Messages in this thread | | | Date | Fri, 13 Nov 2009 13:37:41 +0000 | From | Mel Gorman <> | Subject | Re: [PATCH 0/5] Reduce GFP_ATOMIC allocation failures, candidate fix V3 |
| |
On Fri, Nov 13, 2009 at 01:47:35PM +0100, Tobias Oetiker wrote: > Hi Mel, > > Yesterday Mel Gorman wrote: > > > Sorry for the long delay in posting another version. Testing is extremely > > time-consuming and I wasn't getting to work on this as much as I'd have liked. > > > > Changelog since V2 > > o Dropped the kswapd-quickly-notice-high-order patch. In more detailed > > testing, it made latencies even worse as kswapd slept more on high-order > > congestion causing order-0 direct reclaims. > > o Added changes to how congestion_wait() works > > o Added a number of new patches altering the behaviour of reclaim > > so is there anything promissing for the order 5 allocation problems > in this set? >
Yes. While the change in timing of direct reclaimers might be less important when dm-crypt is not involved, kswapd is more pro-active about maintaining the watermarks.
-- Mel Gorman Part-time Phd Student Linux Technology Center University of Limerick IBM Dublin Software Lab
| |