Merge branch 'for-linus-4.3' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

Pull btrfs fixes from Chris Mason:
 "This is an assorted set I've been queuing up:

  Jeff Mahoney tracked down a tricky one where we ended up starting IO
  on the wrong mapping for special files in btrfs_evict_inode.  A few
  people reported this one on the list.

  Filipe found (and provided a test for) a difficult bug in reading
  compressed extents, and Josef fixed up some quota record keeping with
  snapshot deletion.  Chandan killed off an accounting bug during DIO
  that lead to WARN_ONs as we freed inodes"

* 'for-linus-4.3' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Btrfs: keep dropped roots in cache until transaction commit
  Btrfs: Direct I/O: Fix space accounting
  btrfs: skip waiting on ordered range for special files
  Btrfs: fix read corruption of compressed and shared extents
  Btrfs: remove unnecessary locking of cleaner_mutex to avoid deadlock
  Btrfs: don't initialize a space info as full to prevent ENOSPC

8 files changed, 118 insertions, 42 deletions

diff --git a/fs/btrfs/btrfs_inode.h b/fs/btrfs/btrfs_inode.h
index 81220b2..0ef5cc1 100644
--- a/fs/btrfs/btrfs_inode.h
+++ b/fs/btrfs/btrfs_inode.h
@@ -44,8 +44,6 @@
 #define BTRFS_INODE_IN_DELALLOC_LIST		9
 #define BTRFS_INODE_READDIO_NEED_LOCK		10
 #define BTRFS_INODE_HAS_PROPS		        11
-/* DIO is ready to submit */
-#define BTRFS_INODE_DIO_READY		        12
 /*
  * The following 3 bits are meant only for the btree inode.
  * When any of them is set, it means an error happened while writing an
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 0d98aee..295795a 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -3765,9 +3765,7 @@ void close_ctree(struct btrfs_root *root)
 		 * block groups queued for removal, the deletion will be
 		 * skipped when we quit the cleaner thread.
 		 */
-		mutex_lock(&root->fs_info->cleaner_mutex);
 		btrfs_delete_unused_bgs(root->fs_info);
-		mutex_unlock(&root->fs_info->cleaner_mutex);
 
 		ret = btrfs_commit_super(root);
 		if (ret)
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 5411f0a..9f96042 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -3742,10 +3742,7 @@ static int update_space_info(struct btrfs_fs_info *info, u64 flags,
 	found->bytes_reserved = 0;
 	found->bytes_readonly = 0;
 	found->bytes_may_use = 0;
-	if (total_bytes > 0)
-		found->full = 0;
-	else
-		found->full = 1;
+	found->full = 0;
 	found->force_alloc = CHUNK_ALLOC_NO_FORCE;
 	found->chunk_alloc = 0;
 	found->flush = 0;
@@ -8668,7 +8665,7 @@ int btrfs_drop_snapshot(struct btrfs_root *root,
 	}
 
 	if (test_bit(BTRFS_ROOT_IN_RADIX, &root->state)) {
-		btrfs_drop_and_free_fs_root(tree_root->fs_info, root);
+		btrfs_add_dropped_root(trans, root);
 	} else {
 		free_extent_buffer(root->node);
 		free_extent_buffer(root->commit_root);
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index f1018cf..e2357e3 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -2798,7 +2798,8 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree,
 			      bio_end_io_t end_io_func,
 			      int mirror_num,
 			      unsigned long prev_bio_flags,
-			      unsigned long bio_flags)
+			      unsigned long bio_flags,
+			      bool force_bio_submit)
 {
 	int ret = 0;
 	struct bio *bio;
@@ -2814,6 +2815,7 @@ static int submit_extent_page(int rw, struct extent_io_tree *tree,
 			contig = bio_end_sector(bio) == sector;
 
