Note that this email comes at the end of a really long thread arguing this point (reversion in stable trees) on LKML. The discussion also spilled onto G+. I'm not at all surprised that Linus (among other people) might be losing some patience.