From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 30643 invoked by alias); 23 Sep 2006 03:26:08 -0000 Received: (qmail 30635 invoked by uid 22791); 23 Sep 2006 03:26:08 -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; Sat, 23 Sep 2006 03:26:06 +0000 Received: from drow by nevyn.them.org with local (Exim 4.54) id 1GQy9P-0007Bc-VN; Fri, 22 Sep 2006 23:26:00 -0400 Date: Sat, 23 Sep 2006 03:26:00 -0000 From: Daniel Jacobowitz To: Jim Blandy Cc: Mark Kettenis , Michael.Snyder@palmsource.com, gdb-patches@sourceware.org Subject: Re: [RFC] h8300 "info registers" fix Message-ID: <20060923032559.GA27572@nevyn.them.org> Mail-Followup-To: Jim Blandy , Mark Kettenis , Michael.Snyder@palmsource.com, gdb-patches@sourceware.org References: <0F4DF2E21F33DD46BE7B8CDEEB0E16D307EAB2@ussunex01.palmsource.com> <20060913200917.GA20574@nevyn.them.org> <0F4DF2E21F33DD46BE7B8CDEEB0E16D307EAB3@ussunex01.palmsource.com> <20060913203043.GB21009@nevyn.them.org> <200609230105.k8N15Hf7015324@elgar.sibelius.xs4all.nl> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.13 (2006-08-11) X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2006-09/txt/msg00158.txt.bz2 On Fri, Sep 22, 2006 at 08:21:12PM -0700, Jim Blandy wrote: > > Mark Kettenis writes: > > Not weird at all; it's the same convention we use for storing values. > > > > Bet most debug interfaces you've seen lately actually use > > little-endian byte order. They're frequently organized in integer multiples of the word size, which are documented to return in host uint32_t's. If the underlying "register" is really bigger than that, then the stub is responsible for composing. > No, they're actually documented to return them in host byte order. > You read a register, you get an integer. > > (Well, Daniel's seen more than me, but this is so of the ones I have > seen.) Yep. I can think offhand of four different JTAG-related debug interfaces I've worked with in the last six months, all of which did this. Then, we transform the number from host endianness (on the machine running the stub) to target endianness (the device) before passing it over TCP/IP to GDB. All a bit messy, but you lose some no matter which order you pick :-) -- Daniel Jacobowitz CodeSourcery