From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 7923 invoked by alias); 13 Apr 2007 17:42:42 -0000 Received: (qmail 7912 invoked by uid 22791); 13 Apr 2007 17:42:41 -0000 X-Spam-Check-By: sourceware.org Received: from wr-out-0506.google.com (HELO wr-out-0506.google.com) (64.233.184.224) by sourceware.org (qpsmtpd/0.31) with ESMTP; Fri, 13 Apr 2007 18:42:35 +0100 Received: by wr-out-0506.google.com with SMTP id i22so826089wra for ; Fri, 13 Apr 2007 10:42:33 -0700 (PDT) Received: by 10.114.110.1 with SMTP id i1mr1038407wac.1176486152674; Fri, 13 Apr 2007 10:42:32 -0700 (PDT) Received: by 10.114.52.6 with HTTP; Fri, 13 Apr 2007 10:42:32 -0700 (PDT) Message-ID: Date: Fri, 13 Apr 2007 17:42:00 -0000 From: "Alex Gonzalez" To: "Alex Gonzalez" , gdb@sourceware.org Subject: Re: Invalid hex digit 59 - Linux MIPS kernel debugging In-Reply-To: <20070413172923.GA26847@caradoc.them.org> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <20070413150929.GA20853@caradoc.them.org> <20070413172923.GA26847@caradoc.them.org> X-IsSubscribed: yes Mailing-List: contact gdb-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sourceware.org X-SW-Source: 2007-04/txt/msg00095.txt.bz2 Thanks a lot, that was very helpful. Setting the architecture to mips:isa32 indeed works. And you're right in that the 32bits kernel is running on an RM9000 with 64 bits architecture. Will the CVS version work without having to set the architecture or will I still have to do so? Regards, Alex On 4/13/07, Daniel Jacobowitz wrote: > On Fri, Apr 13, 2007 at 05:40:25PM +0100, Alex Gonzalez wrote: > > Thanks, I'll try that. > > > > If you don't mind me asking, why would gdb expect 64bits registers > > when the target abi is set to o32? > > Because you can run o32 binaries on hardware with 64-bit registers, > and there are some ugly corner cases where the upper half of registers > can cause problems - e.g. for things which have undefined behavior if > the input registers are not correctly sign extended. You've probably > built a 32-bit kernel for a processor that GDB knows is actually > 64-bit. But if your debug agent only supplies 32 bits, then GDB will > get along fine (once it knows that). > > If you don't want to change GDB versions, an appropriate "set arch" > might work - but I recommend the current snapshots. > > -- > Daniel Jacobowitz > CodeSourcery >