lkml.org 
[lkml]   [2014]   [Feb]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH] mm: per-thread vma caching
From
Date
On Fri, 2014-02-21 at 10:13 -0800, Linus Torvalds wrote:
> On Thu, Feb 20, 2014 at 9:28 PM, Davidlohr Bueso <davidlohr@hp.com> wrote:
> > From: Davidlohr Bueso <davidlohr@hp.com>
> >
> > This patch is a continuation of efforts trying to optimize find_vma(),
> > avoiding potentially expensive rbtree walks to locate a vma upon faults.
>
> Ok, so I like this one much better than the previous version.
>
> However, I do wonder if the per-mm vmacache is actually worth it.
> Couldn't the per-thread one replace it entirely?

I think you are right. I just reran some of the tests and things are
pretty much the same, so we could get rid of it. I originally left it
there because I recall seeing a slightly better hit rate for some java
workloads (map/reduce, specifically), but it wasn't a big deal - some
slots endup being redundant with a per mm cache. It does however
guarantee that we access hot vmas immediately, instead of potentially
slightly more reads when we go into per-thread checking. I'm happy with
the results either way.

> Also, the hash you use for the vmacache index is *particularly* odd.
>
> int idx = (addr >> 10) & 3;
>
> you're using the top two bits of the address *within* the page.
> There's a lot of places that round addresses down to pages, and in
> general it just looks really odd to use an offset within a page as an
> index, since in some patterns (linear accesses, whatever), the page
> faults will always be to the beginning of the page, so index 0 ends up
> being special.

Ah, this comes from tediously looking at access patterns. I actually
printed pages of them. I agree that it is weird, and I'm by no means
against changing it. However, the results are just too good, specially
for ebizzy, so I decided to keep it, at least for now. I am open to
alternatives.

Thanks,
Davidlohr



\
 
 \ /
  Last update: 2014-02-21 22:21    [W:0.087 / U:0.084 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site