From 8ccf6f19b67f7e0921063cc309f4672a6afcb528 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Thu, 25 Oct 2012 09:28:04 +0000 Subject: Btrfs: make delalloc inodes be flushed by multi-task This patch introduce a new worker pool named "flush_workers", and if we want to force all the inode with pending delalloc to the disks, we can queue those inodes into the work queue of the worker pool, in this way, those inodes will be flushed by multi-task. Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 7cda51995c1..bd70c2852ba 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2279,6 +2279,10 @@ int open_ctree(struct super_block *sb, fs_info->thread_pool_size, &fs_info->generic_worker); + btrfs_init_workers(&fs_info->flush_workers, "flush_delalloc", + fs_info->thread_pool_size, + &fs_info->generic_worker); + btrfs_init_workers(&fs_info->submit_workers, "submit", min_t(u64, fs_devices->num_devices, fs_info->thread_pool_size), @@ -2350,6 +2354,7 @@ int open_ctree(struct super_block *sb, ret |= btrfs_start_workers(&fs_info->delayed_workers); ret |= btrfs_start_workers(&fs_info->caching_workers); ret |= btrfs_start_workers(&fs_info->readahead_workers); + ret |= btrfs_start_workers(&fs_info->flush_workers); if (ret) { err = -ENOMEM; goto fail_sb_buffer; @@ -2667,6 +2672,7 @@ fail_sb_buffer: btrfs_stop_workers(&fs_info->submit_workers); btrfs_stop_workers(&fs_info->delayed_workers); btrfs_stop_workers(&fs_info->caching_workers); + btrfs_stop_workers(&fs_info->flush_workers); fail_alloc: fail_iput: btrfs_mapping_tree_free(&fs_info->mapping_tree); @@ -3339,6 +3345,7 @@ int close_ctree(struct btrfs_root *root) btrfs_stop_workers(&fs_info->delayed_workers); btrfs_stop_workers(&fs_info->caching_workers); btrfs_stop_workers(&fs_info->readahead_workers); + btrfs_stop_workers(&fs_info->flush_workers); #ifdef CONFIG_BTRFS_FS_CHECK_INTEGRITY if (btrfs_test_opt(root, CHECK_INTEGRITY)) -- cgit v1.2.3-70-g09d2 From d0e1d66b5aa1ec9f556f951aa9a114cc192cd01c Mon Sep 17 00:00:00 2001 From: Namjae Jeon Date: Tue, 11 Dec 2012 16:00:21 -0800 Subject: writeback: remove nr_pages_dirtied arg from balance_dirty_pages_ratelimited_nr() There is no reason to pass the nr_pages_dirtied argument, because nr_pages_dirtied value from the caller is unused in balance_dirty_pages_ratelimited_nr(). Signed-off-by: Namjae Jeon Signed-off-by: Vivek Trivedi Cc: Wu Fengguang Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/btrfs/disk-io.c | 8 ++++---- fs/btrfs/file.c | 3 +-- fs/btrfs/ioctl.c | 2 +- fs/ocfs2/file.c | 5 +---- fs/splice.c | 5 +---- include/linux/writeback.h | 9 +-------- mm/page-writeback.c | 11 +++++------ 7 files changed, 14 insertions(+), 29 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 7cda51995c1..22a0439e5a8 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3416,8 +3416,8 @@ void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) num_dirty = root->fs_info->dirty_metadata_bytes; if (num_dirty > thresh) { - balance_dirty_pages_ratelimited_nr( - root->fs_info->btree_inode->i_mapping, 1); + balance_dirty_pages_ratelimited( + root->fs_info->btree_inode->i_mapping); } return; } @@ -3437,8 +3437,8 @@ void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) num_dirty = root->fs_info->dirty_metadata_bytes; if (num_dirty > thresh) { - balance_dirty_pages_ratelimited_nr( - root->fs_info->btree_inode->i_mapping, 1); + balance_dirty_pages_ratelimited( + root->fs_info->btree_inode->i_mapping); } return; } diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 9ab1bed8811..a8ee75cb96e 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1346,8 +1346,7 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, cond_resched(); - balance_dirty_pages_ratelimited_nr(inode->i_mapping, - dirty_pages); + balance_dirty_pages_ratelimited(inode->i_mapping); if (dirty_pages < (root->leafsize >> PAGE_CACHE_SHIFT) + 1) btrfs_btree_balance_dirty(root, 1); diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index 8fcf9a59c28..5b3429ab8ec 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -1225,7 +1225,7 @@ int btrfs_defrag_file(struct inode *inode, struct file *file, } defrag_count += ret; - balance_dirty_pages_ratelimited_nr(inode->i_mapping, ret); + balance_dirty_pages_ratelimited(inode->i_mapping); mutex_unlock(&inode->i_mutex); if (newer_than) { diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 5a4ee77cec5..dda08980494 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2513,18 +2513,15 @@ static ssize_t ocfs2_file_splice_write(struct pipe_inode_info *pipe, ret = sd.num_spliced; if (ret > 0) { - unsigned long nr_pages; int err; - nr_pages = (ret + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; - err = generic_write_sync(out, *ppos, ret); if (err) ret = err; else *ppos += ret; - balance_dirty_pages_ratelimited_nr(mapping, nr_pages); + balance_dirty_pages_ratelimited(mapping); } return ret; diff --git a/fs/splice.c b/fs/splice.c index 13e5b4776e7..8890604e3fc 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -1024,17 +1024,14 @@ generic_file_splice_write(struct pipe_inode_info *pipe, struct file *out, ret = sd.num_spliced; if (ret > 0) { - unsigned long nr_pages; int err; - nr_pages = (ret + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; - err = generic_write_sync(out, *ppos, ret); if (err) ret = err; else *ppos += ret; - balance_dirty_pages_ratelimited_nr(mapping, nr_pages); + balance_dirty_pages_ratelimited(mapping); } sb_end_write(inode->i_sb); diff --git a/include/linux/writeback.h b/include/linux/writeback.h index 50c3e8fa06a..b82a83aba31 100644 --- a/include/linux/writeback.h +++ b/include/linux/writeback.h @@ -161,14 +161,7 @@ void __bdi_update_bandwidth(struct backing_dev_info *bdi, unsigned long start_time); void page_writeback_init(void); -void balance_dirty_pages_ratelimited_nr(struct address_space *mapping, - unsigned long nr_pages_dirtied); - -static inline void -balance_dirty_pages_ratelimited(struct address_space *mapping) -{ - balance_dirty_pages_ratelimited_nr(mapping, 1); -} +void balance_dirty_pages_ratelimited(struct address_space *mapping); typedef int (*writepage_t)(struct page *page, struct writeback_control *wbc, void *data); diff --git a/mm/page-writeback.c b/mm/page-writeback.c index 830893b2b3c..6f427122449 100644 --- a/mm/page-writeback.c +++ b/mm/page-writeback.c @@ -1069,7 +1069,7 @@ static void bdi_update_bandwidth(struct backing_dev_info *bdi, } /* - * After a task dirtied this many pages, balance_dirty_pages_ratelimited_nr() + * After a task dirtied this many pages, balance_dirty_pages_ratelimited() * will look to see if it needs to start dirty throttling. * * If dirty_poll_interval is too low, big NUMA machines will call the expensive @@ -1436,9 +1436,8 @@ static DEFINE_PER_CPU(int, bdp_ratelimits); DEFINE_PER_CPU(int, dirty_throttle_leaks) = 0; /** - * balance_dirty_pages_ratelimited_nr - balance dirty memory state + * balance_dirty_pages_ratelimited - balance dirty memory state * @mapping: address_space which was dirtied - * @nr_pages_dirtied: number of pages which the caller has just dirtied * * Processes which are dirtying memory should call in here once for each page * which was newly dirtied. The function will periodically check the system's @@ -1449,8 +1448,7 @@ DEFINE_PER_CPU(int, dirty_throttle_leaks) = 0; * limit we decrease the ratelimiting by a lot, to prevent individual processes * from overshooting the limit by (ratelimit_pages) each. */ -void balance_dirty_pages_ratelimited_nr(struct address_space *mapping, - unsigned long nr_pages_dirtied) +void balance_dirty_pages_ratelimited(struct address_space *mapping) { struct backing_dev_info *bdi = mapping->backing_dev_info; int ratelimit; @@ -1484,6 +1482,7 @@ void balance_dirty_pages_ratelimited_nr(struct address_space *mapping, */ p = &__get_cpu_var(dirty_throttle_leaks); if (*p > 0 && current->nr_dirtied < ratelimit) { + unsigned long nr_pages_dirtied; nr_pages_dirtied = min(*p, ratelimit - current->nr_dirtied); *p -= nr_pages_dirtied; current->nr_dirtied += nr_pages_dirtied; @@ -1493,7 +1492,7 @@ void balance_dirty_pages_ratelimited_nr(struct address_space *mapping, if (unlikely(current->nr_dirtied >= ratelimit)) balance_dirty_pages(mapping, current->nr_dirtied); } -EXPORT_SYMBOL(balance_dirty_pages_ratelimited_nr); +EXPORT_SYMBOL(balance_dirty_pages_ratelimited); void throttle_vm_writeout(gfp_t gfp_mask) { -- cgit v1.2.3-70-g09d2 From 292fd7fc39aa06668f3a8db546714e727120cb3e Mon Sep 17 00:00:00 2001 From: Stefan Behrens Date: Tue, 30 Oct 2012 17:16:16 +0000 Subject: Btrfs: don't allow degraded mount if too many devices are missing The current behavior is to allow mounting or remounting a filesystem writeable in degraded mode if at least one writeable device is present. The next failed write access to a missing device which is above the tolerance of the configured level of redundancy results in an read-only enforcement. Even without this, the next time barrier_all_devices() is called and more devices are missing than tolerable, the switch to read-only mode takes place. In order to behave predictably and to provide proper feedback to the user at mount time, this patch compares the number of missing devices with the number of devices that are tolerated to be missing according to the configured RAID level. If more devices are missing than tolerated, e.g. if two devices are missing in case of RAID1, only a read-only mount and remount is allowed. Signed-off-by: Stefan Behrens Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 7 +++++++ fs/btrfs/super.c | 9 +++++++++ 2 files changed, 16 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index bd70c2852ba..064315990f8 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2508,6 +2508,13 @@ retry_root_backup: } fs_info->num_tolerated_disk_barrier_failures = btrfs_calc_num_tolerated_disk_barrier_failures(fs_info); + if (fs_info->fs_devices->missing_devices > + fs_info->num_tolerated_disk_barrier_failures && + !(sb->s_flags & MS_RDONLY)) { + printk(KERN_WARNING + "Btrfs: too many missing devices, writeable mount is not allowed\n"); + goto fail_block_groups; + } fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root, "btrfs-cleaner"); diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index 915ac14c206..acd2df85bed 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -1226,6 +1226,15 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data) goto restore; } + if (fs_info->fs_devices->missing_devices > + fs_info->num_tolerated_disk_barrier_failures && + !