ocfs2: rollback transaction in ocfs2_group_add()
If ocfs2_journal_access_di() fails, group->bg_next_group should rollback.
Otherwise, there would be a inconsistency between group_bh and main_bm_bh.
Signed-off-by: Younger Liu <younger.liu@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.com>
Acked-by: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
diff --git a/fs/ocfs2/resize.c b/fs/ocfs2/resize.c
index ec55add..376750f 100644
--- a/fs/ocfs2/resize.c
+++ b/fs/ocfs2/resize.c
@@ -469,6 +469,7 @@
struct ocfs2_chain_list *cl;
struct ocfs2_chain_rec *cr;
u16 cl_bpc;
+ u64 bg_ptr;
if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb))
return -EROFS;
@@ -538,12 +539,14 @@
}
group = (struct ocfs2_group_desc *)group_bh->b_data;
+ bg_ptr = le64_to_cpu(group->bg_next_group);
group->bg_next_group = cr->c_blkno;
ocfs2_journal_dirty(handle, group_bh);
ret = ocfs2_journal_access_di(handle, INODE_CACHE(main_bm_inode),
main_bm_bh, OCFS2_JOURNAL_ACCESS_WRITE);
if (ret < 0) {
+ group->bg_next_group = cpu_to_le64(bg_ptr);
mlog_errno(ret);
goto out_commit;
}