From mint-bounce@lists.fishpool.fi Fri Feb 22 13:03:52 2008 From: Eero Tamminen Organization: Koti To: mint@lists.fishpool.fi Subject: Re: [MiNT] Automatic -fomit-frame-pointer usage Date: Fri, 22 Feb 2008 19:45:02 +0200 User-Agent: KMail/1.9.6 Cc: Vincent =?iso-8859-1?q?Rivi=E8re?= , mint References: <47BECA07.8020506@freesbee.fr> In-Reply-To: <47BECA07.8020506@freesbee.fr> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Message-Id: <200802221945.02775.oak@helsinkinet.fi> X-ecartis-version: Ecartis v1.0.0 Sender: mint-bounce@lists.fishpool.fi Errors-to: mint-bounce@lists.fishpool.fi X-original-sender: oak@helsinkinet.fi Precedence: bulk List-help: List-unsubscribe: List-Id: X-List-ID: List-subscribe: List-owner: List-post: Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by mail.sparemint.org id m1MI3lX7030612 Hi, On Friday 22 February 2008, Vincent Rivière wrote: > The usage of the frame pointer (actually a6) allows gdb to determine > precisely the call stack when a breakpoint is hit. It is very useful for > debugging, And for callgraphs which help a lot in both memory and performance analysis. :-) > that's why the frame pointer is always used when -O is not > used. When some level of optimization is used, it is very hard to use the > debugger, because the instructions are usually reordered, most variable > values are inaccessible because they are stored in registers. FYI: GCC 4.x and later Gdb versions support locationlists which are supposed to help with this. > And without the frame pointer, it may be impossible to know from which > functions we are called. - Eero