.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 24310c9cb14f..00d59c6a9769 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6207,6 +6207,7 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io unsigned long nr_segs) { int seg; + int i; size_t size; unsigned long addr; unsigned blocksize_mask = root->sectorsize - 1; @@ -6221,8 +6222,22 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io addr = (unsigned long)iov[seg].iov_base; size = iov[seg].iov_len; end += size; - if ((addr & blocksize_mask) || (size & blocksize_mask)) + if ((addr & blocksize_mask) || (size & blocksize_mask)) goto out; + + /* If this is a write we don't need to check anymore */ + if (rw & WRITE) + continue; + + /* + * Check to make sure we don't have duplicate iov_base's in this + * iovec, if so return EINVAL, otherwise we'll get csum errors + * when reading back. + */ + for (i = seg + 1; i < nr_segs; i++) { + if (iov[seg].iov_base == iov[i].iov_base) + goto out; + } } retval = 0; out: -- cgit v1.2.2 From 8fb27640d0e2b43c5584bf0087431b7b8d3c319a Mon Sep 17 00:00:00 2001 From: Yoshinori Sano Date: Sat, 9 Apr 2011 02:30:07 +0000 Subject: Btrfs: fix memory leaks in btrfs_new_inode() This patch fixes memory leaks in btrfs_new_inode(). Signed-off-by: Yoshinori Sano Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index cc6022842e0c..2d1208f964eb 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4526,14 +4526,17 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans, BUG_ON(!path); inode = new_inode(root->fs_info->sb); - if (!inode) + if (!inode) { + btrfs_free_path(path); return ERR_PTR(-ENOMEM); + } if (dir) { trace_btrfs_inode_request(dir); ret = btrfs_set_inode_index(dir, index); if (ret) { + btrfs_free_path(path); iput(inode); return ERR_PTR(ret); } -- cgit v1.2.2 From a1b75f7d961955e697ec377f90115e3517df98f9 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Fri, 8 Apr 2011 15:51:18 +0000 Subject: Btrfs: check for duplicate iov_base's when doing dio reads Apparently it is ok to submit a read to an IDE device with the same target page for different offsets. This is what Windows does under qemu. The problem is under DIO we expect them to be different buffers for checksumming reasons, and so this sort of thing will result in checksum errors, when in reality the file is fine. So when reading, check to make sure that all iov bases are different, and if they aren't fall back to buffered mode, since that will work out right. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 2d1208f964eb..edafc28883af 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6153,6 +6153,7 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io unsigned long nr_segs) { int seg; + int i; size_t size; unsigned long addr; unsigned blocksize_mask = root->sectorsize - 1; @@ -6167,8 +6168,22 @@ static ssize_t check_direct_IO(struct btrfs_root *root, int rw, struct kiocb *io addr = (unsigned long)iov[seg].iov_base; size = iov[seg].iov_len; end += size; - if ((addr & blocksize_mask) || (size & blocksize_mask)) + if ((addr & blocksize_mask) || (size & blocksize_mask)) goto out; + + /* If this is a write we don't need to check anymore */ + if (rw & WRITE) + continue; + + /* + * Check to make sure we don't have duplicate iov_base's in this + * iovec, if so return EINVAL, otherwise we'll get csum errors + * when reading back. + */ + for (i = seg + 1; i < nr_segs; i++) { + if (iov[seg].iov_base == iov[i].iov_base) + goto out; + } } retval = 0; out: -- cgit v1.2.2 From 507903b81840a70cc6a179d4eb03584ad50e8c5b Mon Sep 17 00:00:00 2001 From: Arne Jansen Date: Wed, 6 Apr 2011 10:02:20 +0000 Subject: btrfs: using cached extent_state in set/unlock combinations In several places the sequence (set_extent_uptodate, unlock_extent) is used. This leads to a duplicate lookup of the extent state. This patch lets set_extent_uptodate return a cached extent_state which can be passed to unlock_extent_cached. The occurences of the above sequences are updated to use the cache. Only end_bio_extent_readpage is updated that it first gets a cached state to pass it to the readpage_end_io_hook as the prototype requested and is later on being used for set/unlock. Signed-off-by: Arne Jansen Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index edafc28883af..5a993e0ec865 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5226,7 +5226,7 @@ again: btrfs_mark_buffer_dirty(leaf); } set_extent_uptodate(io_tree, em->start, - extent_map_end(em) - 1, GFP_NOFS); + extent_map_end(em) - 1, NULL, GFP_NOFS); goto insert; } else { printk(KERN_ERR "btrfs unknown found_type %d\n", found_type); -- cgit v1.2.2 From b9e03af0bcc11310f6be4a3951c9ee2c26465011 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Wed, 23 Mar 2011 10:43:58 +0800 Subject: Btrfs: Check if btrfs_next_leaf() returns error in btrfs_real_readdir() btrfs_next_leaf() can return -errno, and we should propagate it to userspace. This also simplifies how we walk the btree path. Signed-off-by: Li Zefan --- fs/btrfs/inode.c | 28 ++++++++++------------------ 1 file changed, 10 insertions(+), 18 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 55a6a0b416d7..b9f7f5258343 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4221,10 +4221,8 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, struct btrfs_key found_key; struct btrfs_path *path; int ret; - u32 nritems; struct extent_buffer *leaf; int slot; - int advance; unsigned char d_type; int over = 0; u32 di_cur; @@ -4267,27 +4265,19 