(*flags & MS_RDONLY)) { + printk(KERN_WARNING + "Btrfs: too many missing devices, writeable remount is not allowed\n"); + ret = -EACCES; + goto restore; + } + if (btrfs_super_log_root(fs_info->super_copy) != 0) { ret = -EINVAL; goto restore; -- cgit v1.2.3-70-g09d2 From 31b1a2bd758f439fc945b3ac5899d890cb7e2dc6 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Sat, 3 Nov 2012 10:58:34 +0000 Subject: fs/btrfs: use WARN Use WARN rather than printk followed by WARN_ON(1), for conciseness. A simplified version of the semantic patch that makes this transformation is as follows: (http://coccinelle.lip6.fr/) // @@ expression list es; @@ -printk( +WARN(1, es); -WARN_ON(1); // Signed-off-by: Julia Lawall Reviewed-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/ctree.c | 19 +++++++------------ fs/btrfs/disk-io.c | 6 ++---- fs/btrfs/extent-tree.c | 7 +++---- fs/btrfs/extent_io.c | 9 +++------ fs/btrfs/inode.c | 3 +-- fs/btrfs/transaction.c | 12 ++++-------- fs/btrfs/volumes.c | 3 +-- 7 files changed, 21 insertions(+), 38 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 100c274a1cf..0e4adb00e9d 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -1359,19 +1359,16 @@ noinline int btrfs_cow_block(struct btrfs_trans_handle *trans, u64 search_start; int ret; - if (trans->transaction != root->fs_info->running_transaction) { - printk(KERN_CRIT "trans %llu running %llu\n", + if (trans->transaction != root->fs_info->running_transaction) + WARN(1, KERN_CRIT "trans %llu running %llu\n", (unsigned long long)trans->transid, (unsigned long long) root->fs_info->running_transaction->transid); - WARN_ON(1); - } - if (trans->transid != root->fs_info->generation) { - printk(KERN_CRIT "trans %llu running %llu\n", + + if (trans->transid != root->fs_info->generation) + WARN(1, KERN_CRIT "trans %llu running %llu\n", (unsigned long long)trans->transid, (unsigned long long)root->fs_info->generation); - WARN_ON(1); - } if (!should_cow_block(trans, root, buf)) { *cow_ret = buf; @@ -3640,11 +3637,9 @@ static noinline int __push_leaf_left(struct btrfs_trans_handle *trans, btrfs_set_header_nritems(left, old_left_nritems + push_items); /* fixup right node */ - if (push_items > right_nritems) { - printk(KERN_CRIT "push items %d nr %u\n", push_items, + if (push_items > right_nritems) + WARN(1, KERN_CRIT "push items %d nr %u\n", push_items, right_nritems); - WARN_ON(1); - } if (push_items < right_nritems) { push_space = btrfs_item_offset_nr(right, push_items - 1) - diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 064315990f8..07a2162cdd6 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3397,14 +3397,12 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf) int was_dirty; btrfs_assert_tree_locked(buf); - if (transid != root->fs_info->generation) { - printk(KERN_CRIT "btrfs transid mismatch buffer %llu, " + if (transid != root->fs_info->generation) + WARN(1, KERN_CRIT "btrfs transid mismatch buffer %llu, " "found %llu running %llu\n", (unsigned long long)buf->start, (unsigned long long)transid, (unsigned long long)root->fs_info->generation); - WARN_ON(1); - } was_dirty = set_extent_buffer_dirty(buf); if (!was_dirty) { spin_lock(&root->fs_info->delalloc_lock); diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c index b495cb4b9b2..0bcb9543da6 100644 --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@ -6292,10 +6292,9 @@ use_block_rsv(struct btrfs_trans_handle *trans, static DEFINE_RATELIMIT_STATE(_rs, DEFAULT_RATELIMIT_INTERVAL, /*DEFAULT_RATELIMIT_BURST*/ 2); - if (__ratelimit(&_rs)) { - printk(KERN_DEBUG "btrfs: block rsv returned %d\n", ret); - WARN_ON(1); - } + if (__ratelimit(&_rs)) + WARN(1, KERN_DEBUG "btrfs: block rsv returned %d\n", + ret); ret = reserve_metadata_bytes(root, block_rsv, blocksize, BTRFS_RESERVE_NO_FLUSH); if (!ret) { diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 472873a94d9..3c062c8d1d7 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -341,12 +341,10 @@ static int insert_state(struct extent_io_tree *tree, { struct rb_node *node; - if (end < start) { - printk(KERN_ERR "btrfs end < start %llu %llu\n", + if (end < start) + WARN(1, KERN_ERR "btrfs end < start %llu %llu\n", (unsigned long long)end, (unsigned long long)start); - WARN_ON(1); - } state->start = start; state->end = end; @@ -4721,10 +4719,9 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start, } if (start + min_len > eb->len) { - printk(KERN_ERR "btrfs bad mapping eb start %llu len %lu, " + WARN(1, KERN_ERR "btrfs bad mapping eb start %llu len %lu, " "wanted %lu %lu\n", (unsigned long long)eb->start, eb->len, start, min_len); - WARN_ON(1); return -EINVAL; } diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 96d20903bee..6dca345fd1b 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -5458,8 +5458,7 @@ again: extent_map_end(em) - 1, NULL, GFP_NOFS); goto insert; } else { - printk(KERN_ERR "btrfs unknown found_type %d\n", found_type); - WARN_ON(1); + WARN(1, KERN_ERR "btrfs unknown found_type %d\n", found_type); } not_found: em->start = start; diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index bc1f5239733..f21f39f0b1a 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -145,16 +145,12 @@ loop: * the log must never go across transaction boundaries. */ smp_mb(); - if (!list_empty(&fs_info->tree_mod_seq_list)) { - printk(KERN_ERR "btrfs: tree_mod_seq_list not empty when " + if (!list_empty(&fs_info->tree_mod_seq_list)) + WARN(1, KERN_ERR "btrfs: tree_mod_seq_list not empty when " "creating a fresh transaction\n"); - WARN_ON(1); - } - if (!RB_EMPTY_ROOT(&fs_info->tree_mod_log)) { - printk(KERN_ERR "btrfs: tree_mod_log rb tree not empty when " + if (!RB_EMPTY_ROOT(&fs_info->tree_mod_log)) + WARN(1, KERN_ERR "btrfs: tree_mod_log rb tree not empty when " "creating a fresh transaction\n"); - WARN_ON(1); - } atomic_set(&fs_info->tree_mod_seq, 0); spin_lock_init(&cur_trans->commit_lock); diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index eeed97d19de..3f4bfee66d7 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3323,9 +3323,8 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, cur = cur->next; if (!device->writeable) { - printk(KERN_ERR + WARN(1, KERN_ERR "btrfs: read-only device in alloc_list\n"); - WARN_ON(1); continue; } -- cgit v1.2.3-70-g09d2 From b53d3f5db2b79637acadc06a330db6c2c60863f5 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 14 Nov 2012 14:34:34 +0000 Subject: Btrfs: cleanup for btrfs_btree_balance_dirty - 'nr' is no more used. - btrfs_btree_balance_dirty() and __btrfs_btree_balance_dirty() can share a bunch of code. Signed-off-by: Liu Bo Signed-off-by: Chris Mason --- fs/btrfs/delayed-inode.c | 5 +---- fs/btrfs/disk-io.c | 29 ++++++++++------------------- fs/btrfs/disk-io.h | 4 ++-- fs/btrfs/file.c | 9 +++------ fs/btrfs/inode.c | 42 +++++++++++------------------------------- fs/btrfs/relocation.c | 22 ++++++---------------- fs/btrfs/transaction.c | 4 +--- 7 files changed, 34 insertions(+), 81 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 0c6dca550ea..34836036f01 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -1257,7 +1257,6 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work) struct btrfs_delayed_node *delayed_node = NULL; struct btrfs_root *root; struct btrfs_block_rsv *block_rsv; - unsigned long nr = 0; int need_requeue = 0; int ret; @@ -1318,11 +1317,9 @@ static void btrfs_async_run_delayed_node_done(struct btrfs_work *work) delayed_node); mutex_unlock(&delayed_node->mutex); - nr = trans->blocks_used; - trans->block_rsv = block_rsv; btrfs_end_transaction_dmeta(trans, root); - __btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty_nodelay(root); free_path: btrfs_free_path(path); out: diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 07a2162cdd6..ff5d259ac27 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3411,7 +3411,8 @@ void btrfs_mark_buffer_dirty(struct extent_buffer *buf) } } -void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) +static void __btrfs_btree_balance_dirty(struct btrfs_root *root, + int flush_delayed) { /* * looks as though older kernels can get into trouble with @@ -3423,7 +3424,8 @@ void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) if (current->flags & PF_MEMALLOC) return; - btrfs_balance_delayed_items(root); + if (flush_delayed) + btrfs_balance_delayed_items(root); num_dirty = root->fs_info->dirty_metadata_bytes; @@ -3434,25 +3436,14 @@ void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) return; } -void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr) +void btrfs_btree_balance_dirty(struct btrfs_root *root) { - /* - * looks as though older kernels can get into trouble with - * this code, they end up stuck in balance_dirty_pages forever - */ - u64 num_dirty; - unsigned long thresh = 32 * 1024 * 1024; - - if (current->flags & PF_MEMALLOC) - return; - - num_dirty = root->fs_info->dirty_metadata_bytes; + __btrfs_btree_balance_dirty(root, 1); +} - if (num_dirty > thresh) { - balance_dirty_pages_ratelimited_nr( - root->fs_info->btree_inode->i_mapping, 1); - } - return; +void btrfs_btree_balance_dirty_nodelay(struct btrfs_root *root) +{ + __btrfs_btree_balance_dirty(root, 0); } int btrfs_read_buffer(struct extent_buffer *buf, u64 parent_transid) diff --git a/fs/btrfs/disk-io.h b/fs/btrfs/disk-io.h index 2025a9132c1..305c33efb0e 100644 --- a/fs/btrfs/disk-io.h +++ b/fs/btrfs/disk-io.h @@ -62,8 +62,8 @@ struct btrfs_root *btrfs_read_fs_root_no_radix(struct btrfs_root *tree_root, struct btrfs_root *btrfs_read_fs_root_no_name(struct btrfs_fs_info *fs_info, struct btrfs_key *location); int btrfs_cleanup_fs_roots(struct btrfs_fs_info *fs_info); -void btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr); -void __btrfs_btree_balance_dirty(struct btrfs_root *root, unsigned long nr); +void btrfs_btree_balance_dirty(struct btrfs_root *root); +void btrfs_btree_balance_dirty_nodelay(struct btrfs_root *root); void btrfs_free_fs_root(struct btrfs_fs_info *fs_info, struct btrfs_root *root); void btrfs_mark_buffer_dirty(struct extent_buffer *buf); int btrfs_buffer_uptodate(struct extent_buffer *buf, u64 parent_transid, diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 883cf826cf2..bd7f1b01e05 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -1349,7 +1349,7 @@ static noinline ssize_t __btrfs_buffered_write(struct file *file, balance_dirty_pages_ratelimited_nr(inode->i_mapping, dirty_pages); if (dirty_pages < (root->leafsize >> PAGE_CACHE_SHIFT) + 1) - btrfs_btree_balance_dirty(root, 1); + btrfs_btree_balance_dirty(root); pos += copied; num_written += copied; @@ -1803,7 +1803,6 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) u64 cur_offset = lockstart; u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); u64 drop_end; - unsigned long nr; int ret = 0; int err = 0; bool same_page = (offset >> PAGE_CACHE_SHIFT) == @@ -1931,9 +1930,8 @@ static int btrfs_punch_hole(struct inode *inode, loff_t offset, loff_t len) break; } - nr = trans->blocks_used; btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); trans = btrfs_start_transaction(root, 3); if (IS_ERR(trans)) { @@ -1969,9 +1967,8 @@ out_trans: trans->block_rsv = &root->fs_info->trans_block_rsv; ret = btrfs_update_inode(trans, root, inode); - nr = trans->blocks_used; btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); out_free: btrfs_free_path(path); btrfs_free_block_rsv(root, rsv); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index e8733fab273..aabf747d056 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -3091,7 +3091,6 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry) struct btrfs_trans_handle *trans; struct inode *inode = dentry->d_inode; int ret; - unsigned long nr = 0; trans = __unlink_start_trans(dir, dentry); if (IS_ERR(trans)) @@ -3111,9 +3110,8 @@ static int