 		if (prev_bio_flags != bio_flags || !contig ||
+		    force_bio_submit ||
 		    merge_bio(rw, tree, page, offset, page_size, bio, bio_flags) ||
 		    bio_add_page(bio, page, page_size, offset) < page_size) {
 			ret = submit_one_bio(rw, bio, mirror_num,
@@ -2910,7 +2912,8 @@ static int __do_readpage(struct extent_io_tree *tree,
 			 get_extent_t *get_extent,
 			 struct extent_map **em_cached,
 			 struct bio **bio, int mirror_num,
-			 unsigned long *bio_flags, int rw)
+			 unsigned long *bio_flags, int rw,
+			 u64 *prev_em_start)
 {
 	struct inode *inode = page->mapping->host;
 	u64 start = page_offset(page);
@@ -2958,6 +2961,7 @@ static int __do_readpage(struct extent_io_tree *tree,
 	}
 	while (cur <= end) {
 		unsigned long pnr = (last_byte >> PAGE_CACHE_SHIFT) + 1;
+		bool force_bio_submit = false;
 
 		if (cur >= last_byte) {
 			char *userpage;
@@ -3008,6 +3012,49 @@ static int __do_readpage(struct extent_io_tree *tree,
 		block_start = em->block_start;
 		if (test_bit(EXTENT_FLAG_PREALLOC, &em->flags))
 			block_start = EXTENT_MAP_HOLE;
+
+		/*
+		 * If we have a file range that points to a compressed extent
+		 * and it's followed by a consecutive file range that points to
+		 * to the same compressed extent (possibly with a different
+		 * offset and/or length, so it either points to the whole extent
+		 * or only part of it), we must make sure we do not submit a
+		 * single bio to populate the pages for the 2 ranges because
+		 * this makes the compressed extent read zero out the pages
+		 * belonging to the 2nd range. Imagine the following scenario:
+		 *
+		 *  File layout
+		 *  [0 - 8K]                     [8K - 24K]
+		 *    |                               |
+		 *    |                               |
+		 * points to extent X,         points to extent X,
+		 * offset 4K, length of 8K     offset 0, length 16K
+		 *
+		 * [extent X, compressed length = 4K uncompressed length = 16K]
+		 *
+		 * If the bio to read the compressed extent covers both ranges,
+		 * it will decompress extent X into the pages belonging to the
+		 * first range and then it will stop, zeroing out the remaining
+		 * pages that belong to the other range that points to extent X.
+		 * So here we make sure we submit 2 bios, one for the first
+		 * range and another one for the third range. Both will target
+		 * the same physical extent from disk, but we can't currently
+		 * make the compressed bio endio callback populate the pages
+		 * for both ranges because each compressed bio is tightly
+		 * coupled with a single extent map, and each range can have
+		 * an extent map with a different offset value relative to the
+		 * uncompressed data of our extent and different lengths. This
+		 * is a corner case so we prioritize correctness over
+		 * non-optimal behavior (submitting 2 bios for the same extent).
+		 */
+		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags) &&
+		    prev_em_start && *prev_em_start != (u64)-1 &&
+		    *prev_em_start != em->orig_start)
+			force_bio_submit = true;
+
+		if (prev_em_start)
+			*prev_em_start = em->orig_start;
+
 		free_extent_map(em);
 		em = NULL;
 
@@ -3057,7 +3104,8 @@ static int __do_readpage(struct extent_io_tree *tree,
 					 bdev, bio, pnr,
 					 end_bio_extent_readpage, mirror_num,
 					 *bio_flags,
-					 this_bio_flag);
+					 this_bio_flag,
+					 force_bio_submit);
 		if (!ret) {
 			nr++;
 			*bio_flags = this_bio_flag;
@@ -3089,6 +3137,7 @@ static inline void __do_contiguous_readpages(struct extent_io_tree *tree,
 	struct inode *inode;
 	struct btrfs_ordered_extent *ordered;
 	int index;
+	u64 prev_em_start = (u64)-1;
 
 	inode = pages[0]->mapping->host;
 	while (1) {
@@ -3104,7 +3153,7 @@ static inline void __do_contiguous_readpages(struct extent_io_tree *tree,
 
