linux-kernel - Re: [PATCH] Give kjournald a IOPRIO_CLASS

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: <20081004141252.GA23398@mit.edu>
Date:	Sat, 4 Oct 2008 10:12:52 -0400
From:	Theodore Tso <tytso@....edu>
To:	Arjan van de Ven <arjan@...radead.org>
Cc:	Andrew Morton <akpm@...ux-foundation.org>,
	Jens Axboe <jens.axboe@...cle.com>,
	linux-kernel@...r.kernel.org, Alan Cox <alan@...rguk.ukuu.org.uk>
Subject: Re: [PATCH] Give kjournald a IOPRIO_CLASS_RT io priority

On Thu, Oct 02, 2008 at 07:33:04AM -0700, Arjan van de Ven wrote:
> > I can whip up a patch which adds some markers which we could use to
> > find out more information about what is happening.
> 
> interesting testcase of the markers concept.

Sorry for the delay, I ran into a minor bug in the Modules.marker
generation support that prevented Systemtap from being able to use
markers.  (It was busted since 2.6.27-rc1, so I guess that gives us
some sense how often developers use Systemtap.  :-)

It looks like Andrew's workaround seems to help you out, but if you're
willing to run this while your mail reader is running, and correlate
it with with the large latency spikes, we might get some interesting
results.

Anyway, here's the patch (against ext4, although could pretty easily
move this to ext3 --- but you can mount an ext3 filesystem as ext4,
although for the moment you do have to run the command "tune2fs -E
test_fs /dev/hdXX" first), and a sample systemtap script.  You'll also
need the markers patch, and of course the latest systemtap from their
git repository.

						- Ted

P.S.  Make sure you install systemtap in /usr/local, instead of trying
to run it out of the build tree.  See for an interesting report from
Roland McGrath about what happens if you make this mistake:

	http://sources.redhat.com/ml/systemtap/2008-q3/msg00809.html

I really think Systemtap has a lot of potential if only they could get
past some "minor usability concerns".  So one thing that I think would
really help the Systemtap folks is if more people gave them more, ah,
"constructive feedback" to their mailing list.

View attachment "add-debugging-markers" of type "text/plain" (3244 bytes)

View attachment "markers-patch" of type "text/plain" (2614 bytes)

View attachment "ext4-marker.stp" of type "text/plain" (775 bytes)