2006-02-13-is-sleep-effected-by-time-changes.html

<!--# set var="title" value="Is sleep(3) effected by time changes?" -->
<!--# set var="date" value="2006-02-13" -->

<!--# include file="include/top.html" -->

<p>It’s an interesting question. I tried it; it isn’t. Here’s my guess at an explanation:</p>

<p><a href="http://www.tin.org/bin/man.cgi?section=3&amp;topic=sleep">sleep(3)</a> is implemented internally on most systems as <a href="http://www.tin.org/bin/man.cgi?section=2&amp;topic=alarm">alarm(2)</a> and <a href="http://www.tin.org/bin/man.cgi?section=2&amp;topic=sigsuspend">sigsuspend(2)</a>. This means that this is an in-kernel question, not a userspace question. However, we know from <a href="http://www.tin.org/bin/man.cgi?section=3&amp;topic=clock_gettime">clock_gettime(3)</a> that the kernel has multiple internal clocks. CLOCK_MONOTONIC is defined as:</p>

<blockquote>
  <p>Clock that cannot be set and represents monotonic time since some unspecified starting point.</p>
</blockquote>

<p>This must be what’s used for alarm(), which makes sense.</p>

<p>On a related note:</p>

<p>Those annoying programs that behave very oddly when you change the system time are broken. They’re using <a href="http://www.tin.org/bin/man.cgi?section=2&amp;topic=gettimeofday">gettimeofday(2)</a> or something that derives from it (i.e. <a href="http://www.tin.org/bin/man.cgi?section=2&amp;topic=time">time(2)</a>) instead of properly calling clock_gettime(CLOCK_MONOTONIC).</p>

<!--# include file="include/bottom.html" -->