Re: kswapd eating too much CPU on ac16/ac18

From: Roger Larsson (
Date: Sat Jun 17 2000 - 16:34:55 EST


The reason for me to ask you to remove it is since there are two
related to this code snippet. (Reported earlier on linux-mm)

* If no zone has pressure - we will loop "forever" since no pages will
this test. (On a 16 MB machine this is the likely scenario)

* If there are no pages of a zone with pressure are on LRU - we will

And since there is no guarantee that shrink_mmap is not called in these

I have released patches (on linux-mm) that tries to handle these
* do_try_to_free_pages avoids to call shrink_mmap with no pressure.
* shrink_mmap tries to determine the bad situation (not in my latest)


Cesar Eduardo Barros wrote:
> > Please try to remove only this test to get a comparable result.
> I nuked the whole block:
> /*
> * Page is from a zone we don't care about.
> * Don't drop page cache entries in vain.
> */
> if (page->zone->free_pages > page->zone->pages_high) {
> /* the page from the wrong zone doesn't count */
> count++;
> goto unlock_continue;
> }
> Commenting it out made ac19 perform almost as good as ac4 (it looked a bit
> faster).
> I don't know how it would affect boxes with more than one zone, but my gut
> feeling is that it won't hurt and might make them even a bit faster.
> --
> Cesar Eduardo Barros
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to For more info on Linux MM,
> see:

Home page:

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to Please read the FAQ at

This archive was generated by hypermail 2b29 : Fri Jun 23 2000 - 21:00:15 EST