xfs
[Top] [All Lists]

direct-io regression [Was: How to track down abysmal performance ata - r

To: Valdis.Kletnieks@xxxxxx, josef@xxxxxxxxxx, chris.mason@xxxxxxxxxx
Subject: direct-io regression [Was: How to track down abysmal performance ata - raid1 - crypto - vg/lv - xfs]
From: Dominik Brodowski <linux@xxxxxxxxxxxxxxxxxxxx>
Date: Thu, 5 Aug 2010 11:31:00 +0200
Cc: Michael Monnerie <michael.monnerie@xxxxxxxxxxxxxxxxxxx>, Christoph Hellwig <hch@xxxxxxxxxxxxx>, linux-raid@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, dm-devel@xxxxxxxxxx
In-reply-to: <15446.1280953986@localhost>
Mail-followup-to: Dominik Brodowski <linux@xxxxxxxxxxxxxxxxxxxx>, Valdis.Kletnieks@xxxxxx, josef@xxxxxxxxxx, chris.mason@xxxxxxxxxx, Michael Monnerie <michael.monnerie@xxxxxxxxxxxxxxxxxxx>, Christoph Hellwig <hch@xxxxxxxxxxxxx>, linux-raid@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, dm-devel@xxxxxxxxxx
References: <20100804073546.GA7494@xxxxxxxxxxxxxxxxxxxxxxxxxx> <20100804085039.GA11671@xxxxxxxxxxxxx> <20100804091317.GA27779@xxxxxxxxxxxxxxxxxx> <20100804092122.GA2998@xxxxxxxxxxxxx> <20100804073546.GA7494@xxxxxxxxxxxxxxxxxxxxxxxxxx> <201008041116.09822@xxxxxx> <20100804102526.GB13766@xxxxxxxxxxxxxxxxxx> <15446.1280953986@localhost>
User-agent: Mutt/1.5.20 (2009-06-14)
Hey,

when attempting to track down insufficient I/O performance, I found the
following reression relating to direct-io on my notebook, where an
ata device, which consists of several partitions, is combined to a lvm
volume, and one logical volume is then encrypted using dm-crypt. Test case
was the following command:

$ dd if=/dev/mapper/vg0-root_crypt of=/dev/zero iflag=direct bs=8k count=131072

2.6.34 results in ~16 MB/s,
2.6.35 results in ~ 3.1 MB/s

The regression was bisected down to the follwoing commit:

commit c2c6ca417e2db7a519e6e92c82f4a933d940d076
Author: Josef Bacik <josef@xxxxxxxxxx>
Date:   Sun May 23 11:00:55 2010 -0400

    direct-io: do not merge logically non-contiguous requests
    
...

How to fix this? I do not use btrfs, but ext3 (and the access was down on
the block level, not on the fs level, so this btrs-related commit should not
cause such a regression).

Best,

        Dominik

<Prev in Thread] Current Thread [Next in Thread>