Oh silly me, thanks Andrew Wang!
I was trapped by the order of patches in
HDFS-6133, I thought rev 9 (the last one in the list) is the one committed, actually it's rev 10 which showed up earlier in the list rather than the last.. Three digits is recommended to represent patch rev number, however, the key is that I should have examined in the committed source code.
Thanks again. New rev to come soon.