From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from netops-testserver-3-out.sgi.com ([192.48.171.28] helo=relay.sgi.com) by pentafluge.infradead.org with esmtp (Exim 4.63 #1 (Red Hat Linux)) id 1Ht5hx-0002sJ-CV for kexec@lists.infradead.org; Tue, 29 May 2007 18:42:11 +0100 Message-ID: <465C6667.2020207@sgi.com> Date: Tue, 29 May 2007 10:44:07 -0700 From: Jay Lan MIME-Version: 1.0 Subject: Re: [PATCH] [makedumpfile] Follow debuginfo link of vmlinux file References: <4655C9B3.7060903@sgi.com> <20070529204311oomichi@mail.jp.nec.com> In-Reply-To: <20070529204311oomichi@mail.jp.nec.com> List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: kexec-bounces@lists.infradead.org Errors-To: kexec-bounces+dwmw2=infradead.org+dwmw2=infradead.org@lists.infradead.org To: Ken'ichi Ohmichi Cc: kexec@lists.infradead.org, Bernhard Walle , Dave Anderson Ken'ichi Ohmichi wrote: > Hi Jay, > > Sorry for the late response. > > 2007/05/24 12:44:26 -0700, Jay Lan wrote: >> Bernhard Walle wrote: >>> Hello, >>> >>> * Dave Anderson [2007-05-24 20:31]: >>>> The crash utility hides the details of there being a separate debug >>>> file in the same way the gdb does when working with a binary executable >>>> that has a separate debuginfo file. It tries *not* to be ambiguous, >>>> i.e., the point is to stay true to the "crash vmlinux vmcore" model. >>> It's documented in the GDB documentation: >>> http://sourceware.org/gdb/current/onlinedocs/gdb_16.html#SEC154 >> Thanks for the pointer, Bernhard. I installed kernel-debuginfo rpm >> but crash still failed to find the debug information... I had to >> copy the kernel-.debug to /boot for it to work. >> >> So, i think it is a good idea for makedumpfile to follow the same >> convension as crash and gdb, Ken'ichi? > > No, I don't think it is worthy that a kernel file is specified *only* > for getting the path to a debuginfo file. It is smart to specify > a debuginfo file directly. The crash utility needs both a kernel > file and a debuginfo file, but makedumpfile needs only a debuginfo > file as I said. > > BTW, I recommend using a makedumpfile's CONFIGFILE instead of a > debuginfo file. If a system has a CONFIGFILE, makedumpfile can > run without a debuginfo file. It is smaller than a debuginfo file, > and it is easy to distribute it to each system. > I don't think each system should have a debuginfo file, because the > file is big and a user may analyze a dumpfile on a remote analysis > system. Ohmichi San, Any way to make kernel debugging easier is great news!!! However... If i run makedumpfile -x option to produce a dumpfile as below: # makedumpfile -c -d 31 -x VMLINUX VMCORE DUMPFILE The size of these files are: 16468 -rw------- 1 jlan wheel 16983716 2007-05-29 10:23 DUMPFILE 1719048 -r-------- 1 jlan wheel 1760303684 2007-05-24 11:35 VMCORE 55488 -rwxr-xr-x 1 jlan wheel 56819470 2007-05-24 11:33 VMLINUX and crash was able to read and analyze the DUMPFILE together with the VMLINUX. Now when i used makedumpfile -g option to generate a CONFIGFILE: # makedumpfile -g CONFIGFILE -x VMLINUX the size of CONFIGFILE is only 696 bytes. Besides, makedumpfile failed to generate a DUMPFILE using the CONFIGFILE: # makedumpfile -c -d 31 -i CONFIGFILE VMCORE DUMPFILE CONFIGFILE and VMCORE don't match. makedumpfile Failed. I think the CONFIGFILE is too small to be good? What did i miss here? Thanks, - jay > > > Thanks > Ken'ichi Ohmichi _______________________________________________ kexec mailing list kexec@lists.infradead.org http://lists.infradead.org/mailman/listinfo/kexec