From: Hubertus Franke <frankeh@watson.ibm.com>
To: Daniel Phillips <phillips@arcor.de>,
davidm@hpl.hp.com, David Mosberger <davidm@napali.hpl.hp.com>,
"David S. Miller" <davem@redhat.com>
Cc: davidm@hpl.hp.com, davidm@napali.hpl.hp.com,
torvalds@transmeta.com, gh@us.ibm.com, Martin.Bligh@us.ibm.com,
wli@holomorpy.com, linux-kernel@vger.kernel.org
Subject: Re: large page patch (fwd) (fwd)
Date: Fri, 9 Aug 2002 15:17:55 -0400 [thread overview]
Message-ID: <200208091517.55019.frankeh@watson.ibm.com> (raw)
In-Reply-To: <E17dEje-0001Ro-00@starship>
On Friday 09 August 2002 02:43 pm, Daniel Phillips wrote:
> On Friday 09 August 2002 20:32, Hubertus Franke wrote:
> > On Friday 09 August 2002 11:20 am, Daniel Phillips wrote:
> > > On Sunday 04 August 2002 19:19, Hubertus Franke wrote:
> > > > "General Purpose Operating System Support for Multiple Page Sizes"
> > > > htpp://www.usenix.org/publications/library/proceedings/usenix98/full_
> > > >pape rs/ganapathy/ganapathy.pdf
> > >
> > > This reference describes roughly what I had in mind for active
> > > defragmentation, which depends on reverse mapping. The main additional
> > > wrinkle I'd contemplated is introducing a new ZONE_LARGE, and
> > > GPF_LARGE, which means the caller promises not to pin the allocation
> > > unit for long periods and does not mind if the underlying physical page
> > > changes spontaneously. Defragmenting in this zone is straightforward.
> >
> > I think the objection to that is that in many cases the cost of
> > defragmentation is to heavy to be recollectable through TLB miss handling
> > alone.
>
> You pay the cost only on transition from a load that doesn't use many large
> pages to one that does, it is not an ongoing cost.
>
Correct. Maybe I misunderstood, when are you doing the coalloction of
adjacent pages (page-clusters, super pages).
Our intend was to do it at page fault time and breakup only during
memory pressure.
> > [...]
> >
> > Defragmenting to me seems a matter of last resort, Copying pages is
> > expensive.
>
> It is the only way to ever have a seamless implementation. Really, I don't
> understand this fear of active defragmentation. Oh well, like davem said,
> code talks.
--
-- Hubertus Franke (frankeh@watson.ibm.com)
next prev parent reply other threads:[~2002-08-09 19:17 UTC|newest]
Thread overview: 110+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <E17ahdi-0001RC-00@w-gerrit2>
2002-08-02 19:34 ` large page patch (fwd) (fwd) Linus Torvalds
2002-08-03 3:19 ` David Mosberger
2002-08-03 3:32 ` Linus Torvalds
2002-08-03 4:17 ` David Mosberger
2002-08-03 4:26 ` Linus Torvalds
2002-08-03 4:39 ` David Mosberger
2002-08-03 5:20 ` David S. Miller
2002-08-03 17:35 ` Linus Torvalds
2002-08-03 19:30 ` David Mosberger
2002-08-03 19:43 ` Linus Torvalds
2002-08-03 21:18 ` David Mosberger
2002-08-03 21:54 ` Hubertus Franke
2002-08-04 0:35 ` David S. Miller
2002-08-04 2:25 ` David Mosberger
2002-08-04 17:19 ` Hubertus Franke
2002-08-09 15:20 ` Daniel Phillips
2002-08-09 15:56 ` Linus Torvalds
2002-08-09 16:15 ` Daniel Phillips
2002-08-09 16:31 ` Rik van Riel
2002-08-09 18:08 ` Daniel Phillips
2002-08-09 16:51 ` Linus Torvalds
2002-08-09 17:11 ` Daniel Phillips
2002-08-09 16:27 ` Rik van Riel
2002-08-09 16:52 ` Linus Torvalds
2002-08-09 17:40 ` yodaiken
2002-08-09 19:15 ` Rik van Riel
2002-08-09 21:20 ` Linus Torvalds
2002-08-09 21:19 ` Marcin Dalecki
2002-08-09 17:46 ` Bill Rugolsky Jr.
