On Mon, 28 Jul 2014, Vlastimil Babka wrote: > Looks like kind of a shotgun approach to me. A single __GFP_NO_KSWAPD bullet > is no longer enough, so we use all the flags and hope for the best. It seems > THP has so many flags it should be unique for now, but I wonder if there is a > better way to say how much an allocation is willing to wait. > We would have to introduce a new __GFP_FAULT bit to distinguish between allocations at pagefault that should not use synchronous memory compaction solely for this case, it's probably not worth it. -- 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=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>