From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Received: (qmail 10557 invoked from network); 10 Jan 2003 21:33:49 -0000 Received: from unknown (HELO crack.them.org) (65.125.64.184) by 209.249.29.67 with SMTP; 10 Jan 2003 21:33:49 -0000 Received: from nevyn.them.org ([66.93.61.169] ident=mail) by crack.them.org with asmtp (Exim 3.12 #1 (Debian)) id 18X8fF-0005HX-00 for ; Fri, 10 Jan 2003 17:34:14 -0600 Received: from drow by nevyn.them.org with local (Exim 3.36 #1 (Debian)) id 18X6mu-0002Gi-00 for ; Fri, 10 Jan 2003 16:34:00 -0500 Date: Fri, 10 Jan 2003 21:33:00 -0000 From: Daniel Jacobowitz To: gdb-patches@sources.redhat.com Subject: Re: [unwind-20030108-branch] Add sentinel-frame + misc Message-ID: <20030110213400.GA8623@nevyn.them.org> Mail-Followup-To: gdb-patches@sources.redhat.com References: <3E1E2D1B.5020100@redhat.com> <3E1E6C23.8050407@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3E1E6C23.8050407@redhat.com> User-Agent: Mutt/1.5.1i X-SW-Source: 2003-01/txt/msg00423.txt.bz2 On Fri, Jan 10, 2003 at 01:45:55AM -0500, Andrew Cagney wrote: > with patch, > > >This adds sentinal-frame.[hc] and then modifies the unwind code in frame.c > >so that it can exploit the fact that there is alway an extra frame beyond > >the inner most frame. This has a dramatic effect on the unwind code - so > >much simpler! > > > >It doesn't do things like ensure that there is always a frame or create > >the inner most frame from the sentinel frame. > > > >To confuse the patch slightly, it also adjusts the id_unwind, pc_unwind > >and register_unwind function types. > > > >Andrew Yeah! I've just got one stylistic comment: > -struct dummy_frame > +struct frame_unwind_cache > { > - struct dummy_frame *next; > + struct frame_unwind_cache *next; > @@ -555,8 +522,14 @@ > If the value isn't here AND a value is needed, try the next inner > most frame. */ > > +struct frame_unwind_cache > +{ > + void *regs[1]; > +}; > + I know it makes some type-checking simpler, but naming all your private types the same thing is a real nuisance when using this lame "GDB" thing on the code afterwards. Must we? -- Daniel Jacobowitz MontaVista Software Debian GNU/Linux Developer