Messages in this thread | | | Date | Wed, 15 Jul 1998 02:36:39 +0200 (MET DST) | From | MOLNAR Ingo <> | Subject | Re: Strange interrupt behaviour |
| |
On Wed, 15 Jul 1998, MOLNAR Ingo wrote:
> i'd do something like this to satisfy higher-order goals: > > if (makes_situation_better(this_candidate_page,order)) > swap_out(this_candidate_page);
makes_situation_better() can be improved further, by possibly adding:
1) an 'unswappables bitmaps', basically shadowing the buddy bitmaps, but showing all the 'impossible to swap out now' pages. Cons: higher allocation overhead, higher cache/memory usage. Pros: makes the effect of unswappable pages exactly predictable.
2) 'goals bitmaps', shadowing the buddy bitmaps again, but showing all the swapouts currently being done to reach one given higher-order goal. Cons: higher cache/memory usage. Pros: eliminates the exponential nature of 'satisfy higher-order request' allocations, focuses swapouts, reduces the number of 'accidentally destroyed lower-order but useful' pages.
i think both of these are still much better than reverse page tables.
-- mingo
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html
| |