[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
- To: misc_(_at_)_openbsd_(_dot_)_org
- Subject: PID reuse
- From: "Johan Lindh" <johan_(_at_)_linkdata_(_dot_)_se>
- Date: Mon, 13 Jan 2003 18:14:20 +0100
- Reply-to: johan_(_at_)_linkdata_(_dot_)_se
While building and testing courier-40.2 on my OpenBSD-3.2-RELEASE machine,
I ran into a problem.
It's described in detail at http://www.linkdata.se/sourcecode.html
The short version is that some programs (courier among them) depend
on the fact that the same PID must not be generated twice within the space of
I've looked around for more info about OpenBSD's PID generation behaviour,
but all I found was that it was random, and that random PID's are A Good Thing
But I can't recall seeing this problem in earlier versions of OpenBSD, and checking
the Changelog for OpenBSD 3.1 I found this statement:
"Simplify PID selection algorithm."
Is this an indication that PID selection is broken for some applications
since 3.1, or is it simply a coincidence?
Given that (many?) programs seem to depend on a certain PID assigment behaviour,
would it be prudent to conform to it, or are there strong and compelling reasons for
keeping PID generation random?
Courier is found at http://www.courier-mta.org/
Visit your host, monkey.org