<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
  <head>

    <meta http-equiv="content-type" content="text/html; charset=ISO-8859-15">
  </head>
  <body bgcolor="#ffffff" text="#000000">
    Hello XFS mailinglist,<br>
    <br>
    On my armv5 device I like to use XFS for my simple home made nas,
    because I think it is ideal for low/medium performance CPUs.<br>
    I searched the mailing archive about the usage of XFS on arm
    architecture. I figured out, that the patchset of James Bottomley
    was applied to the main line. So I expected xfs to run properly on
    arm. Unfortunatly I still run into this (known) error after writing
    some data on an xfs partition and remounting it:<br>
    <a href="mailto:xfs@oss.sgi.com"></a>"<br>
    SGI XFS with ACLs, security attributes, large block/inode numbers,
    no debug enabled<br>
    SGI XFS Quota Management subsystem<br>
    XFS mounting filesystem sda1<br>
    Starting XFS recovery on filesystem: sda1 (logdev: internal)<br>
    XFS: xlog_recover_process_data: bad clientid<br>
    XFS: log mount/recovery failed: error 5<br>
    XFS: log mount failed<br>
    "<br>
    <br>
    Am I still forced to use the "hammer" approach (flushing buffers in
    xfs_buf.c) which was proposed in January 2010? Or did I misinterpret
    the logfile of the xfs component in the kernel (so no arm fixing
    patches were applied)?<br>
    <br>
    Is xfs NOW be known to work on arm (e.g. armv5)? If so I like to
    complain. If not, I'm willing to test patches which might solve this
    issue.<br>
    <br>
    Thanks for reading and any comment,<br>
    Ossy<br>
  </body>
</html>