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, ret = btrfs_search_slot(NULL, root, &key, path, 0, 0); if (ret < 0) goto err; - advance = 0; while (1) { leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); slot = path->slots[0]; - if (advance || slot >= nritems) { - if (slot >= nritems - 1) { - ret = btrfs_next_leaf(root, path); - if (ret) - break; - leaf = path->nodes[0]; - nritems = btrfs_header_nritems(leaf); - slot = path->slots[0]; - } else { - slot++; - path->slots[0]++; - } + if (slot >= btrfs_header_nritems(leaf)) { + ret = btrfs_next_leaf(root, path); + if (ret < 0) + goto err; + else if (ret > 0) + break; + continue; } - advance = 1; item = btrfs_item_nr(leaf, slot); btrfs_item_key_to_cpu(leaf, &found_key, slot); @@ -4296,7 +4286,7 @@ static int btrfs_real_readdir(struct file *filp, void *dirent, if (btrfs_key_type(&found_key) != key_type) break; if (found_key.offset < filp->f_pos) - continue; + goto next; filp->f_pos = found_key.offset; @@ -4349,6 +4339,8 @@ skip: di_cur += di_len; di = (struct btrfs_dir_item *)((char *)di + di_len); } +next: + path->slots[0]++; } /* Reached end of directory/root. Bump pos past the last item. */ -- cgit v1.2.2 From 3153495d8ed6a9bb9f00aea42c18dc488a885dd6 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Wed, 13 Apr 2011 13:19:21 +0800 Subject: Btrfs: Fix incorrect inode nlink in btrfs_link() Link count of the inode is not decreased if btrfs_set_inode_index() fails. Signed-off-by: Miao Xie Singed-off-by: Li Zefan --- fs/btrfs/inode.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index b9f7f5258343..a4157cfdd533 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4846,9 +4846,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, if (inode->i_nlink == ~0U) return -EMLINK; - btrfs_inc_nlink(inode); - inode->i_ctime = CURRENT_TIME; - err = btrfs_set_inode_index(dir, &index); if (err) goto fail; @@ -4864,6 +4861,9 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, goto fail; } + btrfs_inc_nlink(inode); + inode->i_ctime = CURRENT_TIME; + btrfs_set_trans_block_group(trans, dir); ihold(inode); -- cgit v1.2.2 From 8d413713ca744fa00cf4e05d4054d80727b84789 Mon Sep 17 00:00:00 2001 From: Tsutomu Itoh Date: Mon, 25 Apr 2011 19:43:52 -0400 Subject: Btrfs: check return value of kmalloc() The check on the return value of kmalloc() is added to some places. Signed-off-by: Tsutomu Itoh Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index a4157cfdd533..c718d274a352 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -953,6 +953,7 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page, 1, 0, NULL, GFP_NOFS); while (start < end) { async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS); + BUG_ON(!async_cow); async_cow->inode = inode; async_cow->root = root; async_cow->locked_page = locked_page; @@ -5001,6 +5002,8 @@ static noinline int uncompress_inline(struct btrfs_path *path, inline_size = btrfs_file_extent_inline_item_len(leaf, btrfs_item_nr(leaf, path->slots[0])); tmp = kmalloc(inline_size, GFP_NOFS); + if (!tmp) + return -ENOMEM; ptr = btrfs_file_extent_inline_start(item); read_extent_buffer(leaf, tmp, ptr, inline_size); -- cgit v1.2.2 From 64728bbbf892ea7a4aba502c436afbe362217fb9 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 25 Apr 2011 19:43:52 -0400 Subject: Btrfs: put the right bio if we have an error In btrfs_submit_direct_hook if the first btrfs_map_block fails we need to put the orig_bio, not bio. Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index c718d274a352..ad6b515173ac 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -6041,7 +6041,7 @@ static int btrfs_submit_direct_hook(int rw, struct btrfs_dio_private *dip, ret = btrfs_map_block(map_tree, READ, start_sector << 9, &map_length, NULL, 0); if (ret) { - bio_put(bio); + bio_put(orig_bio); return -EIO; } -- cgit v1.2.2 From 7cf96da3ec7ca225acf4f284b0e904a1f5f98821 Mon Sep 17 00:00:00 2001 From: Tsutomu Itoh Date: Mon, 25 Apr 2011 19:43:53 -0400 Subject: Btrfs: cleanup error handling in inode.c The error processing of several places is changed like setting the error number only at the error. Signed-off-by: Tsutomu Itoh Signed-off-by: Chris Mason --- fs/btrfs/inode.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) (limited to 'fs/btrfs/inode.c') diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index ad6b515173ac..870869aab0b8 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -4731,9 +4731,10 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, dentry->d_name.len, dir->i_ino, objectid, BTRFS_I(dir)->block_group, mode, &index); - err = PTR_ERR(inode); - if (IS_ERR(inode)) + if (IS_ERR(inode)) { + err = PTR_ERR(inode); goto out_unlock; + } err = btrfs_init_inode_security(trans, inode, dir); if (err) { @@ -4792,9 +4793,10 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, inode = btrfs_new_inode(trans, root, dir, dentry->d_name.name, dentry->d_name.len, dir->i_ino, objectid, BTRFS_I(dir)->block_group, mode, &index); - err = PTR_ERR(inode); - if (IS_ERR(inode)) + if (IS_ERR(inode)) { + err = PTR_ERR(inode); goto out_unlock; + } err = btrfs_init_inode_security(trans, inode, dir); if (err) { @@ -7278,9 +7280,10 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, dentry->d_name.len, dir->i_ino, objectid, BTRFS_I(dir)->block_group, S_IFLNK|S_IRWXUGO, &index); - err = PTR_ERR(inode); - if (IS_ERR(inode)) + if (IS_ERR(inode)) { + err = PTR_ERR(inode); goto out_unlock; + } err = btrfs_init_inode_security(trans, inode, dir); if (err) { -- cgit v1.2.2