 	for (index = 0; index < nr_pages; index++) {
 		__do_readpage(tree, pages[index], get_extent, em_cached, bio,
-			      mirror_num, bio_flags, rw);
+			      mirror_num, bio_flags, rw, &prev_em_start);
 		page_cache_release(pages[index]);
 	}
 }
@@ -3172,7 +3221,7 @@ static int __extent_read_full_page(struct extent_io_tree *tree,
 	}
 
 	ret = __do_readpage(tree, page, get_extent, NULL, bio, mirror_num,
-			    bio_flags, rw);
+			    bio_flags, rw, NULL);
 	return ret;
 }
 
@@ -3198,7 +3247,7 @@ int extent_read_full_page_nolock(struct extent_io_tree *tree, struct page *page,
 	int ret;
 
 	ret = __do_readpage(tree, page, get_extent, NULL, &bio, mirror_num,
-				      &bio_flags, READ);
+			    &bio_flags, READ, NULL);
 	if (bio)
 		ret = submit_one_bio(READ, bio, mirror_num, bio_flags);
 	return ret;
@@ -3451,7 +3500,7 @@ static noinline_for_stack int __extent_writepage_io(struct inode *inode,
 						 sector, iosize, pg_offset,
 						 bdev, &epd->bio, max_nr,
 						 end_bio_extent_writepage,
-						 0, 0, 0);
+						 0, 0, 0, false);
 			if (ret)
 				SetPageError(page);
 		}
@@ -3754,7 +3803,7 @@ static noinline_for_stack int write_one_eb(struct extent_buffer *eb,
 		ret = submit_extent_page(rw, tree, wbc, p, offset >> 9,
 					 PAGE_CACHE_SIZE, 0, bdev, &epd->bio,
 					 -1, end_bio_extent_buffer_writepage,
-					 0, epd->bio_flags, bio_flags);
+					 0, epd->bio_flags, bio_flags, false);
 		epd->bio_flags = bio_flags;
 		if (ret) {
 			set_btree_ioerr(p);
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index a0fa725..611b66d 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -5084,7 +5084,8 @@ void btrfs_evict_inode(struct inode *inode)
 		goto no_delete;
 	}
 	/* do we really want it for ->i_nlink > 0 and zero btrfs_root_refs? */
-	btrfs_wait_ordered_range(inode, 0, (u64)-1);
+	if (!special_file(inode->i_mode))
+		btrfs_wait_ordered_range(inode, 0, (u64)-1);
 
 	btrfs_free_io_failure_record(inode, 0, (u64)-1);
 
@@ -7408,6 +7409,10 @@ static struct extent_map *create_pinned_em(struct inode *inode, u64 start,
 	return em;
 }
 
+struct btrfs_dio_data {
+	u64 outstanding_extents;
+	u64 reserve;
+};
 
 static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 				   struct buffer_head *bh_result, int create)
@@ -7415,10 +7420,10 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 	struct extent_map *em;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct extent_state *cached_state = NULL;
+	struct btrfs_dio_data *dio_data = NULL;
 	u64 start = iblock << inode->i_blkbits;
 	u64 lockstart, lockend;
 	u64 len = bh_result->b_size;
-	u64 *outstanding_extents = NULL;
 	int unlock_bits = EXTENT_LOCKED;
 	int ret = 0;
 
@@ -7436,7 +7441,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 		 * that anything that needs to check if there's a transction doesn't get
 		 * confused.
 		 */
-		outstanding_extents = current->journal_info;
+		dio_data = current->journal_info;
 		current->journal_info = NULL;
 	}
 
@@ -7568,17 +7573,18 @@ unlock:
 		 * within our reservation, otherwise we need to adjust our inode
 		 * counter appropriately.
 		 */
-		if (*outstanding_extents) {
-			(*outstanding_extents)--;
+		if (dio_data->outstanding_extents) {
+			(dio_data->outstanding_extents)--;
 		} else {
 			spin_lock(&BTRFS_I(inode)->lock);
 			BTRFS_I(inode)->outstanding_extents++;
 			spin_unlock(&BTRFS_I(inode)->lock);
 		}
 
