On Mon, Aug 22, 2016 at 08:55:37PM +1000, Dave Chinner wrote:

Ok, I didn't notice that the fix from 3.14 was further down the
queue. I put a procmail filter in to catch this patch on lkml
so i didn't see it in the context of the entire series (way too much
traffic on lkml to keep up with it). So I probably pulled the
trigger a little early.

No offence taken, don't worry. I prefer to see maintainers fight for
reliability than not care :-)

I agreed that it would be best to combine the two patches so there
isn't a bisection point that could result in corruptions...

yes, I've re-merged them already.

Thanks,
Willy