Message-ID: <38306E6D.A38D917F@crosstor.com>
Date: Mon, 15 Nov 1999 13:34:53 -0700
From: Tom Harrington <tph@crosstor.com>
Organization: CrosStor Software, Inc.
Mime-Version: 1.0
To: new-httpd@apache.org
Subject: Re: recent linux kernels broken?
References: <Pine.BSF.4.20.9911150021070.76772-100000@alive.znep.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: new-httpd-owner@apache.org
Precedence: bulk
Reply-To: new-httpd@apache.org

Marc Slemko wrote:
> 
> I seem to see a lot of people submitting bug reports over the past few
> weeks, using a fairly recent 2.2 kernel, about Apache just hanging after
> some time on their systems.
> 
> Any known kernel bugs that could be causing this?  It seems odd that this
> pops up all of a sudden, all with very similar linux kernel versions...

I don't know where known kernel bugs would be tracked online.  But I have
noticed one case of apparently-buggy networking behavior with 2.2.x
kernels.  The test case I saw was a pair of programs that implemented
UDP send/receive, the sort of thing used in an introductory network
programming class.  In _some_ cases, a call to "recvfrom", with no
incoming data would not block, but would instead return an error.  I
don't recall the error code right now, but it was unexpected behavior
that didn't match documentation.

This happened on 2.2.6 on a PowerPC system and 2.2.12-20smp on some
recent x86-series clone.  It did not happen on 2.0.36 on Pentium, or
(as a sanity check) on any version of Solaris I could get hold of.

Apache shouldn't encounter this specific bug, but it's possible that
what I saw was a manifestation of a deeper problem that would affect
Apache.

-- 
Tom Harrington
CrosStor Software, Inc.
tph@crosstor.com