cifs: remove bogus remapping of error in cifs_filldir()
authorSuresh Jayaraman <sjayaraman@suse.de>
Mon, 6 Dec 2010 08:15:50 +0000 (13:45 +0530)
committerSteve French <sfrench@us.ibm.com>
Wed, 8 Dec 2010 18:47:54 +0000 (18:47 +0000)
As the FIXME points out correctly, now filldir() itself returns -EOVERFLOW if
it not possible to represent the inode number supplied by the filesystem in
the field provided by userspace.

Signed-off-by: Suresh Jayaraman <sjayaraman@suse.de>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
fs/cifs/readdir.c

index 32d300e8f20eb4a6676b8da50187a4b04146fcb8..a73eb9f4bdaf3608c2bf6f5dc124b5fabdd4445b 100644 (file)
@@ -759,18 +759,6 @@ static int cifs_filldir(char *pfindEntry, struct file *file, filldir_t filldir,
        rc = filldir(direntry, qstring.name, qstring.len, file->f_pos,
                     ino, fattr.cf_dtype);
 
-       /*
-        * we can not return filldir errors to the caller since they are
-        * "normal" when the stat blocksize is too small - we return remapped
-        * error instead
-        *
-        * FIXME: This looks bogus. filldir returns -EOVERFLOW in the above
-        * case already. Why should we be clobbering other errors from it?
-        */
-       if (rc) {
-               cFYI(1, "filldir rc = %d", rc);
-               rc = -EOVERFLOW;
-       }
        dput(tmp_dentry);
        return rc;
 }