<?xml version="1.0" encoding="UTF-8" ?>
<rss version="2.0" xmlns:admin="http://webns.net/mvcb/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:creativeCommons="http://backend.userland.com/creativeCommonsRssModule" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:html="http://www.w3.org/1999/html" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:slash="http://purl.org/rss/1.0/modules/slash/"><channel><title>Ross Burton</title><link>http://www.burtonini.com/blog</link><description>A potted account of Ross' life</description><language>en</language><ttl>60</ttl><dc:creator>Ross Burton</dc:creator><admin:generatorAgent rdf:resource="http://pyblosxom.sourceforge.net/"/><admin:errorReportsTo rdf:resource="mailto:ross+web@burtonini.com"/><item><title>In Print</title><guid isPermaLink="false">computers/devilspie/linux-magazine-2006-08-14-20-36</guid><link>http://www.burtonini.com/blog/computers/devilspie/linux-magazine-2006-08-14-20-36</link><description>Those nice people at Linux Magazine have written an article covering Devil's Pie . Apart from the traditional mis-spelling of ...</description><content:encoded><![CDATA[    <p>
      Those nice people at <a href="http://www.linux-magazine.com/">Linux
      Magazine</a> have written <a
      href="http://www.linux-magazine.com/issue/70/Brightside_Devils_Pie.pdf">an
      article covering Devil's Pie</a>.  Apart from the traditional mis-spelling
      of my surname, it's a great article.  Thanks Linux Magazine!
    </p>
]]></content:encoded><category domain="http://www.burtonini.com">/computers/devilspie</category><dc:date>2006-08-14T19:36:00Z</dc:date></item><item><title>Lack of Entropy</title><guid isPermaLink="false">computers/cups-2006-08-14-18-00</guid><link>http://www.burtonini.com/blog/computers/cups-2006-08-14-18-00</link><description>A while back I upgraded CUPS on my NAS device to CUPS 1.2, and as my configuration file had bitrotted ...</description><content:encoded><![CDATA[    <p>
      A while back I upgraded CUPS on my NAS device to CUPS 1.2, and as my
      configuration file had bitrotted I decided to erase it and start again.
      It was looking good, CUPS detected the USB printer like they said it
      would, but then it just stopped working.  I had to upgrade to a SSL
      connection to perform administration tasks, and it just wasn't working.
      <tt>strace</tt> shows the daemon was just running a <tt>select</tt> loop,
      I had no idea what the problem was.
    </p>
    <p>
      Today after a bit more searching the solution hit me.  It's trying to
      create a self-signed SSL key, and SSL keys need random data.  Now, what
      creates random data on Linux?  Mouse activity (the NAS has no mouse),
      keyboard activity (no keyboard), and disk activity (it gets used once a
      week).  A quick poke with sysctl proved my problem:
    </p>
    <pre>$ sysctl  kernel.random.entropy_avail
kernel.random.entropy_avail = 8</pre>
    <p>
      Hm, eight bytes with which to create a SSL key from, that just won't work.
      Solution: <tt>find /</tt>.  I ran that in one terminal with
      <tt>sysctl</tt> in another, and could watch the entropy pool get filled
      with IDE data, to be used by CUPS.  After twenty seconds of disk activity,
      the certificate was finally generated. Yay, a working printer!
    </p>
    <p>
      So, for Google and anyone else out there: if you see <tt>Generating SSL
      server key...</tt> in the CUPS log and then CUPS hangs, check your entropy
      levels.
    </p>]]></content:encoded><category domain="http://www.burtonini.com">/computers</category><dc:date>2006-08-14T17:00:00Z</dc:date></item><item><title>Intel Rockness</title><guid isPermaLink="false">computers/colour-2006-08-14-12-20</guid><link>http://www.burtonini.com/blog/computers/colour-2006-08-14-12-20</link><description>Whilst trying to get the modesettings branch of the Intel X driver working, I noticed this in the LCD probe: ...</description><content:encoded><![CDATA[    <p>
      Whilst trying to get the <tt>modesettings</tt> branch of the Intel X
      driver working, I noticed this in the LCD probe:
    </p>
    <pre>II) I810(0): redX: 0.569 redY: 0.342   greenX: 0.312 greenY: 0.544
(II) I810(0): blueX: 0.149 blueY: 0.132   whiteX: 0.313 whiteY: 0.329</pre>
    <p>
      IBM (this X60 is still branded IBM, not Lenovo) deserve some serious props
      for actually setting all of the data provided by the EDID spec.  I best I
      should hack up a tool to generate ICC profiles from the primaries
      specified at startup.
    </p>
]]></content:encoded><category domain="http://www.burtonini.com">/computers</category><dc:date>2006-08-14T11:20:00Z</dc:date></item></channel></rss>