f2fs: should cover i_xattr_nid with its xattr node page lock

Previously, f2fs_setxattr assigns i_xattr_nid in the inode page inconsistently.

The scenario is:

= Thread 1 =         = Thread 2 =     = fi->i_xattr_nid =  = on-disk nid =

f2fs_setxattr                                   0                 0
  new_node_page                                 X                 0
                   sync_inode_page              X                 X
                   checkpoint                   X                 X -.
    grab_cache_page                             X                 X  |
--> allocate a new xattr node block or -ENOSPC      <----------------'

At this moment, the checkpoint stores inconsistent data where the inode has
i_xattr_nid but actual xattr node block is not allocated yet.

So, we should assign the real i_xattr_nid only after its xattr node block is
allocated.

Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>

1 files changed, 5 insertions, 5 deletions

diff --git a/fs/f2fs/xattr.c b/fs/f2fs/xattr.c
index fb16f71..3bc307c 100644
--- a/fs/f2fs/xattr.c
+++ b/fs/f2fs/xattr.c
@@ -378,23 +378,23 @@ int f2fs_setxattr(struct inode *inode, int name_index, const char *name,
 	if (!fi->i_xattr_nid) {
 		/* Allocate new attribute block */
 		struct dnode_of_data dn;
+		nid_t new_nid;
 
-		if (!alloc_nid(sbi, &fi->i_xattr_nid)) {
+		if (!alloc_nid(sbi, &new_nid)) {
 			error = -ENOSPC;
 			goto exit;
 		}
-		set_new_dnode(&dn, inode, NULL, NULL, fi->i_xattr_nid);
+		set_new_dnode(&dn, inode, NULL, NULL, new_nid);
 		mark_inode_dirty(inode);
 
 		page = new_node_page(&dn, XATTR_NODE_OFFSET, ipage);
 		if (IS_ERR(page)) {
-			alloc_nid_failed(sbi, fi->i_xattr_nid);
-			fi->i_xattr_nid = 0;
+			alloc_nid_failed(sbi, new_nid);
 			error = PTR_ERR(page);
 			goto exit;
 		}
 
-		alloc_nid_done(sbi, fi->i_xattr_nid);
+		alloc_nid_done(sbi, new_nid);
 		base_addr = page_address(page);
 		header = XATTR_HDR(base_addr);
 		header->h_magic = cpu_to_le32(F2FS_XATTR_MAGIC);