ocfs2: set non-default s_time_gran during mount
authorMark Fasheh <mark.fasheh@oracle.com>
Thu, 9 Aug 2007 23:52:30 +0000 (16:52 -0700)
committerMark Fasheh <mark.fasheh@oracle.com>
Fri, 10 Aug 2007 00:27:58 +0000 (17:27 -0700)
We need to manually set this to '1' during mount, otherwise inode_setattr()
will chop off the nanosecond portion of our timestamps.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
fs/ocfs2/super.c

index c3c89e26567ca9ec4c980cc3305223701182b42f..f2fc9a795debe7b3933ca6277100738beb0ef0e7 100644 (file)
@@ -1291,6 +1291,7 @@ static int ocfs2_initialize_super(struct super_block *sb,
        sb->s_fs_info = osb;
        sb->s_op = &ocfs2_sops;
        sb->s_export_op = &ocfs2_export_ops;
+       sb->s_time_gran = 1;
        sb->s_flags |= MS_NOATIME;
        /* this is needed to support O_LARGEFILE */
        cbits = le32_to_cpu(di->id2.i_super.s_clustersize_bits);