btrfs_unlink(struct inode *dir, struct dentry *dentry) } out: - nr = trans->blocks_used; __unlink_end_trans(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); return ret; } @@ -3203,7 +3201,6 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) int err = 0; struct btrfs_root *root = BTRFS_I(dir)->root; struct btrfs_trans_handle *trans; - unsigned long nr = 0; if (inode->i_size > BTRFS_EMPTY_DIR_SIZE) return -ENOTEMPTY; @@ -3232,9 +3229,8 @@ static int btrfs_rmdir(struct inode *dir, struct dentry *dentry) if (!err) btrfs_i_size_write(inode, 0); out: - nr = trans->blocks_used; __unlink_end_trans(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); return err; } @@ -3800,7 +3796,6 @@ void btrfs_evict_inode(struct inode *inode) struct btrfs_root *root = BTRFS_I(inode)->root; struct btrfs_block_rsv *rsv, *global_rsv; u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); - unsigned long nr; int ret; trace_btrfs_inode_evict(inode); @@ -3882,10 +3877,9 @@ void btrfs_evict_inode(struct inode *inode) ret = btrfs_update_inode(trans, root, inode); BUG_ON(ret); - nr = trans->blocks_used; btrfs_end_transaction(trans, root); trans = NULL; - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); } btrfs_free_block_rsv(root, rsv); @@ -3901,9 +3895,8 @@ void btrfs_evict_inode(struct inode *inode) root->root_key.objectid == BTRFS_TREE_RELOC_OBJECTID)) btrfs_return_ino(root, btrfs_ino(inode)); - nr = trans->blocks_used; btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); no_delete: clear_inode(inode); return; @@ -4915,7 +4908,6 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, int err; int drop_inode = 0; u64 objectid; - unsigned long nr = 0; u64 index = 0; if (!new_valid_dev(rdev)) @@ -4965,9 +4957,8 @@ static int btrfs_mknod(struct inode *dir, struct dentry *dentry, d_instantiate(dentry, inode); } out_unlock: - nr = trans->blocks_used; btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); if (drop_inode) { inode_dec_link_count(inode); iput(inode); @@ -4983,7 +4974,6 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, struct inode *inode = NULL; int drop_inode = 0; int err; - unsigned long nr = 0; u64 objectid; u64 index = 0; @@ -5033,13 +5023,12 @@ static int btrfs_create(struct inode *dir, struct dentry *dentry, d_instantiate(dentry, inode); } out_unlock: - nr = trans->blocks_used; btrfs_end_transaction(trans, root); if (drop_inode) { inode_dec_link_count(inode); iput(inode); } - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); return err; } @@ -5050,7 +5039,6 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, struct btrfs_root *root = BTRFS_I(dir)->root; struct inode *inode = old_dentry->d_inode; u64 index; - unsigned long nr = 0; int err; int drop_inode = 0; @@ -5094,14 +5082,13 @@ static int btrfs_link(struct dentry *old_dentry, struct inode *dir, btrfs_log_new_name(trans, inode, NULL, parent); } - nr = trans->blocks_used; btrfs_end_transaction(trans, root); fail: if (drop_inode) { inode_dec_link_count(inode); iput(inode); } - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); return err; } @@ -5114,7 +5101,6 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) int drop_on_err = 0; u64 objectid = 0; u64 index = 0; - unsigned long nr = 1; /* * 2 items for inode and ref @@ -5160,11 +5146,10 @@ static int btrfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) drop_on_err = 0; out_fail: - nr = trans->blocks_used; btrfs_end_transaction(trans, root); if (drop_on_err) iput(inode); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); return err; } @@ -6872,7 +6857,6 @@ static int btrfs_truncate(struct inode *inode) int ret; int err = 0; struct btrfs_trans_handle *trans; - unsigned long nr; u64 mask = root->sectorsize - 1; u64 min_size = btrfs_calc_trunc_metadata_size(root, 1); @@ -6995,9 +6979,8 @@ static int btrfs_truncate(struct inode *inode) break; } - nr = trans->blocks_used; btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); trans = btrfs_start_transaction(root, 2); if (IS_ERR(trans)) { @@ -7031,9 +7014,8 @@ static int btrfs_truncate(struct inode *inode) if (ret && !err) err = ret; - nr = trans->blocks_used; ret = btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); } out: @@ -7594,7 +7576,6 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, unsigned long ptr; struct btrfs_file_extent_item *ei; struct extent_buffer *leaf; - unsigned long nr = 0; name_len = strlen(symname) + 1; if (name_len > BTRFS_MAX_INLINE_DATA_SIZE(root)) @@ -7692,13 +7673,12 @@ static int btrfs_symlink(struct inode *dir, struct dentry *dentry, out_unlock: if (!err) d_instantiate(dentry, inode); - nr = trans->blocks_used; btrfs_end_transaction(trans, root); if (drop_inode) { inode_dec_link_count(inode); iput(inode); } - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); return err; } diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 270f24ffe1b..300e09ac365 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -2025,7 +2025,6 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc, struct btrfs_root_item *root_item; struct btrfs_path *path; struct extent_buffer *leaf; - unsigned long nr; int level; int max_level; int replaced = 0; @@ -2126,10 +2125,9 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc, path->slots[level]); root_item->drop_level = level; - nr = trans->blocks_used; btrfs_end_transaction_throttle(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); if (replaced && rc->stage == UPDATE_DATA_PTRS) invalidate_extent_cache(root, &key, &next_key); @@ -2156,10 +2154,9 @@ out: btrfs_update_reloc_root(trans, root); } - nr = trans->blocks_used; btrfs_end_transaction_throttle(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); if (replaced && rc->stage == UPDATE_DATA_PTRS) invalidate_extent_cache(root, &key, &next_key); @@ -3262,7 +3259,6 @@ static int delete_block_group_cache(struct btrfs_fs_info *fs_info, struct btrfs_path *path; struct btrfs_root *root = fs_info->tree_root; struct btrfs_trans_handle *trans; - unsigned long nr; int ret = 0; if (inode) @@ -3296,9 +3292,8 @@ truncate: ret = btrfs_truncate_free_space_cache(root, trans, path, inode); btrfs_free_path(path); - nr = trans->blocks_used; btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); out: iput(inode); return ret; @@ -3715,7 +3710,6 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc) struct btrfs_trans_handle *trans = NULL; struct btrfs_path *path; struct btrfs_extent_item *ei; - unsigned long nr; u64 flags; u32 item_size; int ret; @@ -3832,9 +3826,8 @@ restart: ret = btrfs_commit_transaction(trans, rc->extent_root); BUG_ON(ret); } else { - nr = trans->blocks_used; btrfs_end_transaction_throttle(trans, rc->extent_root); - btrfs_btree_balance_dirty(rc->extent_root, nr); + btrfs_btree_balance_dirty(rc->extent_root); } trans = NULL; @@ -3864,9 +3857,8 @@ restart: GFP_NOFS); if (trans) { - nr = trans->blocks_used; btrfs_end_transaction_throttle(trans, rc->extent_root); - btrfs_btree_balance_dirty(rc->extent_root, nr); + btrfs_btree_balance_dirty(rc->extent_root); } if (!err) { @@ -3945,7 +3937,6 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info, struct btrfs_trans_handle *trans; struct btrfs_root *root; struct btrfs_key key; - unsigned long nr; u64 objectid = BTRFS_FIRST_FREE_OBJECTID; int err = 0; @@ -3973,9 +3964,8 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info, err = btrfs_orphan_add(trans, inode); out: - nr = trans->blocks_used; btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(root, nr); + btrfs_btree_balance_dirty(root); if (err) { if (inode) iput(inode); diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index f21f39f0b1a..7b297354e73 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -952,7 +952,6 @@ int btrfs_defrag_root(struct btrfs_root *root, int cacheonly) struct btrfs_fs_info *info = root->fs_info; struct btrfs_trans_handle *trans; int ret; - unsigned long nr; if (xchg(&root->defrag_running, 1)) return 0; @@ -964,9 +963,8 @@ int btrfs_defrag_root(struct btrfs_root *root, int cacheonly) ret = btrfs_defrag_leaves(trans, root, cacheonly); - nr = trans->blocks_used; btrfs_end_transaction(trans, root); - btrfs_btree_balance_dirty(info->tree_root, nr); + btrfs_btree_balance_dirty(info->tree_root); cond_resched(); if (btrfs_fs_closing(root->fs_info) || ret != -EAGAIN) -- cgit v1.2.3-70-g09d2 From 5d9640517d92d05843711ea982cbeff42d7ed32d Mon Sep 17 00:00:00 2001 From: Stefan Behrens Date: Mon, 5 Nov 2012 14:59:07 +0100 Subject: Btrfs: Pass fs_info to btrfs_num_copies() instead of mapping_tree This is required for the device replace procedure in a later step. Signed-off-by: Stefan Behrens Signed-off-by: Chris Mason --- fs/btrfs/check-integrity.c | 12 ++++++------ fs/btrfs/disk-io.c | 2 +- fs/btrfs/extent_io.c | 11 +++++------ fs/btrfs/volumes.c | 3 ++- fs/btrfs/volumes.h | 2 +- 5 files changed, 15 insertions(+), 15 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c index 5a3e45db642..58dfac1359a 100644 --- a/fs/btrfs/check-integrity.c +++ b/fs/btrfs/check-integrity.c @@ -723,7 +723,7 @@ static int btrfsic_process_superblock(struct btrfsic_state *state, } num_copies = - btrfs_num_copies(&state->root->fs_info->mapping_tree, + btrfs_num_copies(state->root->fs_info, next_bytenr, state->metablock_size); if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", @@ -903,7 +903,7 @@ static int btrfsic_process_superblock_dev_mirror( } num_copies = - btrfs_num_copies(&state->root->fs_info->mapping_tree, + btrfs_num_copies(state->root->fs_info, next_bytenr, state->metablock_size); if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", @@ -1287,7 +1287,7 @@ static int btrfsic_create_link_to_next_block( *next_blockp = NULL; if (0 == *num_copiesp) { *num_copiesp = - btrfs_num_copies(&state->root->fs_info->mapping_tree, + btrfs_num_copies(state->root->fs_info, next_bytenr, state->metablock_size); if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", @@ -1489,7 +1489,7 @@ static int btrfsic_handle_extent_data( chunk_len = num_bytes; num_copies = - btrfs_num_copies(&state->root->fs_info->mapping_tree, + btrfs_num_copies(state->root->fs_info, next_bytenr, state->datablock_size); if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", @@ -2463,7 +2463,7 @@ static int btrfsic_process_written_superblock( } num_copies = - btrfs_num_copies(&state->root->fs_info->mapping_tree, + btrfs_num_copies(state->root->fs_info, next_bytenr, BTRFS_SUPER_INFO_SIZE); if (state->print_mask & BTRFSIC_PRINT_MASK_NUM_COPIES) printk(KERN_INFO "num_copies(log_bytenr=%llu) = %d\n", @@ -2960,7 +2960,7 @@ static void btrfsic_cmp_log_and_dev_bytenr(struct btrfsic_state *state, struct btrfsic_block_data_ctx block_ctx; int match = 0; - num_copies = btrfs_num_copies(&state->root->fs_info->mapping_tree, + num_copies = btrfs_num_copies(state->root->fs_info, bytenr, state->metablock_size); for (mirror_num = 1; mirror_num <= num_copies; mirror_num++) { diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index ff5d259ac27..ba2b931fd8f 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -387,7 +387,7 @@ static int btree_read_extent_buffer_pages(struct btrfs_root *root, if (test_bit(EXTENT_BUFFER_CORRUPT, &eb->bflags)) break; - num_copies = btrfs_num_copies(&root->fs_info->mapping_tree, + num_copies = btrfs_num_copies(root->fs_info, eb->start, eb->len); if (num_copies == 1) break; diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 3c062c8d1d7..e0b7138909f 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2044,10 +2044,10 @@ static int clean_io_failure(u64 start, struct page *page) spin_unlock(&BTRFS_I(inode)->io_tree.lock); if (state && state->start == failrec->start) { - map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree; - num_copies = btrfs_num_copies(map_tree, failrec->logical, - failrec->len); + num_copies = btrfs_num_copies(BTRFS_I(inode)->root->fs_info, + failrec->logical, failrec->len); if (num_copies > 1) { + map_tree = &BTRFS_I(inode)->root->fs_info->mapping_tree; ret = repair_io_failure(map_tree, start, failrec->len, failrec->logical, page, failrec->failed_mirror); @@ -2157,9 +2157,8 @@ static int bio_readpage_error(struct bio *failed_bio, struct page *page, * clean_io_failure() clean all those errors at once. */ } - num_copies = btrfs_num_copies( - &BTRFS_I(inode)->root->fs_info->mapping_tree, - failrec->logical, failrec->len); + num_copies = btrfs_num_copies(BTRFS_I(inode)->root->fs_info, + failrec->logical, failrec->len); if (num_copies == 1) { /* * we only have a single copy of the data, so don't bother with diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 1483041eb86..5612767b910 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -3785,8 +3785,9 @@ void btrfs_mapping_tree_free(struct btrfs_mapping_tree *tree) } } -int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len) +int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len) { + struct btrfs_mapping_tree *map_tree = &fs_info->mapping_tree; struct extent_map *em; struct map_lookup *map; struct extent_map_tree *em_tree = &map_tree->map_tree; diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 657bb12b306..35ea4424963 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -278,7 +278,7 @@ int btrfs_add_device(struct btrfs_trans_handle *trans, struct btrfs_device *device); int btrfs_rm_device(struct btrfs_root *root, char *device_path); void btrfs_cleanup_fs_uuids(void); -int btrfs_num_copies(struct btrfs_mapping_tree *map_tree, u64 logical, u64 len); +int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len); int btrfs_grow_device(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 new_size); struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, -- cgit v1.2.3-70-g09d2 From aa1b8cd409f05e1489ec77ff219eff6ed4b801b8 Mon Sep 17 00:00:00 2001 From: Stefan Behrens Date: Mon, 5 Nov 2012 17:03:39 +0100 Subject: Btrfs: pass fs_info instead of root A small number of functions that are used in a device replace procedure when the operation is resumed at mount time are unable to pass the same root pointer that would be used in the regular (ioctl) context. And since the root pointer is not required, only the fs_info is, the root pointer argument is replaced with the fs_info pointer argument. Signed-off-by: Stefan Behrens Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 11 ++++---- fs/btrfs/disk-io.c | 4 +-- fs/btrfs/ioctl.c | 8 +++--- fs/btrfs/scrub.c | 76 ++++++++++++++++++++++++------------------------------ fs/btrfs/super.c | 2 +- fs/btrfs/volumes.c | 23 +++++++++-------- fs/btrfs/volumes.h | 2 +- 7 files changed, 60 insertions(+), 66 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index f9a078661eb..f8bb62c82b0 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -3540,15 +3540,16 @@ int btrfs_reloc_post_snapshot(struct btrfs_trans_handle *trans, struct btrfs_pending_snapshot *pending); /* scrub.c */ -int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, - struct btrfs_scrub_progress *progress, int readonly); +int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start, + u64 end, struct btrfs_scrub_progress *progress, + int readonly); void btrfs_scrub_pause(struct btrfs_root *root); void btrfs_scrub_pause_super(struct btrfs_root *root); void btrfs_scrub_continue(struct btrfs_root *root); void btrfs_scrub_continue_super(struct btrfs_root *root); -int __btrfs_scrub_cancel(struct btrfs_fs_info *info); -int btrfs_scrub_cancel(struct btrfs_root *root); -int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev); +int btrfs_scrub_cancel(struct btrfs_fs_info *info); +int btrfs_scrub_cancel_dev(struct btrfs_fs_info *info, + struct btrfs_device *dev); int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid); int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, struct btrfs_scrub_progress *progress); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index ba2b931fd8f..42a8024e935 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3283,9 +3283,9 @@ int close_ctree(struct btrfs_root *root) smp_mb(); /* pause restriper - we want to resume on mount */ - btrfs_pause_balance(root->fs_info); + btrfs_pause_balance(fs_info); - btrfs_scrub_cancel(root); + btrfs_scrub_cancel(fs_info); /* wait for any defraggers to finish */ wait_event(fs_info->transaction_wait, diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index e262cd8c4a7..b40b827f93e 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -1343,7 +1343,7 @@ static noinline int btrfs_ioctl_resize(struct btrfs_root *root, printk(KERN_INFO "btrfs: resizing devid %llu\n", (unsigned long long)devid); } - device = btrfs_find_device(root, devid, NULL, NULL); + device = btrfs_find_device(root->fs_info, devid, NULL, NULL); if (!device) { printk(KERN_INFO "btrfs: resizer unable to find device %llu\n", (unsigned long long)devid); @@ -2332,7 +2332,7 @@ static long btrfs_ioctl_dev_info(struct btrfs_root *root, void __user *arg) s_uuid = di_args->uuid; mutex_lock(&fs_devices->device_list_mutex); - dev = btrfs_find_device(root, di_args->devid, s_uuid, NULL); + dev = btrfs_find_device(root->fs_info, di_args->devid, s_uuid, NULL); mutex_unlock(&fs_devices->device_list_mutex); if (!dev) { @@ -3089,7 +3089,7 @@ static long btrfs_ioctl_scrub(struct btrfs_root *root, void __user *arg) if (IS_ERR(sa)) return PTR_ERR(sa); - ret = btrfs_scrub_dev(root, sa->devid, sa->start, sa->end, + ret = btrfs_scrub_dev(root->fs_info, sa->devid, sa->start, sa->end, &sa->progress, sa->flags & BTRFS_SCRUB_READONLY); if (copy_to_user(arg, sa, sizeof(*sa))) @@ -3104,7 +3104,7 @@ static long btrfs_ioctl_scrub_cancel(struct btrfs_root *root, void __user *arg) if (!capable(CAP_SYS_ADMIN)) return -EPERM; - return btrfs_scrub_cancel(root); + return btrfs_scrub_cancel(root->fs_info); } static long btrfs_ioctl_scrub_progress(struct btrfs_root *root, diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 894bb2732fc..6cf23f4f7bb 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -2262,9 +2262,8 @@ static noinline_for_stack int scrub_supers(struct scrub_ctx *sctx, /* * get a reference count on fs_info->scrub_workers. start worker if necessary */ -static noinline_for_stack int scrub_workers_get(struct btrfs_root *root) +static noinline_for_stack int scrub_workers_get(struct btrfs_fs_info *fs_info) { - struct btrfs_fs_info *fs_info = root->fs_info; int ret = 0; mutex_lock(&fs_info->scrub_lock); @@ -2283,10 +2282,8 @@ out: return ret; } -static noinline_for_stack void scrub_workers_put(struct btrfs_root *root) +static noinline_for_stack void scrub_workers_put(struct btrfs_fs_info *fs_info) { - struct btrfs_fs_info *fs_info = root->fs_info; - mutex_lock(&fs_info->scrub_lock); if (--fs_info->scrub_workers_refcnt == 0) btrfs_stop_workers(&fs_info->scrub_workers); @@ -2294,29 +2291,29 @@ static noinline_for_stack void scrub_workers_put(struct btrfs_root *root) mutex_unlock(&fs_info->scrub_lock); } - -int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, - struct btrfs_scrub_progress *progress, int readonly) +int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start, + u64 end, struct btrfs_scrub_progress *progress, + int readonly) { struct scrub_ctx *sctx; - struct btrfs_fs_info *fs_info = root->fs_info; int ret; struct btrfs_device *dev; - if (btrfs_fs_closing(root->fs_info)) + if (btrfs_fs_closing(fs_info)) return -EINVAL; /* * check some assumptions */ - if (root->nodesize != root->leafsize) { + if (fs_info->chunk_root->nodesize != fs_info->chunk_root->leafsize) { printk(KERN_ERR "btrfs_scrub: size assumption nodesize == leafsize (%d == %d) fails\n", - root->nodesize, root->leafsize); + fs_info->chunk_root->nodesize, + fs_info->chunk_root->leafsize); return -EINVAL; } - if (root->nodesize > BTRFS_STRIPE_LEN) { + if (fs_info->chunk_root->nodesize > BTRFS_STRIPE_LEN) { /* * in this case scrub is unable to calculate the checksum * the way scrub is implemented. Do not handle this @@ -2324,15 +2321,16 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, */ printk(KERN_ERR "btrfs_scrub: size assumption nodesize <= BTRFS_STRIPE_LEN (%d <= %d) fails\n", - root->nodesize, BTRFS_STRIPE_LEN); + fs_info->chunk_root->nodesize, BTRFS_STRIPE_LEN); return -EINVAL; } - if (root->sectorsize != PAGE_SIZE) { + if (fs_info->chunk_root->sectorsize != PAGE_SIZE) { /* not supported for data w/o checksums */ printk(KERN_ERR "btrfs_scrub: size assumption sectorsize != PAGE_SIZE (%d != %lld) fails\n", - root->sectorsize, (unsigned long long)PAGE_SIZE); + fs_info->chunk_root->sectorsize, + (unsigned long long)PAGE_SIZE); return -EINVAL; } @@ -2352,37 +2350,37 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, return -EINVAL; } - ret = scrub_workers_get(root); + ret = scrub_workers_get(fs_info); if (ret) return ret; - mutex_lock(&root->fs_info->fs_devices->device_list_mutex); - dev = btrfs_find_device(root, devid, NULL, NULL); + mutex_lock(&fs_info->fs_devices->device_list_mutex); + dev = btrfs_find_device(fs_info, devid, NULL, NULL); if (!dev || dev->missing) { - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); - scrub_workers_put(root); + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + scrub_workers_put(fs_info); return -ENODEV; } mutex_lock(&fs_info->scrub_lock); if (!dev->in_fs_metadata) { mutex_unlock(&fs_info->scrub_lock); - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); - scrub_workers_put(root); - return -ENODEV; + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + scrub_workers_put(fs_info); + return -EIO; } if (dev->scrub_device) { mutex_unlock(&fs_info->scrub_lock); - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); - scrub_workers_put(root); + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + scrub_workers_put(fs_info); return -EINPROGRESS; } sctx = scrub_setup_ctx(dev); if (IS_ERR(sctx)) { mutex_unlock(&fs_info->scrub_lock); - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); - scrub_workers_put(root); + mutex_unlock(&fs_info->fs_devices->device_list_mutex); + scrub_workers_put(fs_info); return PTR_ERR(sctx); } sctx->readonly = readonly; @@ -2390,7 +2388,7 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, atomic_inc(&fs_info->scrubs_running); mutex_unlock(&fs_info->scrub_lock); - mutex_unlock(&root->fs_info->fs_devices->device_list_mutex); + mutex_unlock(&fs_info->fs_devices->device_list_mutex); down_read(&fs_info->scrub_super_lock); ret = scrub_supers(sctx, dev); @@ -2413,7 +2411,7 @@ int