On Thu, Mar 12, 2015 at 6:10 AM, Mel Gorman [off-list ref] wrote:

I believe you're correct and it matches what was observed. I'm still
travelling and wireless is dirt but managed to queue a test using pmd_dirty

Ok, thanks.

I'm not entirely happy with that change, and I suspect the whole
heuristic should be looked at much more (maybe it should also look at
whether it's executable, for example), but it's a step in the right
direction.

So I committed it and added a comment, and wrote a commit log about
it. I suspect any further work is post-4.0-release, unless somebody
comes up with something small and simple and obviously better.

                                 Linus