From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 7285 invoked by alias); 3 Jan 2013 01:45:26 -0000 Received: (qmail 7274 invoked by uid 22791); 3 Jan 2013 01:45:25 -0000 X-SWARE-Spam-Status: No, hits=-2.7 required=5.0 tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_ENVFROM_END_DIGIT,FREEMAIL_FROM,KHOP_RCVD_TRUST,KHOP_THREADED,RCVD_IN_DNSWL_LOW,RCVD_IN_HOSTKARMA_YE X-Spam-Check-By: sourceware.org Received: from mail-la0-f43.google.com (HELO mail-la0-f43.google.com) (209.85.215.43) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 03 Jan 2013 01:45:20 +0000 Received: by mail-la0-f43.google.com with SMTP id eg20so7167808lab.16 for ; Wed, 02 Jan 2013 17:45:18 -0800 (PST) MIME-Version: 1.0 Received: by 10.112.82.166 with SMTP id j6mr19446470lby.25.1357177518466; Wed, 02 Jan 2013 17:45:18 -0800 (PST) Received: by 10.112.11.38 with HTTP; Wed, 2 Jan 2013 17:45:18 -0800 (PST) In-Reply-To: <50E48011.9010603@redhat.com> References: <50E00806.2090906@tu-dortmund.de> <50E48011.9010603@redhat.com> Date: Thu, 03 Jan 2013 01:45:00 -0000 Message-ID: Subject: Re: Unknown symbol error using stap in 2.6.31 on ARM From: Cipher Chen To: Josh Stone Cc: systemtap@sourceware.org Content-Type: text/plain; charset=ISO-8859-1 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: 2013-q1/txt/msg00001.txt.bz2 On Thu, Jan 3, 2013 at 2:44 AM, Josh Stone wrote: > On 12/30/2012 04:40 AM, Cipher Chen wrote: >> On Sun, Dec 30, 2012 at 5:23 PM, Alexander Lochmann wrote: >>> try to run stap with "--compatible=1.8". >>> Stap needs uprobe support for some builtin functions as weel. Using this >>> option it will use old fashion builtin function. >> >> I'm not sure this would help, since I've already tried version 1.8 >> and snapshot 20122222(?), too(I mean re-install the whole systemtap). >> And it didn't work either.(current used is 2.0) >> Anyway, I'll try that again tomorrow. > > I gather from the name "ntfs_watching" that you're not actually trying > to probe any userspace? In that case, Alex is referring to the > probefunc() change, which --compatible=1.8 ought to fix. I also > recently added a ppfunc() to git which acts like the old way, parsing > the function name out of pp(). > > In any case, it would help if you shared the script you're using, or at > least enough of it that triggers the issue. > In fact, that wasn't a real utrace patch for ARM. I downloaded utrace.patch for 2.6.31, patched it and wrote some similarly functions for ARM by myself which were needed when building the kernel, since that's a really old kernel. ntfs_watching.stp is a simple script, the only user-space probe would be: probe process("/path/to/ntfsmount/").function("*") { printf("%s\n", probefunc()) } Thanks, Cipher > Thanks, > Josh