On Tue, May 03, 2016 at 09:37:57AM +0200, Michal Hocko wrote: > On Mon 02-05-16 19:02:50, Kirill A. Shutemov wrote: > > On Mon, May 02, 2016 at 08:49:03AM -0700, Dave Hansen wrote: > > > On 05/02/2016 08:01 AM, Kirill A. Shutemov wrote: > > > > On Mon, May 02, 2016 at 04:39:35PM +0200, Vlastimil Babka wrote: > > > >> On 04/27/2016 07:11 PM, Dave Hansen wrote: > > > >>> 6. Perhaps don't use the LRU pagevecs for large pages. It limits the > > > >>> severity of the problem. > > > >> > > > >> I think that makes sense. Being large already amortizes the cost per > > > >> base > > > >> page much more than pagevecs do (512 vs ~22 pages?). > > > > > > > > We try to do this already, don't we? Any spefic case where we have THPs > > > > on > > > > pagevecs? > > > > > > Lukas was hitting this on a RHEL 7 era kernel. In his kernel at least, > > > I'm pretty sure THP's were ending up on pagevecs. Are you saying you > > > don't think we're doing that any more? > > > > As Vlastimil pointed, we do. It need to be fixed, I think. > > It seems that offloading the draining to the vmstat context doesn't look > terribly bad. Don't we rather want to go that way?
Maybe. My knowledge about lru cache is limited. -- Kirill A. Shutemov