btrfs_scrub_dev(struct btrfs_root *root, u64 devid, u64 start, u64 end, mutex_unlock(&fs_info->scrub_lock); scrub_free_ctx(sctx); - scrub_workers_put(root); + scrub_workers_put(fs_info); return ret; } @@ -2453,9 +2451,8 @@ void btrfs_scrub_continue_super(struct btrfs_root *root) up_write(&root->fs_info->scrub_super_lock); } -int __btrfs_scrub_cancel(struct btrfs_fs_info *fs_info) +int btrfs_scrub_cancel(struct btrfs_fs_info *fs_info) { - mutex_lock(&fs_info->scrub_lock); if (!atomic_read(&fs_info->scrubs_running)) { mutex_unlock(&fs_info->scrub_lock); @@ -2475,14 +2472,9 @@ int __btrfs_scrub_cancel(struct btrfs_fs_info *fs_info) return 0; } -int btrfs_scrub_cancel(struct btrfs_root *root) +int btrfs_scrub_cancel_dev(struct btrfs_fs_info *fs_info, + struct btrfs_device *dev) { - return __btrfs_scrub_cancel(root->fs_info); -} - -int btrfs_scrub_cancel_dev(struct btrfs_root *root, struct btrfs_device *dev) -{ - struct btrfs_fs_info *fs_info = root->fs_info; struct scrub_ctx *sctx; mutex_lock(&fs_info->scrub_lock); @@ -2514,12 +2506,12 @@ int btrfs_scrub_cancel_devid(struct btrfs_root *root, u64 devid) * does not go away in cancel_dev. FIXME: find a better solution */ mutex_lock(&fs_info->fs_devices->device_list_mutex); - dev = btrfs_find_device(root, devid, NULL, NULL); + dev = btrfs_find_device(fs_info, devid, NULL, NULL); if (!dev) { mutex_unlock(&fs_info->fs_devices->device_list_mutex); return -ENODEV; } - ret = btrfs_scrub_cancel_dev(root, dev); + ret = btrfs_scrub_cancel_dev(fs_info, dev); mutex_unlock(&fs_info->fs_devices->device_list_mutex); return ret; @@ -2532,7 +2524,7 @@ int btrfs_scrub_progress(struct btrfs_root *root, u64 devid, struct scrub_ctx *sctx = NULL; mutex_lock(&root->fs_info->fs_devices->device_list_mutex); - dev = btrfs_find_device(root, devid, NULL, NULL); + dev = btrfs_find_device(root->fs_info, devid, NULL, NULL); if (dev) sctx = dev->scrub_device; if (sctx) diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index acd2df85bed..a1a6c296ddc 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -116,7 +116,7 @@ static void btrfs_handle_error(struct btrfs_fs_info *fs_info) if (fs_info->fs_state & BTRFS_SUPER_FLAG_ERROR) { sb->s_flags |= MS_RDONLY; printk(KERN_INFO "btrfs is forced readonly\n"); - __btrfs_scrub_cancel(fs_info); + btrfs_scrub_cancel(fs_info); // WARN_ON(1); } } diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 6cd8a32c448..d2c0bccca60 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1398,7 +1398,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) disk_super = (struct btrfs_super_block *)bh->b_data; devid = btrfs_stack_device_id(&disk_super->dev_item); dev_uuid = disk_super->dev_item.uuid; - device = btrfs_find_device(root, devid, dev_uuid, + device = btrfs_find_device(root->fs_info, devid, dev_uuid, disk_super->fsid); if (!device) { ret = -ENOENT; @@ -1435,7 +1435,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) spin_unlock(&root->fs_info->free_chunk_lock); device->in_fs_metadata = 0; - btrfs_scrub_cancel_dev(root, device); + btrfs_scrub_cancel_dev(root->fs_info, device); /* * the device list mutex makes sure that we don't change @@ -1492,7 +1492,7 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) * at this point, the device is zero sized. We want to * remove it from the devices list and zero out the old super */ - if (clear_super) { + if (clear_super && disk_super) { /* make sure this device isn't detected as part of * the FS anymore */ @@ -1540,7 +1540,7 @@ int btrfs_find_device_by_path(struct btrfs_root *root, char *device_path, disk_super = (struct btrfs_super_block *)bh->b_data; devid = btrfs_stack_device_id(&disk_super->dev_item); dev_uuid = disk_super->dev_item.uuid; - *device = btrfs_find_device(root, devid, dev_uuid, + *device = btrfs_find_device(root->fs_info, devid, dev_uuid, disk_super->fsid); brelse(bh); if (!*device) @@ -1699,7 +1699,8 @@ next_slot: read_extent_buffer(leaf, fs_uuid, (unsigned long)btrfs_device_fsid(dev_item), BTRFS_UUID_SIZE); - device = btrfs_find_device(root, devid, dev_uuid, fs_uuid); + device = btrfs_find_device(root->fs_info, devid, dev_uuid, + fs_uuid); BUG_ON(!device); /* Logic error */ if (device->fs_devices->seeding) { @@ -4463,13 +4464,13 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, return 0; } -struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, +struct btrfs_device *btrfs_find_device(struct btrfs_fs_info *fs_info, u64 devid, u8 *uuid, u8 *fsid) { struct btrfs_device *device; struct btrfs_fs_devices *cur_devices; - cur_devices = root->fs_info->fs_devices; + cur_devices = fs_info->fs_devices; while (cur_devices) { if (!fsid || !memcmp(cur_devices->fsid, fsid, BTRFS_UUID_SIZE)) { @@ -4567,8 +4568,8 @@ static int read_one_chunk(struct btrfs_root *root, struct btrfs_key *key, read_extent_buffer(leaf, uuid, (unsigned long) btrfs_stripe_dev_uuid_nr(chunk, i), BTRFS_UUID_SIZE); - map->stripes[i].dev = btrfs_find_device(root, devid, uuid, - NULL); + map->stripes[i].dev = btrfs_find_device(root->fs_info, devid, + uuid, NULL); if (!map->stripes[i].dev && !btrfs_test_opt(root, DEGRADED)) { kfree(map); free_extent_map(em); @@ -4686,7 +4687,7 @@ static int read_one_dev(struct btrfs_root *root, return ret; } - device = btrfs_find_device(root, devid, dev_uuid, fs_uuid); + device = btrfs_find_device(root->fs_info, devid, dev_uuid, fs_uuid); if (!device || !device->bdev) { if (!btrfs_test_opt(root, DEGRADED)) return -EIO; @@ -5078,7 +5079,7 @@ int btrfs_get_dev_stats(struct btrfs_root *root, int i; mutex_lock(&fs_devices->device_list_mutex); - dev = btrfs_find_device(root, stats->devid, NULL, NULL); + dev = btrfs_find_device(root->fs_info, stats->devid, NULL, NULL); mutex_unlock(&fs_devices->device_list_mutex); if (!dev) { diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 7eaaf4e6195..802e2ba02f0 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -281,7 +281,7 @@ void btrfs_cleanup_fs_uuids(void); int btrfs_num_copies(struct btrfs_fs_info *fs_info, u64 logical, u64 len); int btrfs_grow_device(struct btrfs_trans_handle *trans, struct btrfs_device *device, u64 new_size); -struct btrfs_device *btrfs_find_device(struct btrfs_root *root, u64 devid, +struct btrfs_device *btrfs_find_device(struct btrfs_fs_info *fs_info, u64 devid, u8 *uuid, u8 *fsid); int btrfs_shrink_device(struct btrfs_device *device, u64 new_size); int btrfs_init_new_device(struct btrfs_root *root, char *path); -- cgit v1.2.3-70-g09d2 From e922e087a35c437acef3bc88ce31e59c699c38bd Mon Sep 17 00:00:00 2001 From: Stefan Behrens Date: Mon, 5 Nov 2012 17:26:40 +0100 Subject: Btrfs: enhance btrfs structures for device replace support Signed-off-by: Stefan Behrens Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 39 +++++++++++++++++++++++++++++++++++++++ fs/btrfs/disk-io.c | 5 +++++ 2 files changed, 44 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index f8bb62c82b0..0781fd4a5c1 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -885,6 +885,42 @@ struct btrfs_dev_stats_item { __le64 values[BTRFS_DEV_STAT_VALUES_MAX]; } __attribute__ ((__packed__)); +#define BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_ALWAYS 0 +#define BTRFS_DEV_REPLACE_ITEM_CONT_READING_FROM_SRCDEV_MODE_AVOID 1 +#define BTRFS_DEV_REPLACE_ITEM_STATE_NEVER_STARTED 0 +#define BTRFS_DEV_REPLACE_ITEM_STATE_STARTED 1 +#define BTRFS_DEV_REPLACE_ITEM_STATE_SUSPENDED 2 +#define BTRFS_DEV_REPLACE_ITEM_STATE_FINISHED 3 +#define BTRFS_DEV_REPLACE_ITEM_STATE_CANCELED 4 + +struct btrfs_dev_replace { + u64 replace_state; /* see #define above */ + u64 time_started; /* seconds since 1-Jan-1970 */ + u64 time_stopped; /* seconds since 1-Jan-1970 */ + atomic64_t num_write_errors; + atomic64_t num_uncorrectable_read_errors; + + u64 cursor_left; + u64 committed_cursor_left; + u64 cursor_left_last_write_of_item; + u64 cursor_right; + + u64 cont_reading_from_srcdev_mode; /* see #define above */ + + int is_valid; + int item_needs_writeback; + struct btrfs_device *srcdev; + struct btrfs_device *tgtdev; + + pid_t lock_owner; + atomic_t nesting_level; + struct mutex lock_finishing_cancel_unmount; + struct mutex lock_management_lock; + struct mutex lock; + + struct btrfs_scrub_progress scrub_progress; +}; + /* different types of block groups (and chunks) */ #define BTRFS_BLOCK_GROUP_DATA (1ULL << 0) #define BTRFS_BLOCK_GROUP_SYSTEM (1ULL << 1) @@ -1471,6 +1507,9 @@ struct btrfs_fs_info { int backup_root_index; int num_tolerated_disk_barrier_failures; + + /* device replace state */ + struct btrfs_dev_replace dev_replace; }; /* diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 42a8024e935..9d1b7106081 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2131,6 +2131,11 @@ int open_ctree(struct super_block *sb, init_rwsem(&fs_info->extent_commit_sem); init_rwsem(&fs_info->cleanup_work_sem); init_rwsem(&fs_info->subvol_sem); + fs_info->dev_replace.lock_owner = 0; + atomic_set(&fs_info->dev_replace.nesting_level, 0); + mutex_init(&fs_info->dev_replace.lock_finishing_cancel_unmount); + mutex_init(&fs_info->dev_replace.lock_management_lock); + mutex_init(&fs_info->dev_replace.lock); spin_lock_init(&fs_info->qgroup_lock); fs_info->qgroup_tree = RB_ROOT; -- cgit v1.2.3-70-g09d2 From 618919236ba54361e93106f4951d233a7ade63cd Mon Sep 17 00:00:00 2001 From: Stefan Behrens Date: Mon, 5 Nov 2012 18:51:52 +0100 Subject: Btrfs: handle errors from btrfs_map_bio() everywhere With the addition of the device replace procedure, it is possible for btrfs_map_bio(READ) to report an error. This happens when the specific mirror is requested which is located on the target disk, and the copy operation has not yet copied this block. Hence the block cannot be read and this error state is indicated by returning EIO. Some background information follows now. A new mirror is added while the device replace procedure is running. btrfs_get_num_copies() returns one more, and btrfs_map_bio(GET_READ_MIRROR) adds one more mirror if a disk location is involved that was already handled by the device replace copy operation. The assigned mirror num is the highest mirror number, e.g. the value 3 in case of RAID1. If btrfs_map_bio() is invoked with mirror_num == 0 (i.e., select any mirror), the copy on the target drive is never selected because that disk shall be able to perform the write requests as quickly as possible. The parallel execution of read requests would only slow down the disk copy procedure. Second case is that btrfs_map_bio() is called with mirror_num > 0. This is done from the repair code only. In this case, the highest mirror num is assigned to the target disk, since it is used last. And when this mirror is not available because the copy procedure has not yet handled this area, an error is returned. Everywhere in the code the handling of such errors is added now. Signed-off-by: Stefan Behrens Signed-off-by: Chris Mason --- fs/btrfs/check-integrity.c | 15 +++++++++++++-- fs/btrfs/compression.c | 6 ++++-- fs/btrfs/disk-io.c | 44 +++++++++++++++++++++++++++----------------- fs/btrfs/extent_io.c | 4 ---- fs/btrfs/inode.c | 27 ++++++++++++++++++++------- fs/btrfs/volumes.c | 2 +- 6 files changed, 65 insertions(+), 33 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c index 8f9abedae2c..badc6f141b6 100644 --- a/fs/btrfs/check-integrity.c +++ b/fs/btrfs/check-integrity.c @@ -1585,6 +1585,18 @@ static int btrfsic_map_block(struct btrfsic_state *state, u64 bytenr, u32 len, ret = btrfs_map_block(state->root->fs_info, READ, bytenr, &length, &multi, mirror_num); + if (ret) { + block_ctx_out->start = 0; + block_ctx_out->dev_bytenr = 0; + block_ctx_out->len = 0; + block_ctx_out->dev = NULL; + block_ctx_out->datav = NULL; + block_ctx_out->pagev = NULL; + block_ctx_out->mem_to_free = NULL; + + return ret; + } + device = multi->stripes[0].dev; block_ctx_out->dev = btrfsic_dev_state_lookup(device->bdev); block_ctx_out->dev_bytenr = multi->stripes[0].physical; @@ -1594,8 +1606,7 @@ static int btrfsic_map_block(struct btrfsic_state *state, u64 bytenr, u32 len, block_ctx_out->pagev = NULL; block_ctx_out->mem_to_free = NULL; - if (0 == ret) - kfree(multi); + kfree(multi); if (NULL == block_ctx_out->dev) { ret = -ENXIO; printk(KERN_INFO "btrfsic: error, cannot lookup dev (#1)!