Messages in this thread |  | | | From | KOSAKI Motohiro <> | | Subject | Re: [PATCH 2/3] mm: shrink_inactive_lis() nr_scan accounting fix fix | | Date | Fri, 17 Jul 2009 09:16:19 +0900 (JST) |
| |
> Not a newly introduced problem, but this early break might under scan > the list, if (max_scan > swap_cluster_max). Luckily the only two > callers all call with (max_scan <= swap_cluster_max). > > What shall we do? The comprehensive solution may be to > - remove the big do-while loop > - replace sc->swap_cluster_max => max_scan > - take care in the callers to not passing small max_scan values > > Or to simply make this function more robust like this?
Sorry, I haven't catch your point. Can you please tell me your worried scenario?
|  |