From: Jeff Garzik <jgarzik@pobox.com>
To: Nuno Silva <nuno.silva@vgertech.com>
Cc: Bartlomiej Zolnierkiewicz <B.Zolnierkiewicz@elka.pw.edu.pl>,
William Lee Irwin III <wli@holomorphy.com>,
Jens Axboe <axboe@suse.de>, Nick Piggin <nickpiggin@yahoo.com.au>,
linux-ide@vger.kernel.org,
Linux Kernel <linux-kernel@vger.kernel.org>,
Andrew Morton <akpm@osdl.org>
Subject: Re: [PATCH] speed up SATA
Date: Sun, 28 Mar 2004 15:02:28 -0500 [thread overview]
Message-ID: <40672F54.1050703@pobox.com> (raw)
In-Reply-To: <40672D07.2060201@vgertech.com>
Nuno Silva wrote:
> With this kind of control we could have /etc/init.d/io-optimize that
> paused the startup for 10 seconds and tests every device|controller in
> fstab and optimizes according to the .conf file for latency or speed...
> Or a daemon that retrieves statistics and adjusts the policies every
> minute?
>
> Also, everybody says "do it in userland". This is doing (some of) it in
> userland :)
An iotune daemon that sits in userland -- probably mlockall'd into
memory -- would be fun. Like the existing irqbalance daemon that Red
Hat (disc: my employer) ships, "iotuned" could run in the background,
adjusting policies every so often. It would be a good place to put some
of the more high-level "laptop mode" or "high throughput mode" tunables,
so that the user wouldn't have to worry about the minute details of such
things.
Just watch the IO statistics and tweak the queue settings... Gotta make
sure the queue settings don't get tweaked beyond what the hardware can
do, though.
A good iotune daemon would probably have to pay attention to various VM
statistics as well, since the VM is often the one that decides when to
start writing out stuff... such an app could (quite justifiably)
mushroom into a io-and-vmtune daemon.
This is a side issue from the topic of finding a decent in-kernel
default, of course...
Jeff
next prev parent reply other threads:[~2004-03-28 20:02 UTC|newest]
Thread overview: 115+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-03-27 22:37 [PATCH] speed up SATA Jeff Garzik
2004-03-27 23:04 ` Stefan Smietanowski
2004-03-27 23:11 ` Jeff Garzik
2004-03-28 7:23 ` Stefan Smietanowski
2004-03-28 15:37 ` Bartlomiej Zolnierkiewicz
2004-03-27 23:32 ` Bartlomiej Zolnierkiewicz
2004-03-27 23:36 ` Jeff Garzik
2004-03-27 23:40 ` Jeff Garzik
2004-03-28 0:13 ` Bartlomiej Zolnierkiewicz
2004-03-28 0:08 ` Jeff Garzik
2004-03-29 11:42 ` Pavel Machek
2004-03-27 23:37 ` Nick Piggin
2004-03-27 23:44 ` Jeff Garzik
2004-03-27 23:47 ` Nick Piggin
2004-03-27 23:59 ` Jeff Garzik
2004-03-28 14:10 ` Jens Axboe
2004-03-28 17:31 ` Jeff Garzik
2004-03-28 17:35 ` Jens Axboe
2004-03-28 17:48 ` Jeff Garzik
2004-03-28 17:54 ` Jens Axboe
2004-03-28 18:08 ` Jamie Lokier
2004-03-28 18:15 ` Jens Axboe
2004-03-28 18:55 ` Jeff Garzik
2004-03-29 8:09 ` Jens Axboe
2004-03-29 12:41 ` Jamie Lokier
2004-03-29 12:44 ` Jens Axboe
2004-03-29 12:50 ` Jamie Lokier
2004-03-29 13:05 ` Arjan van de Ven
2004-03-29 13:08 ` Jens Axboe
2004-03-30 8:13 ` Kurt Garloff
2004-03-30 11:40 ` Jens Axboe
2004-03-29 17:19 ` Craig I. Hagan
2004-03-29 18:19 ` Jeff Garzik
