I am quite shockingly ignorant of the MM code. While looking at this function to figure out how/whether to use it, I noticed the bug, and sent a patch. I assumed the gibberish in the changelog meant something important to people who actually understand this part of the kernel :-) -----Original Message----- From: Andrew Morton [mailto:akpm@xxxxxxxxxxxxxxxxxxxx] Sent: Wednesday, August 13, 2014 1:13 PM To: Wilcox, Matthew R Cc: linux-mm@xxxxxxxxx; Mel Gorman; Rik van Riel; stable@xxxxxxxxxxxxxxx Subject: Re: [PATCH] mm: Actually clear pmd_numa before invalidating On Wed, 13 Aug 2014 20:04:02 +0000 "Wilcox, Matthew R" <matthew.r.wilcox@xxxxxxxxx> wrote: > The commit log for 67f87463d3 explains what the runtime effects should have been. No it doesn't. In fact the sentence "The existing caller of pmdp_invalidate should handle it but it's an inconsistent state for a PMD." makes me suspect there are no end-user visible effects. I don't know why we chose to backport that one into -stable and I don't know why we should backport this one either. Greg (and others) will look at this changelog and wonder "why". It should tell them. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href