Hi... On Sun, Feb 8, 2009 at 6:00 PM, Himanshu Chauhan <[email protected]> wrote: > On Sun, 2009-02-08 at 17:47 +0700, Mulyadi Santosa wrote: >> Hi.. >> >> On Sun, Feb 8, 2009 at 1:46 AM, Himanshu Chauhan <[email protected]> >> wrote: >> > Am I correct when I say that try_to_free_pages is called in direct >> > reclaim, and balance_pgdat is invoked, directly or indirectly, by >> > kswapd? >> >> let me make your question clear first. Do you ask this in the kswapd's >> poin of view? If yes, then I think your conclusion is right. >> >> regards, >> >> Mulyadi. > > Is direct reclaim the path taken by kwapd? I don't think so.
Sorry, I mistakenly understood your question. Direct reclaim does call try_to_free_pages, but kswapd doesn't call it. I was 100% sure it does direct reclaim, but my mistake was by saying kswapd does direct reclaim. This document nicely and quickly sumarize it for us: http://www.cs.cmu.edu/~412/history/2005S/CART/The%20Story%20of%20a%20Page.pdf NB: I also almost think there's another entry point that is something like cache_refill and/or cache_reclaim, but suddenly I recall that's not the main entry point. regards, Mulyadi. -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to [email protected] Please read the FAQ at http://kernelnewbies.org/FAQ