-		current->journal_info = outstanding_extents;
 		btrfs_free_reserved_data_space(inode, len);
-		set_bit(BTRFS_INODE_DIO_READY, &BTRFS_I(inode)->runtime_flags);
+		WARN_ON(dio_data->reserve < len);
+		dio_data->reserve -= len;
+		current->journal_info = dio_data;
 	}
 
 	/*
@@ -7601,8 +7607,8 @@ unlock:
 unlock_err:
 	clear_extent_bit(&BTRFS_I(inode)->io_tree, lockstart, lockend,
 			 unlock_bits, 1, 0, &cached_state, GFP_NOFS);
-	if (outstanding_extents)
-		current->journal_info = outstanding_extents;
+	if (dio_data)
+		current->journal_info = dio_data;
 	return ret;
 }
 
@@ -8329,7 +8335,8 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
 {
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file->f_mapping->host;
-	u64 outstanding_extents = 0;
+	struct btrfs_root *root = BTRFS_I(inode)->root;
+	struct btrfs_dio_data dio_data = { 0 };
 	size_t count = 0;
 	int flags = 0;
 	bool wakeup = true;
@@ -8367,7 +8374,7 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
 		ret = btrfs_delalloc_reserve_space(inode, count);
 		if (ret)
 			goto out;
-		outstanding_extents = div64_u64(count +
+		dio_data.outstanding_extents = div64_u64(count +
 						BTRFS_MAX_EXTENT_SIZE - 1,
 						BTRFS_MAX_EXTENT_SIZE);
 
@@ -8376,7 +8383,8 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
 		 * do the accounting properly if we go over the number we
 		 * originally calculated.  Abuse current->journal_info for this.
 		 */
-		current->journal_info = &outstanding_extents;
+		dio_data.reserve = round_up(count, root->sectorsize);
+		current->journal_info = &dio_data;
 	} else if (test_bit(BTRFS_INODE_READDIO_NEED_LOCK,
 				     &BTRFS_I(inode)->runtime_flags)) {
 		inode_dio_end(inode);
@@ -8391,16 +8399,9 @@ static ssize_t btrfs_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
 	if (iov_iter_rw(iter) == WRITE) {
 		current->journal_info = NULL;
 		if (ret < 0 && ret != -EIOCBQUEUED) {
-			/*
-			 * If the error comes from submitting stage,
-			 * btrfs_get_blocsk_direct() has free'd data space,
-			 * and metadata space will be handled by
-			 * finish_ordered_fn, don't do that again to make
-			 * sure bytes_may_use is correct.
-			 */
-			if (!test_and_clear_bit(BTRFS_INODE_DIO_READY,
-				     &BTRFS_I(inode)->runtime_flags))
-				btrfs_delalloc_release_space(inode, count);
+			if (dio_data.reserve)
+				btrfs_delalloc_release_space(inode,
+							dio_data.reserve);
 		} else if (ret >= 0 && (size_t)ret < count)
 			btrfs_delalloc_release_space(inode,
 						     count - (size_t)ret);
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 2b07b35..11d1eab 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -1658,9 +1658,7 @@ static int btrfs_remount(struct super_block *sb, int *flags, char *data)
 		 * groups on disk until we're mounted read-write again
 		 * unless we clean them up here.
 		 */
-		mutex_lock(&root->fs_info->cleaner_mutex);
 		btrfs_delete_unused_bgs(fs_info);
-		mutex_unlock(&root->fs_info->cleaner_mutex);
 
