<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="/rss.xsl.xml"?>
<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
    <title>Changes in run_ckckp</title>
    <link>http://nxr.netbsd.org/rss/src/regress/sys/fs/lfs/ckckp/run_ckckp</link>
    <description></description>
    <language>en</language>
    <copyright>Copyright 2005</copyright>
    <generator>Java</generator>
    
<item>
    <title>Include testing of the roll-forward agent, when recreating the filesystem<br/>partial-segment by partial-segment.  Each checkpoint should pass fsck_lfs -n<br/>without errors; the results of fsck_lfs -p on non-checkpoints should also<br/>pass fsck_lfs -n without errors.</title>
    <description>/src/regress/sys/fs/lfs/ckckp/run_ckckp - 1.4</description>
    <pubDate>Fri Jul 21 00:29:23 UTC 2006</pubDate>
    <dc:creator>perseant</dc:creator>
</item>

<item>
    <title>Allow the user to specify a file size in the ckckp regression test, including<br/>the value zero (filling the filesystem with directories and empty file<br/>inodes).</title>
    <description>/src/regress/sys/fs/lfs/ckckp/run_ckckp - 1.3</description>
    <pubDate>Fri May 05 19:38:30 UTC 2006</pubDate>
    <dc:creator>perseant</dc:creator>
</item>

<item>
    <title>Regression test improvements:<br/><br/>Move the stop for LFCNWRAPSTOP to the point at which writing at segment 0<br/>is really about to commence, since this is what the test expects (and<br/>incidentally what a snapshotting utility wants as well).<br/><br/>More correctly reconstruct the on-disk state at every checkpoint, rather<br/>than relying on the entire state at the point of wrapping to be accurate<br/>(that is only true the first time we wrap).  Add a "make abort" target to<br/>make rerunning the test more convenient when it has failed and we're done<br/>analyzing the failure.</title>
    <description>/src/regress/sys/fs/lfs/ckckp/run_ckckp - 1.2</description>
    <pubDate>Sat Apr 22 00:10:55 UTC 2006</pubDate>
    <dc:creator>perseant</dc:creator>
</item>

<item>
    <title>Introduce two fcntl calls that freeze the filesystem right at the point<br/>where segment 0 is being considered for writing.  This allows for automated<br/>checkpoint vailidity scanning, and could be used (in conjunction with the<br/>existing LFCNREWIND) for e.g. snapshot dumps as well.<br/><br/>Include a regression test that does such scanning.<br/><br/>When writing the Ifile, loop through the dirty block list three times to<br/>make sure that the checkpoint is always consistent (the first and second<br/>times the Ifile blocks can cross a segment boundary; not so the third time<br/>unless the segments are very small).  Discovered by using the aforementioned<br/>regression test.</title>
    <description>/src/regress/sys/fs/lfs/ckckp/run_ckckp - 1.1</description>
    <pubDate>Mon Apr 17 20:02:35 UTC 2006</pubDate>
    <dc:creator>perseant</dc:creator>
</item>
</channel></rss>