2002-08-12 9:23 ` Helge Hafting
2002-08-13 3:15 ` Bill Davidsen
2002-08-13 3:31 ` Rik van Riel
2002-08-13 7:28 ` Helge Hafting
2002-08-09 21:38 ` Andrew Morton
2002-08-10 18:20 ` Eric W. Biederman
2002-08-10 18:59 ` Daniel Phillips
2002-08-10 19:55 ` Rik van Riel
2002-08-10 19:54 ` Eric W. Biederman
2002-08-09 18:32 ` Hubertus Franke
2002-08-09 18:43 ` Daniel Phillips
2002-08-09 19:17 ` Hubertus Franke [this message]
2002-08-11 20:30 ` Alan Cox
2002-08-11 22:33 ` Daniel Phillips
2002-08-11 22:55 ` Linus Torvalds
2002-08-11 22:56 ` Linus Torvalds
2002-08-11 23:36 ` William Lee Irwin III
2002-08-12 0:46 ` Alan Cox
2002-08-11 23:42 ` Rik van Riel
2002-08-11 23:50 ` Larry McVoy
2002-08-12 8:22 ` Daniel Phillips
2002-08-13 8:40 ` Rob Landley
2002-08-13 15:06 ` Alan Cox
2002-08-13 11:36 ` Rob Landley
2002-08-13 16:51 ` Linus Torvalds
2002-08-13 12:53 ` Rob Landley
2002-08-13 17:14 ` Ruth Ivimey-Cook
2002-08-13 17:29 ` Rik van Riel
2002-08-13 13:18 ` Rob Landley
2002-08-13 18:32 ` Linus Torvalds
2002-08-13 13:50 ` Rob Landley
2002-08-13 17:45 ` Alexander Viro
2002-08-13 17:55 ` Linus Torvalds
2002-08-13 17:59 ` Rik van Riel
2002-08-13 13:35 ` Rob Landley
2002-08-13 19:12 ` Daniel Phillips
2002-08-22 12:03 ` bill davidsen
[not found] ` <Pine.LNX.4.44.0208130942130.7411-100000@home.transmeta.com >
2002-08-13 18:46 ` large page patch (fwd) Mike Galbraith
2002-08-11 23:44 ` large page patch (fwd) (fwd) Daniel Phillips
2002-08-13 8:51 ` Rob Landley
2002-08-13 16:47 ` Daniel Phillips
2002-08-13 13:09 ` Rob Landley
2002-08-11 23:15 ` Larry McVoy
2002-08-12 1:26 ` Linus Torvalds
2002-08-12 5:05 ` Larry McVoy
2002-08-12 10:31 ` Alan Cox
2002-08-04 0:28 ` David S. Miller
2002-08-04 17:31 ` Hubertus Franke
2002-08-04 18:38 ` Linus Torvalds
2002-08-04 19:23 ` Andrew Morton
2002-08-04 19:28 ` Linus Torvalds
2002-08-05 5:42 ` David S. Miller
2002-08-04 19:30 ` Hubertus Franke
2002-08-04 20:23 ` William Lee Irwin III
2002-08-05 16:59 ` David Mosberger
2002-08-05 17:21 ` Hubertus Franke
2002-08-05 21:10 ` Jamie Lokier
2002-08-04 19:41 ` Rik van Riel
2002-08-05 5:40 ` David S. Miller
2002-08-03 18:41 ` Hubertus Franke
2002-08-03 19:39 ` Linus Torvalds
2002-08-04 0:32 ` David S. Miller
2002-08-03 19:41 ` David Mosberger
2002-08-03 20:53 ` Hubertus Franke
2002-08-03 21:26 ` David Mosberger
2002-08-03 21:50 ` Hubertus Franke
2002-08-04 0:34 ` David S. Miller
2002-08-04 0:31 ` David S. Miller
2002-08-04 17:25 ` Hubertus Franke
[not found] <200208041331.24895.frankeh@watson.ibm.com.suse.lists.linux.kernel>
[not found] ` <Pine.LNX.4.44.0208041131380.10314-100000@home.transmeta.com.suse.lists.linux.kernel>
[not found] ` <3D4D7F24.10AC4BDB@zip.com.au.suse.lists.linux.kernel>
2002-08-04 20:20 ` Andi Kleen
2002-08-04 23:51 ` Eric W. Biederman
2002-08-05 23:30 Seth, Rohit
2002-08-06 5:01 ` David Mosberger
2002-08-06 4:58 ` David S. Miller
2002-08-06 5:19 ` David Mosberger
2002-08-06 5:08 ` David S. Miller
2002-08-06 5:32 ` David Mosberger
2002-08-06 19:11 ` Hubertus Franke
2002-08-06 20:38 Luck, Tony
2002-08-06 21:03 ` Hubertus Franke
2002-08-09 17:51 Seth, Rohit
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200208091517.55019.frankeh@watson.ibm.com \
--to=frankeh@watson.ibm.com \
--cc=Martin.Bligh@us.ibm.com \
--cc=davem@redhat.com \
--cc=davidm@hpl.hp.com \
--cc=davidm@napali.hpl.hp.com \
--cc=gh@us.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=phillips@arcor.de \
--cc=torvalds@transmeta.com \
--cc=wli@holomorpy.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).