From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752370AbaLRO27 (ORCPT ); Thu, 18 Dec 2014 09:28:59 -0500 Received: from mx1.redhat.com ([209.132.183.28]:40614 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751341AbaLRO26 (ORCPT ); Thu, 18 Dec 2014 09:28:58 -0500 Date: Thu, 18 Dec 2014 15:28:27 +0100 From: Jiri Olsa To: Peter Zijlstra Cc: mingo@kernel.org, linux-kernel@vger.kernel.org, Frederic Weisbecker , Steven Rostedt , Linus Torvalds Subject: Re: [PATCH] perf ftrace: Factor regs retrieval for function tracer Message-ID: <20141218142827.GB18898@krava.brq.redhat.com> References: <20141216115041.GW3337@twins.programming.kicks-ass.net> <20141217143105.GG18257@krava.brq.redhat.com> <20141217153508.GC30905@twins.programming.kicks-ass.net> <20141217154507.GA20471@krava.brq.redhat.com> <20141218140847.GA18898@krava.brq.redhat.com> <20141218142027.GE30905@twins.programming.kicks-ass.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20141218142027.GE30905@twins.programming.kicks-ass.net> User-Agent: Mutt/1.5.23 (2014-03-12) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Dec 18, 2014 at 03:20:27PM +0100, Peter Zijlstra wrote: > On Thu, Dec 18, 2014 at 03:08:47PM +0100, Jiri Olsa wrote: > > diff --git a/kernel/trace/trace_event_perf.c b/kernel/trace/trace_event_perf.c > > index 6fa484de2ba1..54bffedd28b6 100644 > > --- a/kernel/trace/trace_event_perf.c > > +++ b/kernel/trace/trace_event_perf.c > > @@ -304,7 +304,7 @@ perf_ftrace_function_call(unsigned long ip, unsigned long parent_ip, > > { > > struct ftrace_entry *entry; > > struct hlist_head *head; > > - struct pt_regs regs; > > + struct pt_regs **regsp, *regs = pt_regs; > > int rctx; > > > > head = this_cpu_ptr(event_function.perf_events); > > @@ -316,16 +316,24 @@ perf_ftrace_function_call(unsigned long ip, unsigned long parent_ip, > > > > BUILD_BUG_ON(ENTRY_SIZE > PERF_MAX_TRACE_SIZE); > > > > - perf_fetch_caller_regs(®s); > > + /* > > + * The ftrace_ops is registered with SAVE_REGS_IF_SUPPORTED, > > + * so if we got pt_regs defined, we don't need to retrieve > > + * regs buffer through perf_trace_buf_prepare. > > + */ > > + regsp = pt_regs ? NULL : ®s; > > > > - entry = perf_trace_buf_prepare(ENTRY_SIZE, TRACE_FN, NULL, &rctx); > > + entry = perf_trace_buf_prepare(ENTRY_SIZE, TRACE_FN, regsp, &rctx); > > if (!entry) > > return; > > > > + if (regsp) > > + perf_fetch_caller_regs(regs); > > + > > > Would not something like: > > struct pt_regs *regs; > > entry = perf_trace_buf_prepare(ENTRY_SIZE, TRACE_FN, ®s, &rctx); > if (!pt_regs) { > perf_fetch_caller_regs(regs); > pt_regs = regs; > } > > Be simpler? well, I wanted to omit touching the '__perf_regs' var in perf_trace_buf_prepare: if (regs) *regs = this_cpu_ptr(&__perf_regs[*rctxp]); jirka