From: Robert Love <rml@tech9.net>
To: Chris Wedgwood <cw@f00f.org>
Cc: linux-kernel@vger.kernel.org, akpm@digeo.com
Subject: Re: [PATCH] O_STREAMING - flag for optimal streaming I/O
Date: 08 Oct 2002 15:59:50 -0400 [thread overview]
Message-ID: <1034107190.29467.1533.camel@phantasy> (raw)
In-Reply-To: <20021008195223.GA5040@tapu.f00f.org>
On Tue, 2002-10-08 at 15:52, Chris Wedgwood wrote:
> On Tue, Oct 08, 2002 at 03:17:16PM -0400, Robert Love wrote:
>
> > Yep. Linux treats most "hints" (e.g. madvise) as a requirement - it
> > fails if it cannot do it. That is against the spec most of the
> > time, but oh well...
>
> There is no spec for O_DIRECT... SGI 'invented' this in '93 or perhaps
> earlier (but the idea wasn't new) for IRIX.
I was speaking more of madvise() and in general. I know O_DIRECT does
not have a spec. In general, Linux returns failures on things that many
other operating systems just consider hints (i.e. madvise()).
> O_DIRECT is a very special thing, you shouldn't ask for this unless
> yoy know you want it and how to deal with it --- treating it as
> anything less that a requirement is bogus IMO.
Agreed. Partly why O_STREAMING is needed.
Remember not everything implements O_DIRECT, especially not some odd
device you are streaming into/out of.
I think Andrew summed it up: if O_DIRECT will work in your environment,
and you can rewrite your application, it is probably preferred.
O_STREAMING is a simple solution to solve the pagecache waste which
requires one change to the application.
I did not intend for this to be an O_DIRECT vs. O_STREAMING thread. A
lot of people agree we need something like O_STREAMING - despite never
being implemented, you can find its name referenced often in archives
via google. A much more interesting argument is whether we should not
have an explicit O_STREAMING but instead an intelligent drop-behind
heuristic... but that is a 2.5 issue and the patch is for 2.4.
20% increase in kernel compilation is amazingly nice, for free.
Robert Love
next prev parent reply other threads:[~2002-10-08 19:59 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-10-08 2:38 [PATCH] O_STREAMING - flag for optimal streaming I/O Robert Love
2002-10-08 10:42 ` J.A. Magallon
2002-10-08 18:08 ` Robert Love
2002-10-08 18:38 ` Chris Wedgwood
2002-10-08 18:49 ` Robert Love
2002-10-08 19:05 ` Chris Wedgwood
2002-10-08 19:17 ` Robert Love
2002-10-08 19:30 ` Andrew Morton
2002-10-09 14:14 ` Marco Colombo
2002-10-09 16:30 ` kernel
2002-10-08 19:52 ` Chris Wedgwood
2002-10-08 19:59 ` Robert Love [this message]
2002-10-08 20:01 ` Chris Wedgwood
2002-10-09 8:33 ` Giuliano Pochini
2002-10-09 8:43 ` Andrew Morton
2002-10-09 10:55 ` Giuliano Pochini
2002-10-09 17:05 ` Mark Mielke
2002-10-09 19:36 ` Giuliano Pochini
2002-10-09 22:24 ` Mark Mielke
2002-10-09 23:20 ` Jamie Lokier
2002-10-10 3:07 ` Mark Mielke
2002-10-10 10:55 ` Helge Hafting
2002-10-10 17:50 ` Mark Mielke
2002-10-10 3:29 ` Erik Andersen
2002-10-10 3:37 ` Robert Love
2002-10-10 13:39 ` Giuliano Pochini
2002-10-10 22:50 ` Mike Fedyk
2002-10-10 22:58 ` Erik Andersen
2002-10-11 8:26 ` Giuliano Pochini
2002-10-11 8:32 ` Helge Hafting
2002-10-10 8:33 ` Giuliano Pochini
2002-10-10 9:10 ` Erik Andersen
2002-10-10 9:38 ` Giuliano Pochini
2002-10-10 10:40 ` Miquel van Smoorenburg
2002-10-10 11:01 ` Helge Hafting
2002-10-10 12:29 ` Xavier Bestel
2002-10-10 13:17 ` Giuliano Pochini
2002-10-10 22:44 ` Mike Fedyk
2002-10-11 8:13 ` Giuliano Pochini
2002-10-10 11:38 ` O_STREAMING has insufficient info - how about fadvise() ? Alan Cox
2002-10-10 11:47 ` William Lee Irwin III
2002-10-10 15:34 ` Andrew Morton
2002-10-10 16:08 ` Alan Cox
2002-10-10 16:49 ` Oliver Xymoron
2002-10-10 15:37 ` [PATCH] O_STREAMING - flag for optimal streaming I/O Gerhard Mack
2002-10-10 22:47 ` Mike Fedyk
2002-10-11 2:14 ` Gerhard Mack
2002-10-11 8:10 ` Chris Wedgwood
2002-10-10 9:14 ` David Lang
2002-10-10 14:51 ` Denis Vlasenko
2002-10-08 19:53 ` Matthias Schniedermeyer
2002-10-08 19:59 ` Chris Wedgwood
2002-10-08 20:03 ` Andrew Morton
2002-10-08 20:34 ` Matthias Schniedermeyer
2002-10-08 20:42 ` Andrew Morton
2002-10-08 20:37 ` Larry McVoy
2002-10-09 11:53 ` Roy Sigurd Karlsbakk
2002-10-09 14:10 ` Marco Colombo
2002-10-09 14:14 ` Robert Love
2002-10-09 14:33 ` Richard B. Johnson
2002-10-09 15:27 ` Andreas Dilger
2002-10-09 23:17 ` Jamie Lokier
2002-10-09 23:46 ` Rik van Riel
2002-10-10 0:16 ` Jamie Lokier
2002-10-10 2:39 ` Erik Andersen
2002-10-10 10:33 ` Marco Colombo
2002-10-10 20:00 ` Erik Andersen
2002-10-11 4:16 Hank Leininger
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=1034107190.29467.1533.camel@phantasy \
--to=rml@tech9.net \
--cc=akpm@digeo.com \
--cc=cw@f00f.org \
--cc=linux-kernel@vger.kernel.org \
/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).