 		btrfs_dev_replace_suspend_for_unmount(fs_info);
 		btrfs_scrub_cancel(fs_info);
diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 8f259b3..74bc333 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -117,6 +117,18 @@ static noinline void switch_commit_roots(struct btrfs_transaction *trans,
 			btrfs_unpin_free_ino(root);
 		clear_btree_io_tree(&root->dirty_log_pages);
 	}
+
+	/* We can free old roots now. */
+	spin_lock(&trans->dropped_roots_lock);
+	while (!list_empty(&trans->dropped_roots)) {
+		root = list_first_entry(&trans->dropped_roots,
+					struct btrfs_root, root_list);
+		list_del_init(&root->root_list);
+		spin_unlock(&trans->dropped_roots_lock);
+		btrfs_drop_and_free_fs_root(fs_info, root);
+		spin_lock(&trans->dropped_roots_lock);
+	}
+	spin_unlock(&trans->dropped_roots_lock);
 	up_write(&fs_info->commit_root_sem);
 }
 
@@ -255,11 +267,13 @@ loop:
 	INIT_LIST_HEAD(&cur_trans->pending_ordered);
 	INIT_LIST_HEAD(&cur_trans->dirty_bgs);
 	INIT_LIST_HEAD(&cur_trans->io_bgs);
+	INIT_LIST_HEAD(&cur_trans->dropped_roots);
 	mutex_init(&cur_trans->cache_write_mutex);
 	cur_trans->num_dirty_bgs = 0;
 	spin_lock_init(&cur_trans->dirty_bgs_lock);
 	INIT_LIST_HEAD(&cur_trans->deleted_bgs);
 	spin_lock_init(&cur_trans->deleted_bgs_lock);
+	spin_lock_init(&cur_trans->dropped_roots_lock);
 	list_add_tail(&cur_trans->list, &fs_info->trans_list);
 	extent_io_tree_init(&cur_trans->dirty_pages,
 			     fs_info->btree_inode->i_mapping);
@@ -336,6 +350,24 @@ static int record_root_in_trans(struct btrfs_trans_handle *trans,
 }
 
 
+void btrfs_add_dropped_root(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root)
+{
+	struct btrfs_transaction *cur_trans = trans->transaction;
+
+	/* Add ourselves to the transaction dropped list */
+	spin_lock(&cur_trans->dropped_roots_lock);
+	list_add_tail(&root->root_list, &cur_trans->dropped_roots);
+	spin_unlock(&cur_trans->dropped_roots_lock);
+
+	/* Make sure we don't try to update the root at commit time */
+	spin_lock(&root->fs_info->fs_roots_radix_lock);
+	radix_tree_tag_clear(&root->fs_info->fs_roots_radix,
+			     (unsigned long)root->root_key.objectid,
+			     BTRFS_ROOT_TRANS_TAG);
+	spin_unlock(&root->fs_info->fs_roots_radix_lock);
+}
+
 int btrfs_record_root_in_trans(struct btrfs_trans_handle *trans,
 			       struct btrfs_root *root)
 {
diff --git a/fs/btrfs/transaction.h b/fs/btrfs/transaction.h
index edc2fbc..87964bf 100644
--- a/fs/btrfs/transaction.h
+++ b/fs/btrfs/transaction.h
@@ -65,6 +65,7 @@ struct btrfs_transaction {
 	struct list_head switch_commits;
 	struct list_head dirty_bgs;
 	struct list_head io_bgs;
+	struct list_head dropped_roots;
 	u64 num_dirty_bgs;
 
 	/*
@@ -76,6 +77,7 @@ struct btrfs_transaction {
 	spinlock_t dirty_bgs_lock;
 	struct list_head deleted_bgs;
 	spinlock_t deleted_bgs_lock;
+	spinlock_t dropped_roots_lock;
 	struct btrfs_delayed_ref_root delayed_refs;
 	int aborted;
 	int dirty_bg_run;
@@ -216,5 +218,6 @@ int btrfs_transaction_blocked(struct btrfs_fs_info *info);
 int btrfs_transaction_in_commit(struct btrfs_fs_info *info);
 void btrfs_put_transaction(struct btrfs_transaction *transaction);
 void btrfs_apply_pending_changes(struct btrfs_fs_info *fs_info);
-
+void btrfs_add_dropped_root(struct btrfs_trans_handle *trans,
+			    struct btrfs_root *root);
 #endif