From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 10526 invoked by alias); 29 Nov 2011 14:30:30 -0000 Received: (qmail 10518 invoked by uid 22791); 29 Nov 2011 14:30:28 -0000 X-SWARE-Spam-Status: No, hits=0.1 required=5.0 tests=AWL,BAYES_00,RP_MATCHES_RCVD,SPF_HELO_PASS X-Spam-Check-By: sourceware.org Received: from lo.gmane.org (HELO lo.gmane.org) (80.91.229.12) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 29 Nov 2011 14:30:11 +0000 Received: from list by lo.gmane.org with local (Exim 4.69) (envelope-from ) id 1RVOhG-0004rY-1A for systemtap@sources.redhat.com; Tue, 29 Nov 2011 15:30:10 +0100 Received: from wardy-en0.manor.fubra.net ([87.124.34.36]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Tue, 29 Nov 2011 15:30:09 +0100 Received: from matthew.ward by wardy-en0.manor.fubra.net with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Tue, 29 Nov 2011 15:30:09 +0100 To: systemtap@sources.redhat.com From: Matthew Ward Subject: usertap kernel read faults - can I ignore them somehow? Date: Tue, 29 Nov 2011 14:30:00 -0000 Message-ID: Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.6; rv:8.0) Gecko/20111105 Thunderbird/8.0 X-IsSubscribed: yes Mailing-List: contact systemtap-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Post: List-Help: , Sender: systemtap-owner@sourceware.org X-SW-Source: 2011-q4/txt/msg00234.txt.bz2 I'm having an issue with our usertap tapset (https://github.com/fubralimited/usertap) on a high-load server where trying to find an IP address for a particular socket is causing systemtap to crash out on a kernel read fault. This tapset always builds and runs on a low-traffic server, but after testing on one of our high-traffic nodes it sometimes builds and runs without a problem, and sometimes refuses to build at all. There's a gist of the error we're getting here: https://gist.github.com/1404868. The relevant line we're getting is: ERROR: kernel read fault at 0x0000000100000264 (addr) near identifier '@cast' at /usr/share/systemtap/tapset/ip.stp:45:11 I presume that that file is called/included by the usertap tapset as it's asking for the IP address associated with a socket, but this seems to be causing an error in certain situations (presumably where it's requesting an IP address for something that doesn't have one). Is there any way I can tell our systemtap script to silence or ignore the call for the IP address if that lookup fails, so that it will start anyway like it does normally? System info: systemtap-runtime-1.4-6.el6_1.2.x86_64 kernel-2.6.32-131.17.1.el6.x86_64 Source code for usertap: https://github.com/fubralimited/usertap/blob/master/src/usertap.stp --- Matthew Ward e: matthew.ward@fubra.com