Re: 2.1.76, nfs client, and memory fragmentation
kwrohrer@enteract.com
Wed, 31 Dec 1997 22:59:16 -75000 (CST)
And lo, Linus Torvalds saith unto me:
> In article <199712291047.EAA28278@jadrek.kwr>, <kwrohrer@enteract.com> wrote:
> >I'll try it, but I don't think this will solve the fragmentation. It will
> >adjust the reservation to try to hold on to free chunks of larger orders,
> >but once a high-priority allocation snags the last one we're back where we
> >started, with no way of bringing high-order free chunks back into the
> >system.
>
> Correct. Also, the patch probably doesn't mix all that nicely with the
> slab allocator which has this nasty habit of wanting larger areas than
> it really needs (because it can do a better job of cache coloring with a
> larger area to play with, and it also has a smaller wasted memory
> impact). That may undo some of the advantage of the patch - back when I
> tried it originally the slab code didn't exist.
>
> Anyway, I'd like to know what people think about the patch (even just
> subjective "feels less snappy" or similar)
I can definitely feel pauses (it feels like more pauses than there ought
to be) but they are of finite length so far.
Meanwhile, is there a semblance of reverse page table I haven't found yet?
I can find the pages I want to shuffle around by physical address, but I
need the task and page structs in order to do it, and brute-force searching
just isn't worth it...
Keith