linux-kernel - Re: General question (scheduler) with SSDs?

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20090220094356.GB26124@silver.sucs.org>
Date:	Fri, 20 Feb 2009 09:43:56 +0000
From:	Sitsofe Wheeler <sitsofe@...oo.com>
To:	Justin Piszcz <jpiszcz@...idpixels.com>
Cc:	linux-kernel@...r.kernel.org, linux-ide@...r.kernel.org,
	linux-raid@...r.kernel.org
Subject: Re: General question (scheduler) with SSDs?

Whoops, I didn't mean to send that previous half formed mail :) Sorry.

On Thu, Feb 19, 2009 at 08:38:38AM -0500, Justin Piszcz wrote:
> 
> I was curious if anyone had done any benchmarks and/or has conclusive 
> information, what is the best Linux scheduler to use with SSDs?
> 
> Noop?
> CFQ?
> AS?
> Deadline?

As mentioned in another mail there was a discussion on
http://lkml.org/lkml/2009/1/31/28 and long thread talking about the
introduction of the rotational flag here
http://lkml.org/lkml/2009/1/5/340 . Cheapo SSDs or even USB keys are not
auto detected as non-rotational devices by the kernel and after a bit of
poking about I've come up with the following udev rules for my
particular cases:

SUBSYSTEM=="block", TEST=="/sys$devpath/queue/rotational", ATTRS{model}=="ASUS-PHISON *", RUN+="/bin/sh -c 'echo 0 > /sys$devpath/queue/rotational'"
SUBSYSTEM=="block", TEST=="/sys$devpath/queue/rotational", ATTRS{idVendor}=="0951", ATTRS{idProduct}=="1606", RUN+="/bin/sh -c 'echo 0 > /sys$devpath/queue/rotational'"
SUBSYSTEM=="block", TEST=="/sys$devpath/queue/rotational", ATTRS{manufacturer}=="SanDisk", ATTRS{product}=="Cruzer Micro", RUN+="/bin/sh -c 'echo 0 > /sys$devpath/queue/rotational'"

Ever since the rotational option appeared I've been trying cfq but prior
to that I was using noop or deadline. However it doesn't look like
anyone has sat down and run the numbers to see what affect the
ioscheulder/rotational flag is having on cheapo SSDs - all the
suggestions are anecdotal. Could you run some benchmarks with these
different options and report back the results?

-- 
Sitsofe | http://sucs.org/~sits/
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/