\n"); diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index c6467aa88be..94ab2f80e7e 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -687,7 +687,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0); - BUG_ON(ret); /* -ENOMEM */ + if (ret) + bio_endio(comp_bio, ret); bio_put(comp_bio); @@ -712,7 +713,8 @@ int btrfs_submit_compressed_read(struct inode *inode, struct bio *bio, } ret = btrfs_map_bio(root, READ, comp_bio, mirror_num, 0); - BUG_ON(ret); /* -ENOMEM */ + if (ret) + bio_endio(comp_bio, ret); bio_put(comp_bio); return 0; diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 9d1b7106081..0e410478ad2 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -852,11 +852,16 @@ static int __btree_submit_bio_done(struct inode *inode, int rw, struct bio *bio, int mirror_num, unsigned long bio_flags, u64 bio_offset) { + int ret; + /* * when we're called for a write, we're already in the async * submission context. Just jump into btrfs_map_bio */ - return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, mirror_num, 1); + ret = btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, mirror_num, 1); + if (ret) + bio_endio(bio, ret); + return ret; } static int check_async_write(struct inode *inode, unsigned long bio_flags) @@ -878,7 +883,6 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, int ret; if (!(rw & REQ_WRITE)) { - /* * called for a read, do the setup so that checksum validation * can happen in the async kernel threads @@ -886,26 +890,32 @@ static int btree_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, ret = btrfs_bio_wq_end_io(BTRFS_I(inode)->root->fs_info, bio, 1); if (ret) - return ret; - return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, - mirror_num, 0); + goto out_w_error; + ret = btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, + mirror_num, 0); } else if (!async) { ret = btree_csum_one_bio(bio); if (ret) - return ret; - return btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, - mirror_num, 0); + goto out_w_error; + ret = btrfs_map_bio(BTRFS_I(inode)->root, rw, bio, + mirror_num, 0); + } else { + /* + * kthread helpers are used to submit writes so that + * checksumming can happen in parallel across all CPUs + */ + ret = btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info, + inode, rw, bio, mirror_num, 0, + bio_offset, + __btree_submit_bio_start, + __btree_submit_bio_done); } - /* - * kthread helpers are used to submit writes so that checksumming - * can happen in parallel across all CPUs - */ - return btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info, - inode, rw, bio, mirror_num, 0, - bio_offset, - __btree_submit_bio_start, - __btree_submit_bio_done); + if (ret) { +out_w_error: + bio_endio(bio, ret); + } + return ret; } #ifdef CONFIG_MIGRATION diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 62ec6e45f70..1b319df29ee 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -2462,10 +2462,6 @@ btrfs_bio_alloc(struct block_device *bdev, u64 first_sector, int nr_vecs, return bio; } -/* - * Since writes are async, they will only return -ENOMEM. - * Reads can return the full range of I/O error conditions. - */ static int __must_check submit_one_bio(int rw, struct bio *bio, int mirror_num, unsigned long bio_flags) { diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 5d1675a8c9e..d7bf2e7ee8a 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -1602,7 +1602,12 @@ static int __btrfs_submit_bio_done(struct inode *inode, int rw, struct bio *bio, u64 bio_offset) { struct btrfs_root *root = BTRFS_I(inode)->root; - return btrfs_map_bio(root, rw, bio, mirror_num, 1); + int ret; + + ret = btrfs_map_bio(root, rw, bio, mirror_num, 1); + if (ret) + bio_endio(bio, ret); + return ret; } /* @@ -1626,15 +1631,17 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, if (!(rw & REQ_WRITE)) { ret = btrfs_bio_wq_end_io(root->fs_info, bio, metadata); if (ret) - return ret; + goto out; if (bio_flags & EXTENT_BIO_COMPRESSED) { - return btrfs_submit_compressed_read(inode, bio, - mirror_num, bio_flags); + ret = btrfs_submit_compressed_read(inode, bio, + mirror_num, + bio_flags); + goto out; } else if (!skip_sum) { ret = btrfs_lookup_bio_sums(root, inode, bio, NULL); if (ret) - return ret; + goto out; } goto mapit; } else if (!skip_sum) { @@ -1642,15 +1649,21 @@ static int btrfs_submit_bio_hook(struct inode *inode, int rw, struct bio *bio, if (root->root_key.objectid == BTRFS_DATA_RELOC_TREE_OBJECTID) goto mapit; /* we're doing a write, do the async checksumming */ - return btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info, + ret = btrfs_wq_submit_bio(BTRFS_I(inode)->root->fs_info, inode, rw, bio, mirror_num, bio_flags, bio_offset, __btrfs_submit_bio_start, __btrfs_submit_bio_done); + goto out; } mapit: - return btrfs_map_bio(root, rw, bio, mirror_num, 0); + ret = btrfs_map_bio(root, rw, bio, mirror_num, 0); + +out: + if (ret < 0) + bio_endio(bio, ret); + return ret; } /* diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 31f7af878d9..415862885b6 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -4435,7 +4435,7 @@ int btrfs_map_bio(struct btrfs_root *root, int rw, struct bio *bio, ret = btrfs_map_block(root->fs_info, rw, logical, &map_length, &bbio, mirror_num); - if (ret) /* -ENOMEM */ + if (ret) return ret; total_devs = bbio->num_stripes; -- cgit v1.2.3-70-g09d2 From 8dabb7420f014ab0f9f04afae8ae046c0f48b270 Mon Sep 17 00:00:00 2001 From: Stefan Behrens Date: Tue, 6 Nov 2012 13:15:27 +0100 Subject: Btrfs: change core code of btrfs to support the device replace operations This commit contains all the essential changes to the core code of Btrfs for support of the device replace procedure. Signed-off-by: Stefan Behrens Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 24 +++++++++++++++++++++- fs/btrfs/reada.c | 17 ++++++++++++++++ fs/btrfs/scrub.c | 7 ++++++- fs/btrfs/super.c | 13 ++++++++++++ fs/btrfs/transaction.c | 7 ++++++- fs/btrfs/volumes.c | 54 ++++++++++++++++++++++++++++++++++++++++---------- fs/btrfs/volumes.h | 3 ++- 7 files changed, 111 insertions(+), 14 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 0e410478ad2..76b82506bf9 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -45,6 +45,7 @@ #include "inode-map.h" #include "check-integrity.h" #include "rcu-string.h" +#include "dev-replace.h" #ifdef CONFIG_X86 #include @@ -2438,7 +2439,11 @@ int open_ctree(struct super_block *sb, goto fail_tree_roots; } - btrfs_close_extra_devices(fs_devices); + /* + * keep the device that is marked to be the target device for the + * dev_replace procedure + */ + btrfs_close_extra_devices(fs_info, fs_devices, 0); if (!fs_devices->latest_bdev) { printk(KERN_CRIT "btrfs: failed to read devices on %s\n", @@ -2510,6 +2515,14 @@ retry_root_backup: goto fail_block_groups; } + ret = btrfs_init_dev_replace(fs_info); + if (ret) { + pr_err("btrfs: failed to init dev_replace: %d\n", ret); + goto fail_block_groups; + } + + btrfs_close_extra_devices(fs_info, fs_devices, 1); + ret = btrfs_init_space_info(fs_info); if (ret) { printk(KERN_ERR "Failed to initial space info: %d\n", ret); @@ -2658,6 +2671,13 @@ retry_root_backup: return ret; } + ret = btrfs_resume_dev_replace_async(fs_info); + if (ret) { + pr_warn("btrfs: failed to resume dev_replace\n"); + close_ctree(tree_root); + return ret; + } + return 0; fail_qgroup: @@ -3300,6 +3320,8 @@ int close_ctree(struct btrfs_root *root) /* pause restriper - we want to resume on mount */ btrfs_pause_balance(fs_info); + btrfs_dev_replace_suspend_for_unmount(fs_info); + btrfs_scrub_cancel(fs_info); /* wait for any defraggers to finish */ diff --git a/fs/btrfs/reada.c b/fs/btrfs/reada.c index 9f363e17ec7..c705a48e676 100644 --- a/fs/btrfs/reada.c +++ b/fs/btrfs/reada.c @@ -27,6 +27,7 @@ #include "volumes.h" #include "disk-io.h" #include "transaction.h" +#include "dev-replace.h" #undef DEBUG @@ -331,6 +332,7 @@ static struct reada_extent *reada_find_extent(struct btrfs_root *root, int nzones = 0; int i; unsigned long index = logical >> PAGE_CACHE_SHIFT; + int dev_replace_is_ongoing; spin_lock(&fs_info->reada_lock); re = radix_tree_lookup(&fs_info->reada_tree, index); @@ -392,6 +394,7 @@ static struct reada_extent *reada_find_extent(struct btrfs_root *root, } /* insert extent in reada_tree + all per-device trees, all or nothing */ + btrfs_dev_replace_lock(&fs_info->dev_replace); spin_lock(&fs_info->reada_lock); ret = radix_tree_insert(&fs_info->reada_tree, index, re); if (ret == -EEXIST) { @@ -399,13 +402,17 @@ static struct reada_extent *reada_find_extent(struct btrfs_root *root, BUG_ON(!re_exist); re_exist->refcnt++; spin_unlock(&fs_info->reada_lock); + btrfs_dev_replace_unlock(&fs_info->dev_replace); goto error; } if (ret) { spin_unlock(&fs_info->reada_lock); + btrfs_dev_replace_unlock(&fs_info->dev_replace); goto error; } prev_dev = NULL; + dev_replace_is_ongoing = btrfs_dev_replace_is_ongoing( + &fs_info->dev_replace); for (i = 0; i < nzones; ++i) { dev = bbio->stripes[i].dev; if (dev == prev_dev) { @@ -422,6 +429,14 @@ static struct reada_extent *reada_find_extent(struct btrfs_root *root, /* cannot read ahead on missing device */ continue; } + if (dev_replace_is_ongoing && + dev == fs_info->dev_replace.tgtdev) { + /* + * as this device is selected for reading only as + * a last resort, skip it for read ahead. + */ + continue; + } prev_dev = dev; ret = radix_tree_insert(&dev->reada_extents, index, re); if (ret) { @@ -434,10 +449,12 @@ static struct reada_extent *reada_find_extent(struct btrfs_root *root, BUG_ON(fs_info == NULL); radix_tree_delete(&fs_info->reada_tree, index); spin_unlock(&fs_info->reada_lock); + btrfs_dev_replace_unlock(&fs_info->dev_replace); goto error; } } spin_unlock(&fs_info->reada_lock); + btrfs_dev_replace_unlock(&fs_info->dev_replace); kfree(bbio); return re; diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 61157a26cf2..30cbf6921c0 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -2843,12 +2843,17 @@ int btrfs_scrub_dev(struct btrfs_fs_info *fs_info, u64 devid, u64 start, return -EIO; } - if (dev->scrub_device) { + btrfs_dev_replace_lock(&fs_info->dev_replace); + if (dev->scrub_device || + (!is_dev_replace && + btrfs_dev_replace_is_ongoing(&fs_info->dev_replace))) { + btrfs_dev_replace_unlock(&fs_info->dev_replace); mutex_unlock(&fs_info->scrub_lock); mutex_unlock(&fs_info->fs_devices->device_list_mutex); scrub_workers_put(fs_info); return -EINPROGRESS; } + btrfs_dev_replace_unlock(&fs_info->dev_replace); sctx = scrub_setup_ctx(dev, is_dev_replace); if (IS_ERR(sctx)) { mutex_unlock(&fs_info->scrub_lock); diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index ad4380684b9..def4f24b58d 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -55,6 +55,7 @@ #include "export.h" #include "compression.h" #include "rcu-string.h" +#include "dev-replace.h" #define CREATE_TRACE_POINTS #include @@ -1225,8 +1226,15 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data) return 0; if (*flags & MS_RDONLY) { + /* + * this also happens on 'umount -rf' or on shutdown, when + * the filesystem is busy. + */ sb->s_flags |= MS_RDONLY; + btrfs_dev_replace_suspend_for_unmount(fs_info); + btrfs_scrub_cancel(fs_info); + ret = btrfs_commit_super(root); if (ret) goto restore; @@ -1263,6 +1271,11 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data) if (ret) goto restore; + ret = btrfs_resume_dev_replace_async(fs_info); + if (ret) { + pr_warn("btrfs: failed to resume dev_replace\n"); + goto restore; + } sb->s_flags &= ~MS_RDONLY; } diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c index 7b297354e73..bcc6b65be3b 100644 --- a/fs/btrfs/transaction.c +++ b/fs/btrfs/transaction.c @@ -30,6 +30,7 @@ #include "tree-log.h" #include "inode-map.h" #include "volumes.h" +#include "dev-replace.h" #define BTRFS_ROOT_TRANS_TAG 0 @@ -845,7 +846,9 @@ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans, return ret; ret = btrfs_run_dev_stats(trans, root->fs_info); - BUG_ON(ret); + WARN_ON(ret); + ret = btrfs_run_dev_replace(trans, root->fs_info); + WARN_ON(ret); ret = btrfs_run_qgroups(trans, root->fs_info); BUG_ON(ret); @@ -868,6 +871,8 @@ static noinline int commit_cowonly_roots(struct btrfs_trans_handle *trans, switch_commit_root(fs_info->extent_root); up_write(&fs_info->extent_commit_sem); + btrfs_after_dev_replace_commit(fs_info); + return 0; } diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index 5777e6a9aab..a4e0963bf45 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -36,6 +36,7 @@ #include "check-integrity.h" #include "rcu-string.h" #include "math.h" +#include "dev-replace.h" static int init_first_rw_device(struct btrfs_trans_handle *trans, struct btrfs_root *root, @@ -505,7 +506,8 @@ error: return ERR_PTR(-ENOMEM); } -void btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices) +void btrfs_close_extra_devices(struct btrfs_fs_info *fs_info, + struct btrfs_fs_devices *fs_devices, int step) { struct btrfs_device *device, *next; @@ -528,6 +530,21 @@ again: continue; } + if (device->devid == BTRFS_DEV_REPLACE_DEVID) { + /* + * In the first step, keep the device which has + * the correct fsid and the devid that is used + * for the dev_replace procedure. + * In the second step, the dev_replace state is + * read from the device tree and it is known + * whether the procedure is really active or + * not, which means whether this device is + * used or whether it should be removed. + */ + if (step == 0 || device->is_tgtdev_for_dev_replace) { + continue; + } + } if (device->bdev) { blkdev_put(device->bdev, device->mode); device->bdev = NULL; @@ -536,7 +553,8 @@ again: if (device->writeable) { list_del_init(&device->dev_alloc_list); device->writeable = 0; - fs_devices->rw_devices--; + if (!device->is_tgtdev_for_dev_replace) + fs_devices->rw_devices--; } list_del_init(&device->dev_list); fs_devices->num_devices--; @@ -594,7 +612,7 @@ static int __btrfs_close_devices(struct btrfs_fs_devices *fs_devices) if (device->bdev) fs_devices->open_devices--; - if (device->writeable) { + if (device->writeable && !device->is_tgtdev_for_dev_replace) { list_del_init(&device->dev_alloc_list); fs_devices->rw_devices--; } @@ -718,7 +736,7 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices, fs_devices->rotating = 1; fs_devices->open_devices++; - if (device->writeable) { + if (device->writeable && !device->is_tgtdev_for_dev_replace) { fs_devices->rw_devices++; list_add(&device->dev_alloc_list, &fs_devices->alloc_list); @@ -1350,16 +1368,22 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path) root->fs_info->avail_system_alloc_bits | root->fs_info->avail_metadata_alloc_bits; - if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && - root->fs_info->fs_devices->num_devices <= 4) { + num_devices = root->fs_info->fs_devices->num_devices; + btrfs_dev_replace_lock(&root->fs_info->dev_replace); + if (btrfs_dev_replace_is_ongoing(&root->fs_info->dev_replace)) { + WARN_ON(num_devices < 1); + num_devices--; + } + btrfs_dev_replace_unlock(&root->fs_info->dev_replace); + + if ((all_avail & BTRFS_BLOCK_GROUP_RAID10) && num_devices <= 4) { printk(KERN_ERR "btrfs: unable to go below four devices " "on raid10\n"); ret = -EINVAL; goto out; } - if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && - root->fs_info->fs_devices->num_devices <= 2) { + if ((all_avail & BTRFS_BLOCK_GROUP_RAID1) && num_devices <= 2) { printk(KERN_ERR "btrfs: unable to go below two " "devices on raid1\n"); ret = -EINVAL; @@ -2935,6 +2959,7 @@ int btrfs_balance(struct btrfs_balance_control *bctl, u64 allowed; int mixed = 0; int ret; + u64 num_devices; if (btrfs_fs_closing(fs_info) || atomic_read(&fs_info->balance_pause_req) || @@ -2963,10 +2988,17 @@ int btrfs_balance(struct btrfs_balance_control *bctl, } } + num_devices = fs_info->fs_devices->num_devices; + btrfs_dev_replace_lock(&fs_info->dev_replace); + if (btrfs_dev_replace_is_ongoing(&fs_info->dev_replace)) { + BUG_ON(num_devices < 1); + num_devices--; + } + btrfs_dev_replace_unlock(&fs_info->dev_replace); allowed = BTRFS_AVAIL_ALLOC_BIT_SINGLE; - if (fs_info->fs_devices->num_devices == 1) + if (num_devices == 1) allowed |= BTRFS_BLOCK_GROUP_DUP; - else if (fs_info->fs_devices->num_devices < 4) + else if (num_devices < 4) allowed |= (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1); else allowed |= (BTRFS_BLOCK_GROUP_RAID0 | BTRFS_BLOCK_GROUP_RAID1 | @@ -3591,6 +3623,7 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans, devices_info[ndevs].total_avail = total_avail; devices_info[ndevs].dev = device; ++ndevs; + WARN_ON(ndevs > fs_devices->rw_devices); } /* @@ -4773,6 +4806,7 @@ static void fill_device_from_item(struct extent_buffer *leaf, device->io_align = btrfs_device_io_align(leaf, dev_item); device->io_width = btrfs_device_io_width(leaf, dev_item); device->sector_size = btrfs_device_sector_size(leaf, dev_item); + WARN_ON(device->devid == BTRFS_DEV_REPLACE_DEVID); device->is_tgtdev_for_dev_replace = 0; ptr = (unsigned long)btrfs_device_uuid(dev_item); diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h index 58d79375dea..37d0157167b 100644 --- a/fs/btrfs/volumes.h +++ b/fs/btrfs/volumes.h @@ -268,7 +268,8 @@ int btrfs_open_devices(struct btrfs_fs_devices *fs_devices, int btrfs_scan_one_device(const char *path, fmode_t flags, void *holder, struct btrfs_fs_devices **fs_devices_ret); int btrfs_close_devices(struct btrfs_fs_devices *fs_devices); -void btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices); +void btrfs_close_extra_devices(struct btrfs_fs_info *fs_info, + struct btrfs_fs_devices *fs_devices, int step); int btrfs_find_device_missing_or_by_path(struct btrfs_root *root, char *device_path, struct btrfs_device **device); -- cgit v1.2.3-70-g09d2 From 26176e7c2aa923327becdc25b5aca2cb907ac932 Mon Sep 17 00:00:00 2001 From: Miao Xie Date: Mon, 26 Nov 2012 09:26:20 +0000 Subject: Btrfs: restructure btrfs_run_defrag_inodes() This patch restructure btrfs_run_defrag_inodes() and make the code of the auto defragment more readable. Signed-off-by: Miao Xie Signed-off-by: Chris Mason --- fs/btrfs/ctree.h | 1 + fs/btrfs/disk-io.c | 2 +- fs/btrfs/file.c | 197 +++++++++++++++++++++++++++++------------------------ 3 files changed, 109 insertions(+), 91 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 389c05715ea..6ba56aea5b6 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -3510,6 +3510,7 @@ void btrfs_auto_defrag_exit(void); int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans, struct inode *inode); int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info); +void btrfs_cleanup_defrag_inodes(struct btrfs_fs_info *fs_info); int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync); void btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end, int skip_pinned); diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 76b82506bf9..3229531af8c 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -3329,7 +3329,7 @@ int close_ctree(struct btrfs_root *root) (atomic_read(&fs_info->defrag_running) == 0)); /* clear out the rbtree of defraggable inodes */ - btrfs_run_defrag_inodes(fs_info); + btrfs_cleanup_defrag_inodes(fs_info); if (!(fs_info->sb->s_flags & MS_RDONLY)) { ret = btrfs_commit_super(root); diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 00918321e39..3c6f7479cd5 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -216,11 +216,11 @@ out: } /* - * must be called with the defrag_inodes lock held + * pick the defragable inode that we want, if it doesn't exist, we will get + * the next one. */ -struct inode_defrag *btrfs_find_defrag_inode(struct btrfs_fs_info *info, - u64 root, u64 ino, - struct rb_node **next) +static struct inode_defrag * +btrfs_pick_defrag_inode(struct btrfs_fs_info *fs_info, u64 root, u64 ino) { struct inode_defrag *entry = NULL; struct inode_defrag tmp; @@ -231,7 +231,8 @@ struct inode_defrag *btrfs_find_defrag_inode(struct btrfs_fs_info *info, tmp.ino = ino; tmp.root = root; - p = info->defrag_inodes.rb_node; + spin_lock(&fs_info->defrag_inodes_lock); + p = fs_info->defrag_inodes.rb_node; while (p) { parent = p; entry = rb_entry(parent, struct inode_defrag, rb_node); @@ -242,52 +243,128 @@ struct inode_defrag *btrfs_find_defrag_inode(struct btrfs_fs_info *info, else if (ret > 0) p = parent->rb_right; else - return entry; + goto out; } - if (next) { - while (parent && __compare_inode_defrag(&tmp, entry) > 0) { - parent = rb_next(parent); + if (parent && __compare_inode_defrag(&tmp, entry) > 0) { + parent = rb_next(parent); + if (parent) entry = rb_entry(parent, struct inode_defrag, rb_node); - } - *next = parent; + else + entry = NULL; } - return NULL; +out: + if (entry) + rb_erase(parent, &fs_info->defrag_inodes); + spin_unlock(&fs_info->defrag_inodes_lock); + return entry; } -/* - * run through the list of inodes in the FS that need - * defragging - */ -int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) +void btrfs_cleanup_defrag_inodes(struct btrfs_fs_info *fs_info) { struct inode_defrag *defrag; + struct rb_node *node; + + spin_lock(&fs_info->defrag_inodes_lock); + node = rb_first(&fs_info->defrag_inodes); + while (node) { + rb_erase(node, &fs_info->defrag_inodes); + defrag = rb_entry(node, struct inode_defrag, rb_node); + kmem_cache_free(btrfs_inode_defrag_cachep, defrag); + + if (need_resched()) { + spin_unlock(&fs_info->defrag_inodes_lock); + cond_resched(); + spin_lock(&fs_info->defrag_inodes_lock); + } + + node = rb_first(&fs_info->defrag_inodes); + } + spin_unlock(&fs_info->defrag_inodes_lock); +} + +#define BTRFS_DEFRAG_BATCH 1024 + +static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info, + struct inode_defrag *defrag) +{ struct btrfs_root *inode_root; struct inode *inode; - struct rb_node *n; struct btrfs_key key; struct btrfs_ioctl_defrag_range_args range; - u64 first_ino = 0; - u64 root_objectid = 0; int num_defrag; - int defrag_batch = 1024; + /* get the inode */ + key.objectid = defrag->root; + btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY); + key.offset = (u64)-1; + inode_root = btrfs_read_fs_root_no_name(fs_info, &key); + if (IS_ERR(inode_root)) { + kmem_cache_free(btrfs_inode_defrag_cachep, defrag); + return PTR_ERR(inode_root); + } + + key.objectid = defrag->ino; + btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY); + key.offset = 0; + inode = btrfs_iget(fs_info->sb, &key, inode_root, NULL); + if (IS_ERR(inode)) { + kmem_cache_free(btrfs_inode_defrag_cachep, defrag); + return PTR_ERR(inode); + } + + /* do a chunk of defrag */ + clear_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags); memset(&range, 0, sizeof(range)); range.len = (u64)-1; + range.start = defrag->last_offset; + num_defrag = btrfs_defrag_file(inode, NULL, &range, defrag->transid, + BTRFS_DEFRAG_BATCH); + /* + * if we filled the whole defrag batch, there + * must be more work to do. Queue this defrag + * again + */ + if (num_defrag == BTRFS_DEFRAG_BATCH) { + defrag->last_offset = range.start; + btrfs_requeue_inode_defrag(inode, defrag); + } else if (defrag->last_offset && !defrag->cycled) { + /* + * we didn't fill our defrag batch, but + * we didn't start at zero. Make sure we loop + * around to the start of the file. + */ + defrag->last_offset = 0; + defrag->cycled = 1; + btrfs_requeue_inode_defrag(inode, defrag); + } else { + kmem_cache_free(btrfs_inode_defrag_cachep, defrag); + } + + iput(inode); + return 0; +} + +/* + * run through the list of inodes in the FS that need + * defragging + */ +int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) +{ + struct inode_defrag *defrag; + u64 first_ino = 0; + u64 root_objectid = 0; atomic_inc(&fs_info->defrag_running); - spin_lock(&fs_info->defrag_inodes_lock); while(1) { - n = NULL; + if (!__need_auto_defrag(fs_info->tree_root)) + break; /* find an inode to defrag */ - defrag = btrfs_find_defrag_inode(fs_info, root_objectid, - first_ino, &n); + defrag = btrfs_pick_defrag_inode(fs_info, root_objectid, + first_ino); if (!defrag) { - if (n) { - defrag = rb_entry(n, struct inode_defrag, - rb_node); - } else if (root_objectid || first_ino) { + if (root_objectid || first_ino) { root_objectid = 0; first_ino = 0; continue; @@ -296,71 +373,11 @@ int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info) } } - /* remove it from the rbtree */ first_ino = defrag->ino + 1; root_objectid = defrag->root; - rb_erase(&defrag->rb_node, &fs_info->defrag_inodes); - - if (btrfs_fs_closing(fs_info)) - goto next_free; - spin_unlock(&fs_info->defrag_inodes_lock); - - /* get the inode */ - key.objectid = defrag->root; - btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY); - key.offset = (u64)-1; - inode_root = btrfs_read_fs_root_no_name(fs_info, &key); - if (IS_ERR(inode_root)) - goto next; - - key.objectid = defrag->ino; - btrfs_set_key_type(&key, BTRFS_INODE_ITEM_KEY); - key.offset = 0; - - inode = btrfs_iget(fs_info->sb, &key, inode_root, NULL); - if (IS_ERR(inode)) - goto next; - - /* do a chunk of defrag */ - clear_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags); - range.start = defrag->last_offset; - num_defrag = btrfs_defrag_file(inode, NULL, &range, defrag->transid, - defrag_batch); - /* - * if we filled the whole defrag batch, there - * must be more work to do. Queue this defrag - * again - */ - if (num_defrag == defrag_batch) { - defrag->last_offset = range.start; - btrfs_requeue_inode_defrag(inode, defrag); - /* - * we don't want to kfree defrag, we added it back to - * the rbtree - */ - defrag = NULL; - } else if (defrag->last_offset && !defrag->cycled) { - /* - * we didn't fill our defrag batch, but - * we didn't start at zero. Make sure we loop - * around to the start of the file. - */ - defrag->last_offset = 0; - defrag->cycled = 1; - btrfs_requeue_inode_defrag(inode, defrag); - defrag = NULL; - } - - iput(inode); -next: - spin_lock(&fs_info->defrag_inodes_lock); -next_free: - if (defrag) - kmem_cache_free(btrfs_inode_defrag_cachep, defrag); + __btrfs_run_defrag_inode(fs_info, defrag); } - spin_unlock(&fs_info->defrag_inodes_lock); - atomic_dec(&fs_info->defrag_running); /* -- cgit v1.2.3-70-g09d2 From 5f3ab90a72f98adbf00c50ac2d4d2b47cf4a9685 Mon Sep 17 00:00:00 2001 From: Anand Jain Date: Fri, 7 Dec 2012 09:28:54 +0000 Subject: Btrfs: rename root_times_lock to root_item_lock Originally root_times_lock was introduced as part of send/receive code however newly developed patch to label the subvol reused the same lock, so renaming it for a meaningful name. Signed-off-by: Anand Jain Signed-off-by: Chris Mason --- fs/btrfs/ctree.c | 16 ++++++++-------- fs/btrfs/ctree.h | 2 +- fs/btrfs/disk-io.c | 2 +- fs/btrfs/root-tree.c | 4 ++-- fs/btrfs/send.c | 8 ++++---- 5 files changed, 16 insertions(+), 16 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c index 5c2cf992e71..01efcbc80df 100644 --- a/fs/btrfs/ctree.c +++ b/fs/btrfs/ctree.c @@ -5114,13 +5114,13 @@ int btrfs_compare_trees(struct btrfs_root *left_root, right_path->search_commit_root = 1; right_path->skip_locking = 1; - spin_lock(&left_root->root_times_lock); + spin_lock(&left_root->root_item_lock); left_start_ctransid = btrfs_root_ctransid(&left_root->root_item); - spin_unlock(&left_root->root_times_lock); + spin_unlock(&left_root->root_item_lock); - spin_lock(&right_root->root_times_lock); + spin_lock(&right_root->root_item_lock); right_start_ctransid = btrfs_root_ctransid(&right_root->root_item); - spin_unlock(&right_root->root_times_lock); + spin_unlock(&right_root->root_item_lock); trans = btrfs_join_transaction(left_root); if (IS_ERR(trans)) { @@ -5215,15 +5215,15 @@ int btrfs_compare_trees(struct btrfs_root *left_root, goto out; } - spin_lock(&left_root->root_times_lock); + spin_lock(&left_root->root_item_lock); ctransid = btrfs_root_ctransid(&left_root->root_item); - spin_unlock(&left_root->root_times_lock); + spin_unlock(&left_root->root_item_lock); if (ctransid != left_start_ctransid) left_start_ctransid = 0; - spin_lock(&right_root->root_times_lock); + spin_lock(&right_root->root_item_lock); ctransid = btrfs_root_ctransid(&right_root->root_item); - spin_unlock(&right_root->root_times_lock); + spin_unlock(&right_root->root_item_lock); if (ctransid != right_start_ctransid) right_start_ctransid = 0; diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h index 6ba56aea5b6..313a6adfde5 100644 --- a/fs/btrfs/ctree.h +++ b/fs/btrfs/ctree.h @@ -1645,7 +1645,7 @@ struct btrfs_root { int force_cow; - spinlock_t root_times_lock; + spinlock_t root_item_lock; }; struct btrfs_ioctl_defrag_range_args { diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 3229531af8c..faf182691b4 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1204,7 +1204,7 @@ static void __setup_root(u32 nodesize, u32 leafsize, u32 sectorsize, root->root_key.objectid = objectid; root->anon_dev = 0; - spin_lock_init(&root->root_times_lock); + spin_lock_init(&root->root_item_lock); } static int __must_check find_and_setup_root(struct btrfs_root *tree_root, diff --git a/fs/btrfs/root-tree.c b/fs/btrfs/root-tree.c index eb923d087da..668af537a3e 100644 --- a/fs/btrfs/root-tree.c +++ b/fs/btrfs/root-tree.c @@ -548,9 +548,9 @@ void btrfs_update_root_times(struct btrfs_trans_handle *trans, struct btrfs_root_item *item = &root->root_item; struct timespec ct = CURRENT_TIME; - spin_lock(&root->root_times_lock); + spin_lock(&root->root_item_lock); item->ctransid = cpu_to_le64(trans->transid); item->ctime.sec = cpu_to_le64(ct.tv_sec); item->ctime.nsec = cpu_to_le32(ct.tv_nsec); - spin_unlock(&root->root_times_lock); + spin_unlock(&root->root_item_lock); } diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index e78b297b0b0..54454542ad4 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -4397,9 +4397,9 @@ static int full_send_tree(struct send_ctx *sctx) if (!path) return -ENOMEM; - spin_lock(&send_root->root_times_lock); + spin_lock(&send_root->root_item_lock); start_ctransid = btrfs_root_ctransid(&send_root->root_item); - spin_unlock(&send_root->root_times_lock); + spin_unlock(&send_root->root_item_lock); key.objectid = BTRFS_FIRST_FREE_OBJECTID; key.type = BTRFS_INODE_ITEM_KEY; @@ -4422,9 +4422,9 @@ join_trans: * Make sure the tree has not changed after re-joining. We detect this * by comparing start_ctransid and ctransid. They should always match. */ - spin_lock(&send_root->root_times_lock); + spin_lock(&send_root->root_item_lock); ctransid = btrfs_root_ctransid(&send_root->root_item); - spin_unlock(&send_root->root_times_lock); + spin_unlock(&send_root->root_item_lock); if (ctransid != start_ctransid) { WARN(1, KERN_WARNING "btrfs: the root that you're trying to " -- cgit v1.2.3-70-g09d2 From bb146eb265091f472ada52a3419d41e9b0ff1f7d Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 15 Oct 2012 13:30:43 -0400 Subject: Btrfs: move checks in set_page_dirty under DEBUG This is a high traffic function, let's try and do as little as possible during normal operations shall we? Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index faf182691b4..b8f7f04a640 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1001,6 +1001,7 @@ static void btree_invalidatepage(struct page *page, unsigned long offset) static int btree_set_page_dirty(struct page *page) { +#ifdef DEBUG struct extent_buffer *eb; BUG_ON(!PagePrivate(page)); @@ -1009,6 +1010,7 @@ static int btree_set_page_dirty(struct page *page) BUG_ON(!test_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)); BUG_ON(!atomic_read(&eb->refs)); btrfs_assert_tree_locked(eb); +#endif return __set_page_dirty_nobuffers(page); } -- cgit v1.2.3-70-g09d2 From ed7b63eb8afd0bb8d978a23184d70c105b54aa26 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 15 Oct 2012 13:33:54 -0400 Subject: Btrfs: only clear dirty on the buffer if it is marked as dirty No reason to set the path blocking or loop through all of the pages if the extent buffer isn't actually marked dirty. Thanks, Signed-off-by: Josef Bacik Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index b8f7f04a640..65f03670a95 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1142,11 +1142,11 @@ void clean_tree_block(struct btrfs_trans_handle *trans, struct btrfs_root *root, root->fs_info->dirty_metadata_bytes); } spin_unlock(&root->fs_info->delalloc_lock); - } - /* ugh, clear_extent_buffer_dirty needs to lock the page */ - btrfs_set_lock_blocking(buf); - clear_extent_buffer_dirty(buf); + /* ugh, clear_extent_buffer_dirty needs to lock the page */ + btrfs_set_lock_blocking(buf); + clear_extent_buffer_dirty(buf); + } } } -- cgit v1.2.3-70-g09d2