From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: [PATCH 1/3] ftrace: graph of a single function Date: Thu, 4 Dec 2008 09:39:17 +0100 Message-ID: <20081204083917.GF32594@elte.hu> References: <20081203203656.776893226@goodmis.org> <20081203203828.920111079@goodmis.org> <20081203130741.473c1262.akpm@linux-foundation.org> <20081203133242.4e80bbe6.akpm@linux-foundation.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Steven Rostedt Cc: Andrew Morton , linux-kernel@vger.kernel.org, fweisbec@gmail.com, peterz@infradead.org, arjan@infradead.org, dave@linux.vnet.ibm.com, containers@lists.osdl.org, ebiederm@xmission.com, sukadev@linux.vnet.ibm.com, serue@us.ibm.com, srostedt@redhat.com List-Id: containers.vger.kernel.org * Steven Rostedt wrote: > > On Wed, 3 Dec 2008, Andrew Morton wrote: > > > On Wed, 3 Dec 2008 16:10:38 -0500 (EST) > > Steven Rostedt wrote: > > > > > + > > > > > + seq_printf(m, "%s\n", str); > > > > > > > > Can we use %pF here? > > > > > > If there's a way to not print the "+offset". > > > > Could be added, I guess. I wonder if it would be > > reused elsewhere. > > There's lots of places in ftrace that would use it, and probably clean > up a bunch of code in the process. Well, we do eventually want to have a trace_option that extends all function names with the +offset/size portion - and one that switches them to raw RIPs. In rare occasions, when the same function has multiple call sites of the same child function, it can be useful. I ran into such scenarios with the latency tracer and it had this capability to do 'verbose' symbol printing. So plain %pF wont cut it - please abstract out the "print function symbol string" bit within the ftrace infrastructure. And the default trace_option for this should be to print without +offset/size spam, emphatically :) Ingo