From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 29219 invoked by alias); 11 May 2006 14:53:56 -0000 Received: (qmail 29209 invoked by uid 22791); 11 May 2006 14:53:55 -0000 X-Spam-Check-By: sourceware.org Received: from nevyn.them.org (HELO nevyn.them.org) (66.93.172.17) by sourceware.org (qpsmtpd/0.31.1) with ESMTP; Thu, 11 May 2006 14:53:53 +0000 Received: from drow by nevyn.them.org with local (Exim 4.54) id 1FeCY1-0001oH-TA; Thu, 11 May 2006 10:53:50 -0400 Date: Thu, 11 May 2006 15:00:00 -0000 From: Daniel Jacobowitz To: Brendan Kehoe Cc: gdb@sourceware.org Subject: Re: support for multiple ARM register banks in GDB Message-ID: <20060511145349.GA6923@nevyn.them.org> Mail-Followup-To: Brendan Kehoe , gdb@sourceware.org References: <44634F3D.1050401@zen.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <44634F3D.1050401@zen.org> User-Agent: Mutt/1.5.8i X-IsSubscribed: yes Mailing-List: contact gdb-help@sourceware.org; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sourceware.org X-SW-Source: 2006-05/txt/msg00142.txt.bz2 On Thu, May 11, 2006 at 03:50:37PM +0100, Brendan Kehoe wrote: > Hi folks, > > Is anyone already looking at GDB's port to the ARM, and in particular > what it believes is the single register bank on the chip? For > processors like the ARM7TDMI, there are a few register banks depending > on the current mode of the ARM. In order to debug (not C++) exceptions > on an ARM embedded target, it's necessary to be able to look at a > specific bank's register like r13_irq. This sort of stuff appears to be > in place in the simulator, but not in GDB itself. I've implemented banked register support, as a sample application of the XML-based register description work I've posted about several times. The GDB client and the RDI-based GDB stub we ship in the 2006-Q1 CodeSourcery release includes support for banked registers. Of course, if your target isn't RDI-based, that doesn't directly help you. But it shouldn't be hard to do for a custom debug stub either. (It took all of an hour or two once the XML bits were in place, so I'd say it was a successful test run). -- Daniel Jacobowitz CodeSourcery