2004-03-28 19:06 ` Jeff Garzik
2004-03-28 18:12 ` William Lee Irwin III
2004-03-28 18:17 ` Jens Axboe
2004-03-28 18:30 ` Bartlomiej Zolnierkiewicz
2004-03-28 18:30 ` Jens Axboe
2004-03-28 18:45 ` Bartlomiej Zolnierkiewicz
2004-03-28 18:59 ` Jeff Garzik
2004-03-28 20:32 ` Andrew Morton
2004-03-28 20:45 ` Jeff Garzik
2004-03-29 0:55 ` Andrea Arcangeli
2004-03-29 4:02 ` Jeff Garzik
2004-03-29 13:04 ` Andrea Arcangeli
2004-03-29 19:45 ` Jeff Garzik
2004-03-30 11:09 ` Jens Axboe
2004-03-30 15:54 ` Timothy Miller
2004-03-30 16:20 ` Jeff Garzik
2004-03-30 18:05 ` Timothy Miller
2004-03-30 17:50 ` Jeff Garzik
2004-03-30 18:19 ` Timothy Miller
2004-03-29 4:29 ` Wim Coekaerts
2004-03-29 7:32 ` Denis Vlasenko
2004-03-29 8:13 ` Jens Axboe
2004-03-29 13:05 ` Andrea Arcangeli
2004-03-29 4:31 ` William Lee Irwin III
2004-03-29 4:57 ` Jeff Garzik
2004-03-28 19:52 ` Nuno Silva
2004-03-28 20:02 ` Jeff Garzik [this message]
2004-03-28 0:06 ` Jeff Garzik
2004-03-28 0:15 ` Nick Piggin
2004-03-28 0:49 ` Jeff Garzik
2004-03-28 1:02 ` Andrew Morton
2004-03-28 1:09 ` Jeff Garzik
2004-03-28 13:59 ` Jens Axboe
2004-03-28 17:29 ` Jeff Garzik
2004-03-28 17:31 ` Jens Axboe
2004-03-28 13:51 ` Jamie Lokier
2004-03-28 17:24 ` Jeff Garzik
2004-03-28 17:36 ` Jamie Lokier
2004-03-28 17:54 ` Jeff Garzik
2004-03-28 20:50 ` Eric D. Mudama
2004-04-02 10:11 ` Jeremy Higdon
2004-04-02 16:11 ` Jamie Lokier
2004-04-03 10:48 ` Jeremy Higdon
2004-04-03 13:49 ` Jamie Lokier
2004-03-28 17:40 ` Jens Axboe
2004-03-28 17:49 ` Jeff Garzik
2004-03-28 17:55 ` Jens Axboe
2004-03-28 18:04 ` Jeff Garzik
2004-03-28 18:09 ` Jens Axboe
2004-03-28 20:12 ` Jeff Garzik
2004-03-28 20:54 ` Eric D. Mudama
2004-03-28 7:32 ` Stefan Smietanowski
2004-03-28 20:25 ` Jeff Garzik
2004-03-28 21:16 ` Stefan Smietanowski
2004-03-28 21:26 ` Jeff Garzik
2004-03-28 14:08 ` Jens Axboe
2004-03-28 17:38 ` Jeff Garzik
2004-03-28 17:45 ` Jens Axboe
2004-03-28 20:21 ` Jeff Garzik
2004-03-28 0:07 ` Andrew Morton
2004-03-28 0:21 ` Nick Piggin
2004-03-28 4:40 ` Eric D. Mudama
2004-03-28 6:56 ` Nick Piggin
2004-03-28 20:33 ` Eric D. Mudama
2004-03-28 20:59 ` Eric D. Mudama
2004-03-29 1:30 ` Nick Piggin
2004-03-29 5:24 ` Eric D. Mudama
2004-03-29 13:03 ` Jamie Lokier
2004-03-29 11:36 ` Pavel Machek
2004-03-29 18:46 ` David Lang
2004-03-29 20:13 ` Jeff Garzik
2004-03-30 5:55 ` Eric D. Mudama
2004-03-30 11:54 ` Marc Bevand
2004-03-30 13:07 ` Jens Axboe
2004-03-30 13:48 ` Marc Bevand
2004-03-30 13:49 ` Jens Axboe
2004-03-30 15:31 ` Jeff Garzik
2004-03-30 17:42 ` Jeff Garzik
2004-03-31 9:12 ` Marc Bevand
-- strict thread matches above, loose matches on Subject: below --
2004-03-31 5:47 Marcus Hartig
2004-03-31 6:56 ` Jeff Garzik
2004-03-31 16:07 ` Marcus Hartig
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=40672F54.1050703@pobox.com \
--to=jgarzik@pobox.com \
--cc=B.Zolnierkiewicz@elka.pw.edu.pl \
--cc=akpm@osdl.org \
--cc=axboe@suse.de \
--cc=linux-ide@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=nickpiggin@yahoo.com.au \
--cc=nuno.silva@vgertech.com \
--cc=wli@holomorphy.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