From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 30174 invoked by alias); 16 May 2002 18:35:25 -0000 Mailing-List: contact gdb-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sources.redhat.com Received: (qmail 30151 invoked from network); 16 May 2002 18:35:21 -0000 Received: from unknown (HELO neptune.kirkland.local) (12.104.72.61) by sources.redhat.com with SMTP; 16 May 2002 18:35:21 -0000 Received: by neptune.kirkland.local with Internet Mail Service (5.5.2650.21) id ; Thu, 16 May 2002 11:35:19 -0700 Message-ID: <43CB1396676FD4119F03001083FD2994F5F222@neptune.kirkland.local> From: "Kevin \"Squail\" Endres" To: 'Daniel Jacobowitz' , Richard.Earnshaw@arm.com Cc: gdb@sources.redhat.com Subject: RE: cross targeted gdb and corefiles Date: Thu, 16 May 2002 11:35:00 -0000 MIME-Version: 1.0 Content-Type: text/plain X-SW-Source: 2002-05/txt/msg00196.txt.bz2 getting todays source and configuring... keep on timing out- will this change work with the released 5.2? as to the change I am not that familure with gdb source and structures. I am assuming that by "link in corelow/core-regset as target instead of nat objects" you are directing me to alter the NATDEPLIST list in gdb's Makefile - removing all *nat*.o references and add reference to corelow.o and core-regset.o. I am puzzled why post config the first line of NATDEPLIST is commented out... as for "move supply_gregset and supply_fpregset to the tdep file" - move from where(assumed a nat file) to which tdep file(arm-linux-tdep.c or arm-tdep.c??) I appreciate your help and apologize for the "newbie" questions.... :]k -----Original Message----- From: Daniel Jacobowitz [mailto:drow@mvista.com] Sent: Thursday, May 16, 2002 10:42 AM To: Richard.Earnshaw@arm.com Cc: Kevin Squail Endres; gdb@sources.redhat.com Subject: Re: cross targeted gdb and corefiles On Thu, May 16, 2002 at 06:24:53PM +0100, Richard Earnshaw wrote: > > I am using a gdb built from a cvs snapshot from 4-29 that is i386 hosted and > > ARM targeted. > > > > this gdb does not support reading arm cores. I would like to get a backtrace > > for an arm failure (indeterminant - cannot set a breakpoint) without running > > an arm native gdb (or running my application under gdbserver). Is that > > possible?? > > Not at present. > > > > > if not - is it a feature slated to be added to cross targeted gdb soon? I > > know i run my application under gdbserver... > > Maybe. Some targets can now do this. Precisely which target OS are you > using? He's on GNU/Linux, IIRC... Doing cross coredumps is trivial; I put all the BFD machinery in place for a number of GNU/Linux targets, and others did it for the various BSDs. I'd post my ARM patches for this but I doubt even a line of them applies any more after your extensive ARM cleanups. The only thing you have to do is move supply_gregset and supply_fpregset to the tdep file, and then link in corelow/core-regset as target instead of nat objects. This means a little bit of grubbing about in target header files to copy out definitions for regsets, but only a little. -- Daniel Jacobowitz Carnegie Mellon University MontaVista Software Debian GNU/Linux Developer