1 files changed, 76 insertions, 35 deletions
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 1fd827b99c1b..30e6aef9739f 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -1848,7 +1848,6 @@ static ssize_t check_direct_IO(struct btrfs_fs_info *fs_info,
 
 static ssize_t btrfs_direct_write(struct kiocb *iocb, struct iov_iter *from)
 {
-	const bool is_sync_write = (iocb->ki_flags & IOCB_DSYNC);
 	struct file *file = iocb->ki_filp;
 	struct inode *inode = file_inode(file);
 	struct btrfs_fs_info *fs_info = btrfs_sb(inode->i_sb);
@@ -1902,15 +1901,6 @@ relock:
 	}
 
 	/*
-	 * We remove IOCB_DSYNC so that we don't deadlock when iomap_dio_rw()
-	 * calls generic_write_sync() (through iomap_dio_complete()), because
-	 * that results in calling fsync (btrfs_sync_file()) which will try to
-	 * lock the inode in exclusive/write mode.
-	 */
-	if (is_sync_write)
-		iocb->ki_flags &= ~IOCB_DSYNC;
-
-	/*
 	 * The iov_iter can be mapped to the same file range we are writing to.
 	 * If that's the case, then we will deadlock in the iomap code, because
 	 * it first calls our callback btrfs_dio_iomap_begin(), which will create
@@ -1964,13 +1954,6 @@ again:
 
 	btrfs_inode_unlock(inode, ilock_flags);
 
-	/*
-	 * Add back IOCB_DSYNC. Our caller, btrfs_file_write_iter(), will do
-	 * the fsync (call generic_write_sync()).
-	 */
-	if (is_sync_write)
-		iocb->ki_flags |= IOCB_DSYNC;
-
 	/* If 'err' is -ENOTBLK then it means we must fallback to buffered IO. */
 	if ((err < 0 && err != -ENOTBLK) || !iov_iter_count(from))
 		goto out;
@@ -2038,7 +2021,7 @@ ssize_t btrfs_do_write_iter(struct kiocb *iocb, struct iov_iter *from,
 	struct file *file = iocb->ki_filp;
 	struct btrfs_inode *inode = BTRFS_I(file_inode(file));
 	ssize_t num_written, num_sync;
-	const bool sync = iocb->ki_flags & IOCB_DSYNC;
+	const bool sync = iocb_is_dsync(iocb);
 
 	/*
 	 * If the fs flips readonly due to some impossible error, although we
@@ -2323,25 +2306,62 @@ int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
 	 */
 	btrfs_inode_unlock(inode, BTRFS_ILOCK_MMAP);
 
-	if (ret != BTRFS_NO_LOG_SYNC) {
+	if (ret == BTRFS_NO_LOG_SYNC) {
+		ret = btrfs_end_transaction(trans);
+		goto out;
+	}
+
+	/* We successfully logged the inode, attempt to sync the log. */
+	if (!ret) {
+		ret = btrfs_sync_log(trans, root, &ctx);
 		if (!ret) {
-			ret = btrfs_sync_log(trans, root, &ctx);
-			if (!ret) {
-				ret = btrfs_end_transaction(trans);
-				goto out;
-			}
-		}
-		if (!full_sync) {
-			ret = btrfs_wait_ordered_range(inode, start, len);
-			if (ret) {
-				btrfs_end_transaction(trans);
-				goto out;
-			}
+			ret = btrfs_end_transaction(trans);
+			goto out;
 		}
-		ret = btrfs_commit_transaction(trans);
-	} else {
+	}
+
+	/*
+	 * At this point we need to commit the transaction because we had
+	 * btrfs_need_log_full_commit() or some other error.
+	 *
+	 * If we didn't do a full sync we have to stop the trans handle, wait on
+	 * the ordered extents, start it again and commit the transaction.  If
+	 * we attempt to wait on the ordered extents here we could deadlock with
+	 * something like fallocate() that is holding the extent lock trying to
+	 * start a transaction while some other thread is trying to commit the
+	 * transaction while we (fsync) are currently holding the transaction
+	 * open.
+	 */
+	if (!full_sync) {
 		ret = btrfs_end_transaction(trans);
+		if (ret)
+			goto out;
+		ret = btrfs_wait_ordered_range(inode, start, len);
+		if (ret)
+			goto out;
+
+		/*
+		 * This is safe to use here because we're only interested in
+		 * making sure the transaction that had the ordered extents is
+		 * committed.  We aren't waiting on anything past this point,
+		 * we're purely getting the transaction and committing it.
+		 */
+		trans = btrfs_attach_transaction_barrier(root);
+		if (IS_ERR(trans)) {
+			ret = PTR_ERR(trans);
+
+			/*
+			 * We committed the transaction and there's no currently
+			 * running transaction, this means everything we care
+			 * about made it to disk and we are done.
+			 */
+			if (ret == -ENOENT)
+				ret = 0;
+			goto out;
+		}
 	}
+
+	ret = btrfs_commit_transaction(trans);
 out:
 	ASSERT(list_empty(&ctx.list));
 	err = file_check_and_advance_wb_err(file);
@@ -2719,7 +2739,8 @@ int btrfs_replace_file_extents(struct btrfs_inode *inode,
 
 	ret = btrfs_block_rsv_migrate(&fs_info->trans_block_rsv, rsv,
 				      min_size, false);
-	BUG_ON(ret);
+	if (WARN_ON(ret))
+		goto out_trans;
 	trans->block_rsv = rsv;
 
 	cur_offset = start;
@@ -2803,6 +2824,25 @@ int btrfs_replace_file_extents(struct btrfs_inode *inode,
 			extent_info->file_offset += replace_len;
 		}
 
+		/*
+		 * We are releasing our handle on the transaction, balance the
+		 * dirty pages of the btree inode and flush delayed items, and
+		 * then get a new transaction handle, which may now point to a
+		 * new transaction in case someone else may have committed the
+		 * transaction we used to replace/drop file extent items. So
+		 * bump the inode's iversion and update mtime and ctime except
+		 * if we are called from a dedupe context. This is because a
+		 * power failure/crash may happen after the transaction is
+		 * committed and before we finish replacing/dropping all the
+		 * file extent items we need.
+		 */
+		inode_inc_iversion(&inode->vfs_inode);
+
+		if (!extent_info || extent_info->update_times) {
+			inode->vfs_inode.i_mtime = current_time(&inode->vfs_inode);
+			inode->vfs_inode.i_ctime = inode->vfs_inode.i_mtime;
+		}
+
 		ret = btrfs_update_inode(trans, root, inode);
 		if (ret)
 			break;
@@ -2819,7 +2859,8 @@ int btrfs_replace_file_extents(struct btrfs_inode *inode,
 
 		ret = btrfs_block_rsv_migrate(&fs_info->trans_block_rsv,
 					      rsv, min_size, false);
-		BUG_ON(ret);	/* shouldn't happen */
+		if (WARN_ON(ret))
+			break;
 		trans->block_rsv = rsv;
 
 		cur_offset = drop_args.drop_end;